Thursday, 2015-02-26

« Wednesday, 2015-02-25 Index Friday, 2015-02-27 »
jamielennoxclarkb: it's not supposed to contain 2 entries for compute00:00
morganfainbergjamielennox, i think that is what we need to say, you've asked for something we found many of, be more specific. that can be either novaclient's job or the user's job00:00
clarkbjamielennox: then keystone should throw an error here00:00
morganfainbergclarkb, ^00:00
clarkbjamielennox: amd we will all yell at rax until they fix it00:00
jamielennoxclarkb: it contains one entry and then novaclient can go off, query that and then figure out where to go next00:00
*** woodster_ has quit IRC00:00
jrollI don't see why returning two endpoints is so crazy, personally00:00
morganfainbergclarkb, i'm fine with going and yelling at RAX about this.00:00
jamielennoxclarkb: the problem was people started jamming like /v2.0 and /v3 and volumev2 and other crap in it00:00
jroll(not in defense of rackspace's abomination of a catalog)00:00
*** MarkAtwood has joined #openstack-infra00:01
morganfainbergjroll, because the api we defined said there'd be one service with that name. and we're just dealign with the fallout of not being specific about the catalog data definition00:01
mordredclarkb: I'm going to fix up shrews' patch for your comments real quick00:01
clarkbmordred: ok00:01
jamielennoxmordred: ++00:01
mordredclarkb: I think for the second one, perhaps an inline todo comment - because I also don't know the answer to your question00:01
clarkbmordred: wfm00:01
jamielennoxmorganfainberg: to be honest i don't know how endpoint filtering works here..00:01
*** abhirc has quit IRC00:02
morganfainbergjamielennox, for RAX it doesn00:02
morganfainberg't00:02
morganfainbergthey don't use keystone ;)00:02
jamielennoxmorganfainberg: in general00:02
morganfainbergoh00:02
morganfainbergyou mean our catalog filering thing? i'll discuss that in -keystone, doesn't need to bug poor -infra people00:02
openstackgerritMerged openstack-infra/shade: Import from v2 instead of v1_1  https://review.openstack.org/15750900:02
jamielennoxi looked at it ages ago, i think it does the right thing00:02
openstackgerritMerged openstack-infra/shade: Do not allow to pass *-cache on init  https://review.openstack.org/15775200:02
openstackgerritMonty Taylor proposed openstack-infra/shade: Add support for volume attach/detach  https://review.openstack.org/15362300:04
mordredclarkb: ^^00:04
*** baoli has quit IRC00:04
*** yamamoto has joined #openstack-infra00:05
clarkbmordred: +2 thanks00:05
*** doug-fish has left #openstack-infra00:05
mordredclarkb: yay!00:05
*** woodster_ has joined #openstack-infra00:07
clarkbmordred: I think I have reviewed all changes that don't otherwise hae -1's00:08
clarkb(in shade)00:08
mordredclarkb: yes. you are a machine and I am going to hand you a very cuddly aardvark00:09
*** sabeen1 has quit IRC00:09
*** dangers is now known as dangers_away00:09
mordredclarkb: or, you know, something more pleasant than that00:09
mordredunless you want to try making smoked aardvark ...00:09
clarkbjeblair: see my wip comment on 158891, yay for self reviewes00:09
clarkbmordred: smoked aadvark does sound intriguing00:10
*** gordc has quit IRC00:10
morganfainbergsmoked aardvark00:10
morganfainberg... interesting ...00:11
*** sdake_ has joined #openstack-infra00:11
jeblairclarkb: silly me for believing the commit message!  :)00:12
*** sdake_ has quit IRC00:12
jeblairclarkb: sometimes they are more like theorems that need to be proved, rather than axioms... :)00:12
*** sdake_ has joined #openstack-infra00:12
*** gyee has joined #openstack-infra00:13
clarkbjeblair: indeed :)00:13
*** cpallares has quit IRC00:13
mordredjeblair: what if we started writing them that way?00:13
*** david-lyle is now known as david-lyle_afk00:14
clarkbmordred: 158177 appears to be missing a depends on for something that installs the new ansible-puppet role to /etc/ansible/roles00:15
clarkbmordred: or am I missing something with how ansible works?00:15
morganfainbergmordred, you know your terminology is spreading, i just caught myself saying crazypants00:16
anteayacrazypants00:16
mordredclarkb: looking00:17
morganfainberganteaya, i see what you did there...00:17
SpamapSwell this is fun00:17
rhe00can you guess stand another CI question from me? :)00:17
SpamapSI hadn't used hpcloud since they deprecated the legacy cloud00:18
rhe00guys, not guess00:18
anteayamorganfainberg: can't get anything past you00:18
SpamapSand the new us west refuses to associate floating ips with instances00:18
clarkbmordred: we have the new repo it just doesn't seem to have anything that installs it, maybe 158177 should do that with a vcsrepo?00:18
mordredclarkb: you are right - there is a missing depends on -00:18
clarkbmordred: or whatever tool ansible wants you to use for that00:18
mordrednope -Ive got a patch for it00:18
clarkbmordred: is that patch merged? if not I will -1 for the depends on00:18
anteayarhe00: I'm not a guy00:18
mordredclarkb: https://review.openstack.org/#/c/158175/00:19
mordredclarkb: please -1 for the depends on00:19
clarkbSpamapS: did you create a router, a network, a subnet, and attach them all and so on?00:19
rhe00anteaya: sorry. scratch the gender specification. :)00:19
mordredclarkb: I think I depends-on'd the wrong thing00:19
mordredSpamapS: yeah - you have to do all the neutron things00:19
clarkbmordred: done00:19
rhe00I have a simple noop-check-communication job set up: http://paste.openstack.org/show/182175/00:20
clarkbSpamapS: its much easier to do via horizon because getting order right via the cli is almost impossible00:20
SpamapSI thought I did00:20
SpamapSand I did do it via horizon yes00:20
clarkbSpamapS: did you attach your router to Ext-Net?00:20
clarkbSpamapS: floating ip won' work without that iirc00:20
SpamapSclarkb: it doesn't show as an option00:20
anteayarhe00: that is more likely you to get teh results you are after, yeah00:20
rhe00jenkins complains that it can't find any nodes with the label '' when it is trying to build the job00:20
clarkband router needs to be attached to your internal network as well00:20
anteayarhe00: can we get a stacktrace00:21
anteayarhe00: we love stacktraces00:21
SpamapSclarkb: when I say "add interface" to the router I only get the private network00:21
openstackgerritMonty Taylor proposed openstack-infra/puppet-ansible: Remove embedded ansible role  https://review.openstack.org/15817700:21
mordredclarkb: ^^00:21
*** Sukhdev has joined #openstack-infra00:21
mordredSpamapS: I've been meaning to finish the ansible network modules so that I can just make an ansible playbook that does this for you00:22
mordredbut, you know, one thing at a time00:22
SpamapSyeah I've done it via cli before in tripleo clouds00:23
SpamapSshared networks like ext-net should be showing in the add interface, but they're not00:23
rhe00anteaya: http://paste.openstack.org/show/182181/00:23
*** mbacchi has joined #openstack-infra00:24
SpamapSfound it00:24
SpamapSclarkb: it's entirely different attaching to ext-net as it is attaching to internal doh00:24
clarkbmordred: that just made a loop00:24
clarkbSpamapS: of course, its neutron00:24
clarkbSpamapS: this would be another reason I am a proponent of nova network00:25
SpamapSwe..00:25
clarkbSpamapS: neutron gives you all the control but I need almost noen of it00:25
SpamapSreally need to fix this cloud thing00:25
SpamapSlike this should just happen00:25
SpamapSI didn't ask for the floating ip association in hopes that I'd find out if the router wasn't available.00:25
mordredSpamapS: your 158906 is in merge conflict, btw00:26
mordredSpamapS: also ++ to fix this cloud thing00:26
clarkbmordred: so 158175 needs to remove its depends on now00:26
SpamapSoh this is amazing00:27
SpamapSall this and now I see that ec2 metadata failed00:27
SpamapSso no ssh keys for you00:27
*** marun has quit IRC00:27
clarkbSpamapS: yes, I highly recommend that dib thing00:27
clarkbSpamapS: :) though apparently there is a way to enable config drive00:27
anteayarhe00: I'm not seeing the error, sorry00:27
*** zz_dimtruck is now known as dimtruck00:28
anteayarhe00: I can't find the word node in that paste00:28
SpamapSHow can ec2 metadata not work? :-P00:28
clarkbSpamapS: it turns out when you do DNAT without SNAT rules it almost never works00:29
clarkbSpamapS: last time I was debugging ec2 metadata trouble I was pretty sure that ebtables was killing asynchronously routed and sourced packets00:29
rhe00anteaya: I see that in the jenkins GUI. Trying to find a text log showing the message so that I can copy it to a paste.00:30
SpamapSclarkb: not sure what you're saying. Why would ebtables come into the picture?00:31
openstackgerritMerged openstack-infra/shade: Add support for volume attach/detach  https://review.openstack.org/15362300:32
SpamapSthis should be packets going from my vm, to the overlay, to the gateway, hitting DNAT, to the neutron-metadata-agent, to nova-api00:32
anteayarhe00: oh okay thanks00:32
clarkbSpamapS: yes and when they don't return by the same path things break00:32
SpamapSclarkb: are you saying that the source IP is wrong, so it drops the replies?00:32
clarkbSpamapS: and let me tell you they almost certainly do not return by the same path00:33
SpamapSThats amazing. :-P00:33
anteayarhe00: I can't see anything in your stacktrace to tell me where the next clue is00:33
*** sdake_ has quit IRC00:33
SpamapSclarkb: ok I get it, so the neutron-metadata-agent machine inside hpcloud should be SNAT'ing00:34
SpamapSso that nova-api just replies to it.00:34
*** hdd has joined #openstack-infra00:34
*** sdake_ has joined #openstack-infra00:34
bhunter71only as it might pertain,   if you include diskimage-builder/cloud-init-datasources, you can define any combination of export `DIB_CLOUD_INIT_DATASOURCES`="Ec2, ConfigDrive, OpenStack"00:34
SpamapSreally weird00:34
*** tmcpeak has joined #openstack-infra00:34
*** shashankhegde has quit IRC00:34
clarkbSpamapS: at least that was about where I got to when debugging these type of issues in the gate. We only get DNAT rules by default, this means return packets go via default route, packets drop00:35
mordredclarkb: DRAT00:35
rhe00anteaya: I will dig a bit more and see if I can get a bit more information to share. thanks!00:35
mordredclarkb: no - I meant to do the original thing00:35
*** tmcpeak has quit IRC00:35
clarkbmordred: the original thing is wrong though00:35
mordrednope00:35
mordredit's right00:35
clarkbmordred: you are saing use this stuff over here and its not managed at all00:35
anteayarhe00: sure, artifacts are good, so far I don't know enough of what you are experiencing to offer anything except emotional support00:35
mordredyes. that's right- because they both put files in the same place00:35
clarkbmordred: you ma need a third change00:35
anteayarhe00: go you00:35
jeblairSpamapS, jhesketh: i spent about an hour today reviewing 137664.  I think it's a correct fix, with a possibly more correct variant.  let me know what you think.  It's so old, perhaps we should just merge it and then I should propose my change on top of it.00:36
mordredclarkb: so, the first one stops putting it in place, but does not delete it00:36
mordredthe second one starts putting it in place again00:36
clarkbmordred: I think you need, puppet stops managing /etc/ansbile/roles, new change manages /etc/ansible/roles, last change uses /etc/ansible/roles00:36
*** tmcpeak has joined #openstack-infra00:36
jeblair(it's the shortest patches that always take the most time to review, isn't it?)00:36
mordredclarkb: but it uses /etc/ansible/role already00:36
SpamapSjeblair: Somebody needs to go get a pHd figuring out how to reliably unit test for race conditions. :-P00:37
clarkbmordred: so maybe thats why this is confusing00:37
clarkbmordred: we should've done that last :)00:37
mordredclarkb: we've been using it for months00:37
mordredit's the place where things go00:37
openstackgerritRamy Asselin proposed openstack-infra/infra-specs: Add spec for reusable OpenStack CI solution  https://review.openstack.org/13974500:37
mordredwe've just been installing files there that were embedded in a puppet module00:37
clarkbmordred: didn't we just change to use /etc/ansible/roles00:37
clarkbwe were using a library ibefore00:37
clarkbanyways I can update my votes00:37
mordredhang on - let me go make sure I'm not high00:38
mordredmulti-repo dependency switches like this are hard for a reason :)00:38
openstackgerritDiego proposed openstack-infra/git-review: Small changes to better support HTTPS scheme including the following: Add support for passwords via gitreview.password attribute in http environments Added message on how to configure password in gitreview attribute Blank password if using ssh as scheme O  https://review.openstack.org/15930500:38
krotscheckAnyone knows who else out there uses apscheduler?00:38
jeblairkrotscheck: zuul and nodepool do00:38
krotscheckOn SQLA?00:38
jeblairkrotscheck: nodepool does00:39
krotscheckjeblair: Any special handling I have to do to not have it yell at me because of unicode reasons?00:39
jeblairkrotscheck: i haven't run into that, but i'm not sure why the two would interact00:39
*** tmcpeak has quit IRC00:40
krotscheckjeblair: I don't either. It started happening with a recent release of... ugh, _something_ in our db dependencies. I didn't think to take a snapshot.00:40
*** jaypipes has quit IRC00:40
jheskethjeblair: yeah your comment makes sense, but I agree its not overly harmful as it is. Up to you if you want to merge and reiterate or re-roll00:41
SpamapSjeblair: I took about 10 minutes to review it, with 9:30 of that spent drawing a time line on a piece of paper. I think you're right that the counters will be off because my piece of paper doesn't have the counters considered.00:41
mordredclarkb: https://review.openstack.org/#/c/158177/2/manifests/init.pp <-- this is where we stop having puppet instal stuff into /etc/ansible/roles - but it was being used before00:41
jeblairkrotscheck: previous jenkins runs might have "pip freeze" output in the logs, so you might be able to compare versions00:41
*** yamamoto has quit IRC00:42
*** mbacchi has quit IRC00:42
jeblairjhesketh: cool, i'll flag that change to remind me to write the follow up patch tomorrow, and approve it after i have done so :)00:42
jheskethshiny00:42
jeblairSpamapS: you are fast.  you should review all changes involving race conditions!  (though in fairness, i did keep getting interrupted and having to start over)00:43
SpamapSjeblair: I am fast and not careful. ;)00:43
SpamapSjeblair: the timeline thing is the only way I've ever been able to do it. Started with upstart jobs. :-P00:44
clarkbmordred: so we are good then with the old deps? and I should just jump in without worry? :)00:44
openstackgerritMonty Taylor proposed openstack-infra/puppet-ansible: Remove embedded ansible role  https://review.openstack.org/15817700:44
jeblairSpamapS: i'm a fan, i've done the same sort of thing before00:44
mordredclarkb: I've updated them - the deps are now all correct00:45
mordredclarkb: or, rather, the deps are all as intended00:45
mordredclarkb: it is possible that as part of reviewing you will find that I'm wrong :)00:45
clarkbmordred: ok looking again00:45
mordredclarkb: because honestly, it's not my FAVORITE transition, but I'm not sure there is a great way00:46
*** aysyd has joined #openstack-infra00:46
mordredclarkb: how does keypair extension work for us on hpcloud if ec2 metadata doesn't ever work00:47
openstackgerritClint 'SpamapS' Byrum proposed openstack-infra/shade: Utilize dogpile.cache for caching  https://review.openstack.org/15890600:47
mordredclarkb: because we're not using config-drive in production yet00:47
SpamapSok, with that, time to unjack from the matrix for a bit00:47
mordredSpamapS: the matrix will miss you00:47
*** yamamoto has joined #openstack-infra00:47
SpamapSand later: time to make puppet my bitch00:48
clarkbmordred: it often did not work, broken metadata service is a very common reason for our image builds to fail00:48
clarkbmordred: but with dib in hpcloud that has mostly been negated for a subset of the images00:48
mordredclarkb: oh righ t- but because our image build build in keys00:48
mordredwe _only_ notice it on image builds00:48
clarkbmordred: SpamapS it should be noted I think the ec2 thing isn't a 100% break because it depends on host specific routing00:48
mordredsee - this is the answer - build in keys00:48
mordred_totally_ scalable00:49
clarkbthat said, for a long time centos6 in my personal account could not boot ever with working ec200:49
SpamapSmordred: as we most recently learned.. scalable isn't always super _usable_ ;)00:49
clarkb Ifiled at least one ticket about it and never got it fixed00:49
clarkbgave over lots of uuids in the support channel too00:49
mordredSpamapS: ++00:49
* clarkb reads ansible-galaxy docs00:50
*** aysyd has quit IRC00:54
clarkbmordred: http://docs.ansible.com/galaxy.html#advanced-control-over-role-requirements-files doesn't tell me what -f does, so then I did http://paste.openstack.org/show/182193/ and now I am going to go get a beer00:54
*** yamamoto has quit IRC00:55
clarkbgreghaynes: SpamapS ^ you know ansible what does -f do?00:56
greghaynesclarkb: number of parallel processes to use00:56
clarkbgreghaynes: sorry with ansible-galaxy00:57
greghaynes-f, --force           Force overwriting an existing role00:57
greghaynes*shrug*00:57
greghayneshelpful greg is helpful00:58
*** SumitNaiksatam has quit IRC00:58
clarkbok so I think mordred needs a -r then according to the docs?00:58
*** salv-orlando has quit IRC01:00
mordredclarkb: BLEH01:00
mordredthis is what I get for writing this on a plane01:00
*** mtanino has quit IRC01:01
*** shashankhegde has joined #openstack-infra01:02
*** oomichi has joined #openstack-infra01:04
mordredclarkb: ansible-galaxy install -r roles.yaml01:05
mordredit turns out01:05
openstackgerritMonty Taylor proposed openstack-infra/system-config: Install standalone ansible roles  https://review.openstack.org/15817501:05
openstackgerritMonty Taylor proposed openstack-infra/system-config: Install standalone ansible roles  https://review.openstack.org/15817501:07
openstackgerritMonty Taylor proposed openstack-infra/system-config: Install standalone ansible roles  https://review.openstack.org/15817501:07
mordredok - emprical testing says it does not matter, but the docs say the file needs to end in .yml - so I changed that too01:07
greghayneswat, .yml01:07
mordredansible likes its yaml files to have a .yml extension usually01:08
mordredI have just learned to not fight01:08
clarkbpretty sure ingy said its yaml once01:08
mordredhowever, I did test that it works fine with .yaml too in this case01:09
clarkball other suffixes are therefore less correct01:09
greghaynesI think yaml is one of those formats that define the file extension01:09
greghaynesand it is defined as the color blue01:09
morganfainbergmordred, when / if you get a chance see if https://review.openstack.org/#/c/159314/ makes shade with your wonky catalog better. it *should* work.01:09
morganfainbergmordred, or well just in general the environment workign with nodes in RAX for -infra :)01:09
mordredmorganfainberg: how do I consume that?01:10
morganfainbergmordred, easiest way, checkout keystoneclient, and apply the 1 line diff before installing01:10
mordrednono - lemme ask again01:10
*** dims_ has joined #openstack-infra01:11
morganfainbergoh. how do you consume the thing from ksc's catalog?01:11
*** dims_ has quit IRC01:11
mordredwould the programmatic consuption module then be "for endpoint in get_Endpoint(): try: do something; break; except EndpointNotFound: continue" ?01:11
*** dims_ has joined #openstack-infra01:11
morganfainberguhm..01:12
*** dims has quit IRC01:12
morganfainbergmordred, well it should work like it did before - we're just collapsing the list of endpoints if 2 services claim to both be "compute" (you know... what you'd expect)01:13
mordredmorganfainberg: http://git.openstack.org/cgit/openstack-infra/shade/tree/shade/__init__.py#n17601:13
mordredthis is my workaround right now01:13
mordredmorganfainberg: namely, I pass in a service_name instead of service_type01:13
morganfainbergmordred, yeah that works01:14
morganfainbergthat will still work01:14
mordredyes01:14
mordredI'm being partially daft because sometimes things like "you know... what you'd expect" may not be what I expect01:14
morganfainbergwhat we're doing is in the case of passing a service type, we just say "any service that claims to be type='x'" will have it's endpoints added to the service catalog list01:15
morganfainbergso if you say i want type=compute, you get everything that is type=compute01:15
*** MarkAtwood has quit IRC01:15
mordredand then novaclient will do the right thing with that list of things?01:15
morganfainbergit should01:15
mordredcool. I will try that and verify01:15
mordredas long as _I_ don't have to do anything new01:16
mordred:)01:16
morganfainbergnope01:16
morganfainbergnot breaking what you're doing01:16
* mordred groks now01:16
morganfainbergfixing what you were trying to do01:16
morganfainbergbtw: https://review.openstack.org/#/c/159315/ jamie's is a little better01:16
morganfainbergbut same net effect01:16
morganfainberg[he even does tests in his patch]01:16
morganfainbergnow rax case is still a bit insane01:16
morganfainbergbecause compute != compute01:16
morganfainbergbut that is something i'm happy to go yell at someone about01:17
*** cpallares has joined #openstack-infra01:17
morganfainbergbecause type == 'compute' except when i say type == 'compute' is really something else, isn't fair to users :P01:17
morganfainberg[or us developers when we try and make something work]01:17
*** gyee has quit IRC01:18
*** patrickeast has quit IRC01:22
*** cinerama has quit IRC01:23
*** rlandy has quit IRC01:25
*** marun has joined #openstack-infra01:25
*** ashleighfarnham has quit IRC01:25
*** cinerama has joined #openstack-infra01:27
*** rwsu is now known as rwsu-afk01:27
*** otter768 has joined #openstack-infra01:30
*** armax has joined #openstack-infra01:30
*** marun has quit IRC01:31
*** Sukhdev has quit IRC01:33
*** sambetts has quit IRC01:34
*** otter768 has quit IRC01:34
*** sambetts has joined #openstack-infra01:35
*** dims has joined #openstack-infra01:37
*** asettle-mtg is now known as asettle01:39
*** dims_ has quit IRC01:40
*** melwitt has quit IRC01:41
dimsis git slow for anyone else?01:41
*** bdpayne has quit IRC01:43
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Replaced python-crontab with apscheduler  https://review.openstack.org/15895501:43
*** sigmavirus24 is now known as sigmavirus24_awa01:44
krotscheckCould it be feasible to create a gh-pages style repository in the post queue?01:49
krotscheckand/or maintain one?01:49
*** ddieterly has joined #openstack-infra01:49
krotscheck(I know, I know, it's one of those WHyyyyyy questions)01:49
krotscheck(humor me)01:50
mordredkrotscheck: so - I just recently hit this myself01:51
krotscheckmordred: Oh? Context?01:51
*** tiswanso has joined #openstack-infra01:51
mordredkrotscheck: yeah - I decided that I did not want to host my free-software talks on github01:51
mordredbecause, you know, free software01:51
mordredbut Id' been using their gh-pages feature :)01:51
krotscheckAnd now they're on inaugust.01:52
mordredkrotscheck: yah - I decided an rsync script to my server was just fine01:52
mordredANYWHO - let me aska  follow up questio01:52
krotscheckI can anticipate the question, and have a followup anser.01:52
mordredis the thing you want to publish some arbitrary html? or is it to commit some html to a git branch?01:52
*** tiswanso has quit IRC01:53
*** tiswanso has joined #openstack-infra01:53
krotscheckI want to publish code in a way that can be consumed by bower. Bower makes a couple of assumtpions.01:54
krotscheck1- That versions are tagged.01:54
mordredcool. that part is easy!01:54
krotscheck2- That anything in a tag is "released" code that's gone through minification/concatenation.01:54
mordredah01:54
krotscheckThat's the tricky bit.01:54
krotscheckIt's gh-page-ish01:54
krotscheckInsofar as the content of that tag has NOTHING to do with the actual source.01:55
krotschecks/source/master/01:55
mordredyeah - so you want to $do_something and have the results of that go into a branch with a tag applied, potentially01:55
krotscheckyep01:55
mordredor - you want to $do_something and put that into a branch01:55
mordredand from time to time, you might want to tag that branch01:55
krotscheckThe option of "hey let's just create a repo that's just for release code"01:55
* mkj_wk crosses his fingers, hopes to have found the bottom of the rabbit hole01:55
krotscheckis also available01:55
mordredyah01:55
krotscheckmordred: I think your latter example might be better....?01:56
*** shashankhegde has quit IRC01:56
mordredkrotscheck: well, congratulations, I have not considered that particular combination of actions01:56
openstackgerritJoshua Hesketh proposed openstack-infra/infra-manual: Document Cross-Repo Dependencies  https://review.openstack.org/15885501:56
krotscheckmordred: Well, it's a thing that javascript people do because their release code is relatively easily diffable.01:56
mordredand it's an interesting question - I _think_ it might bear a passing resemblence to some of the things git-buildpackage --pristine-tar wants to do01:56
mordredkrotscheck: and the assumption is that you want to consume someone else's minification, rather tahn to locally minify on build yourself, yeah?01:57
krotscheckcert sig fail01:57
krotscheckmordred: Well, sortof. What I want to build is a javascript library that has all kinds of individual service files that map to, say, nova services. And then have those minified into ng_nova.js, etc.01:58
krotscheckAnd release those.01:58
krotscheckSo that someone else could roll a pure-JS UI on top of services that are tested in gate.01:59
mordredoh - sorry, I totally failed at writing english01:59
mordredwhat I meant to ask01:59
mordredis that "what javascript developers consuming things in this manner want to do is consume your pre-minified things, rather than downloading your expanded things and minifying them themselves"01:59
*** ghostpl_ has quit IRC02:00
mordredpronouns are hard02:00
krotscheckmordred: Yes.02:00
*** che-arne has quit IRC02:00
mordredawesome. I'm learning things02:00
krotscheckInclude one file. Do neat nova thigns from javascript.02:00
* krotscheck is going to fight the mod_headers battle later.02:00
mordredso - we've thus-far avoided having gerrit/zuul commit or push code to itself02:00
clarkbabd they expect to consume via git? my first thought was use a tarball but seems like git is important02:00
mordredlargely because of skynet02:00
mordredclarkb: yah. git seems important here02:00
krotscheckclarkb: npm runs on tarballs. bower runs off git.02:01
mordredkrotscheck: off the top of my head, I have a three suggestions02:01
mordredeach varying degrees of time and suck02:01
krotscheckdetails on how that works: http://bower.io/docs/creating-packages/02:02
*** ddieterly has quit IRC02:02
*** ddieterl_ has joined #openstack-infra02:02
krotscheckIn this particular context, npm gets our build tools, and we don't care about that.02:02
krotscheckOk, so what are the options?02:03
mordredEASIEST is to manually push a review of minified output to a branch each time before you want to release, have that be approved and land that. it takes 0 extra stuff, but also is the most manual and doesn' leverage much of anything02:03
krotscheckmordred: That seems.... prone to human error.02:03
*** dannywilson has joined #openstack-infra02:03
mordredSLIGHTLY HARDER is to do the same thing that translations and requirements sync do - which is to have a job that responds to a post queue, runs a script to generate a commit and submit that to gerrit for review02:03
mordredthe logic we have there knows how to overwrite an outstanding review02:03
mordredso you could ignore most of the reviews on the publish branch02:04
mordreduntil you cared02:04
*** sdake_ has quit IRC02:04
krotscheckThat seems saner.02:04
mordredit takes a job on a trusted worker - but there is a clear pattern of other jobs taht do that that should be very easy to work from02:04
*** SergK has quit IRC02:05
mordredone could also keep the idea of a team who can release things02:05
krotscheckGiven that these services won't be CD-style, and attached to the openstack release cadence...02:05
mordredas the people who could approve +2 thigns on the release branch in addition to being the people who can tag things02:05
mordredyah02:05
krotscheckWhat's option 3?02:06
mordredHARDEST is tied to the currently-as-yet-non-existent zuul-pushes-merges-back-to-gerrit which is where we have zuul push rather than click the submit button02:06
*** sdake_ has joined #openstack-infra02:06
mordredand would be extending that, when it exists, to allow for pushing arbitrary modified things to arbitrary locations02:06
*** armax has quit IRC02:06
mordredthat would be kinda cool, and I think would enable some interesting dpkg workflows too02:07
*** vilobhmm1 has quit IRC02:07
mordredbut good golly not going to happen this month or next02:07
krotscheckInteresting.02:07
krotscheckSo that would involve teaching zuul how to git?02:07
mordredwell, actually, zuul already gits which is one of the reasons we've talked about this02:08
*** vilobhmm has joined #openstack-infra02:08
mordredit already pre-merges everythign for you before your job, and then passes the pre-merged ref to the jobs02:08
mordredso one of the ideas is to, at job success, just have zuul push the ref back to gerrit, rather than asking gerrit to do the merge again02:08
clarkbalso would avoid gerrit jgit merge weirdness02:08
mordredyup02:09
mordredbut we've not taken the step to robots that can directly push code yet02:09
mordredjust mainly because, well, honestly, skynet02:09
mordredwe have a lot of machines aroudn here02:09
mordredI'm pretty sure theyre going to kill us one day02:09
*** Ryan_Lane has joined #openstack-infra02:09
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Replaced python-crontab with apscheduler  https://review.openstack.org/15895502:10
clarkbmordred shoelaces kill my robot02:10
clarkbso keep your shoelaces habdy02:10
mordredclarkb: I mean, they're not going to kill us tomorrow :)02:11
krotscheckYes, well, I want nova.js02:11
mordred++02:11
mordredkrotscheck: btw - as you work on that ...02:11
krotscheckAnd I'm perfectly willing to teach robots how to push git.02:11
krotscheckmordred: Well, before you get ahead of yourself, I still have a storyboard thing that I need to, you know, justify the existence of.02:12
mordredkrotscheck: I'd make sure to look at shade - it's ugly in places, but has a bunch of ugly knowledge about novas in the real world02:12
mordredkrotscheck: there is that02:12
krotscheckmordred: Though I know a bunch of people in the next cube over would LOVE having js things that talk to nova.02:12
*** dannywilson has quit IRC02:13
mordredkrotscheck: I do not understand why such things do not exist already02:13
*** mayurig has quit IRC02:13
*** dannywilson has joined #openstack-infra02:13
krotscheckmordred: I do. Explaining that however might involve 'inflammatory language'.02:13
mordred:)02:13
clarkbbecause smartos docker by joyent02:14
clarkbwhich isnt really docker02:14
krotscheckWait what?02:14
clarkband isnt linux either02:14
krotscheckThis does not seem related.02:14
mordredkrotscheck: if you do work on such a thing at some point, it would be neat if it could also consume the same config format that os-client-config is consuming02:14
clarkbjoyent has a cloud thing02:14
clarkbthey node.js02:14
krotscheckOh.02:15
greghayneshahah02:15
krotscheckYes.02:15
greghaynessmart datacenter02:15
clarkbclearly js people only do joyent cloud02:15
krotscheckAnd the community has taken their ball and gone over to io.js02:15
mordredI learn more things today ...02:15
krotscheckmordred: https://iojs.org/en/index.html02:15
mordredso - this tells me that node has been forked then?02:15
clarkbyes02:15
krotscheckmordred: Yep.02:16
krotscheckTwice, actually, this is the most recent one.02:16
mkj_wkargh, still can't comment further on https://storyboard.openstack.org/#!/story/883176 -- particularly annoying since previous comment I made is much in need of update with new information... ☺02:16
greghaynesArent they also trying to form a node foundation or somesuch to get it under control?02:16
mordredso when do we move etherpad to not-node02:16
clarkbgreghaynes yes they made a foundation already02:16
*** cpallares has quit IRC02:16
krotscheckmkj_wk: We're looking for someone willing to fix the paging on storyboard's comment threads. Think you have some time for that?02:16
*** weshay has quit IRC02:17
clarkbmordred well who knows aiui theybarent really compatible02:17
*** koolhead17 has joined #openstack-infra02:17
mkj_wkkrotscheck: I just want to be able to post a comment... :P02:17
*** che-arne has joined #openstack-infra02:17
krotscheckmkj_wk: Are you getting an error, or is the comment just not showing up?02:18
mkj_wkkrotscheck: I suppose you are telling me that I've been posting various comments over and over and they are all there but not showing up?02:18
dstufftclarkb: io.js isn't compat with node.js?02:18
krotscheckmkj_wk: It's likely.02:18
krotscheckvery likely02:18
*** dannywilson has quit IRC02:18
mkj_wkkrotscheck: ugh, I guess I'll quit trying then...02:18
mkj_wkkrotscheck: thanks!02:19
clarkbdstufft aiui not really anymore but I could be wrong02:19
krotscheckmkj_wk: Looking forward to seeing your patch!02:19
dstufftclarkb: cool, so node gets their own 2 vs 3 divide02:19
mkj_wklast two failing test cases fail only in whole suite, not when running individual tests02:20
mkj_wksince it depends on sys.stdout.isatty() value I have no idea how it used to pass :)02:20
clarkbdstufft I read things like npm compatible but node node compatible02:20
dstufftclarkb: I'm not sure what "npm" compatible means02:20
clarkbdstufft so possible that they are judging against a subset of code? I dunno02:20
dstufftis that like being pip compatible02:20
clarkbexactly :P02:21
* mkj_wk hopes @mock.patch(sys.stdout) is his friend ☺02:22
clarkbbut reallh dont take my word its very much from an outsider who doesnt follow too closely02:22
*** vilobhmm has quit IRC02:24
clarkbat the very least it will be a super set. new modules, new fsatures, newer v8 etc02:24
*** ashleighfarnham has joined #openstack-infra02:26
*** thedodd has joined #openstack-infra02:28
*** baoli has joined #openstack-infra02:29
*** ddieterl_ has quit IRC02:30
mkj_wkyeah, passed test suite locally02:31
mkj_wkcleaning up commit...02:31
*** hdd has quit IRC02:32
*** ashleighfarnham has quit IRC02:33
*** tsg has quit IRC02:33
*** armax has joined #openstack-infra02:34
*** Ryan_Lane has quit IRC02:37
*** dalgaaf has quit IRC02:37
*** chlong_ has joined #openstack-infra02:43
*** markvoelker has quit IRC02:47
*** Shrews has quit IRC02:47
*** markvoelker has joined #openstack-infra02:47
*** carl_baldwin has joined #openstack-infra02:48
openstackgerritMichael Johnson proposed openstack-infra/git-review: Choose correct branch for rebase when submitting  https://review.openstack.org/15887702:49
*** stevemar has joined #openstack-infra02:49
*** markvoelker has quit IRC02:51
*** yamahata has quit IRC02:52
*** koolhead17 has quit IRC02:53
*** sdake_ has quit IRC02:54
*** ddieterly has joined #openstack-infra02:58
*** sdake_ has joined #openstack-infra02:59
*** jerryz has joined #openstack-infra03:00
*** unicell has quit IRC03:00
*** ddieterly has quit IRC03:02
*** davidlenwell has quit IRC03:05
*** rm_work has quit IRC03:10
*** davidlenwell has joined #openstack-infra03:11
*** carl_baldwin has quit IRC03:13
*** carl_baldwin has joined #openstack-infra03:15
*** ghostpl_ has joined #openstack-infra03:15
*** baoli has quit IRC03:16
*** dims has quit IRC03:24
*** hdd has joined #openstack-infra03:24
*** armax has quit IRC03:26
*** sdake_ has quit IRC03:27
*** sdake_ has joined #openstack-infra03:28
*** ghostpl_ has quit IRC03:29
*** otter768 has joined #openstack-infra03:31
*** coolsvap_ is now known as coolsvap03:33
*** baoli has joined #openstack-infra03:33
*** otter768 has quit IRC03:36
*** shrews has joined #openstack-infra03:40
*** mayurig has joined #openstack-infra03:41
*** markvoelker has joined #openstack-infra03:43
*** yamamoto_ has joined #openstack-infra03:43
*** shrews is now known as Shrews03:46
*** mayurig has quit IRC03:49
*** cpallares has joined #openstack-infra03:50
*** carl_baldwin has quit IRC03:51
*** mayurig has joined #openstack-infra03:52
*** Shrews has quit IRC03:53
*** Shrews has joined #openstack-infra03:53
*** tsg has joined #openstack-infra03:57
*** carl_baldwin has joined #openstack-infra03:58
*** baoli has quit IRC03:59
*** baoli has joined #openstack-infra03:59
*** jerryz has quit IRC04:04
*** baoli has quit IRC04:04
*** wendar has quit IRC04:04
*** wendar has joined #openstack-infra04:04
*** baoli has joined #openstack-infra04:05
*** jerryz has joined #openstack-infra04:05
*** jvrbanac has quit IRC04:05
*** jlk has quit IRC04:05
*** nijaba has quit IRC04:06
*** jlk has joined #openstack-infra04:06
*** jvrbanac has joined #openstack-infra04:06
*** armax has joined #openstack-infra04:06
*** MarkAtwood has joined #openstack-infra04:07
*** nijaba has joined #openstack-infra04:08
*** nijaba has joined #openstack-infra04:08
*** MarkAtwood has quit IRC04:09
*** baoli has quit IRC04:09
*** thedodd has quit IRC04:10
*** jlk has quit IRC04:11
*** jlk has joined #openstack-infra04:11
*** shashankhegde has joined #openstack-infra04:12
*** shashankhegde has quit IRC04:13
*** mayurig has quit IRC04:13
*** carl_baldwin has quit IRC04:14
openstackgerritIan Wienand proposed openstack-infra/system-config: Use nodepool 'generate-log-config' command to generate log config  https://review.openstack.org/15645704:15
*** thedodd has joined #openstack-infra04:18
*** thedodd has quit IRC04:18
*** sdake__ has joined #openstack-infra04:18
*** signed8bit_ZZZzz has quit IRC04:21
*** sdake_ has quit IRC04:22
*** dims has joined #openstack-infra04:25
*** sdake__ has quit IRC04:25
*** yamamoto_ has quit IRC04:26
*** ghostpl_ has joined #openstack-infra04:26
*** spzala has quit IRC04:26
*** bhunter71 has quit IRC04:27
*** Ryan_Lane has joined #openstack-infra04:30
*** dims has quit IRC04:30
*** ghostpl_ has quit IRC04:31
*** MarkAtwood has joined #openstack-infra04:32
*** patrickeast has joined #openstack-infra04:33
*** jerryz has quit IRC04:35
*** rm_work|away has joined #openstack-infra04:36
*** rm_work|away is now known as rm_work04:36
*** rm_work has joined #openstack-infra04:36
*** jerryz has joined #openstack-infra04:36
*** markvoelker has quit IRC04:44
*** markvoelker has joined #openstack-infra04:45
*** deepakcs has joined #openstack-infra04:47
*** markvoelker has quit IRC04:49
openstackgerritBharat Kumar Kobagana proposed openstack-infra/project-config: Run tempest-dsvm-full-glusterfs job on ubuntu  https://review.openstack.org/15921704:50
deepakcsAJaeger, good morning04:52
*** MarkAtwood has quit IRC04:52
openstackgerritBharat Kumar Kobagana proposed openstack-infra/project-config: Run tempest-dsvm-full-glusterfs job on ubuntu  https://review.openstack.org/15921704:53
*** ivar-laz_ has joined #openstack-infra04:56
*** ivar-laz_ has quit IRC04:56
*** coolsvap is now known as coolsvap_04:57
*** patrickeast has quit IRC04:58
*** ivar-lazzaro has quit IRC04:59
*** ddieterly has joined #openstack-infra05:00
*** unicell has joined #openstack-infra05:00
*** sdake_ has joined #openstack-infra05:01
*** yamahata has joined #openstack-infra05:01
*** sdake__ has joined #openstack-infra05:03
*** xyang1 has quit IRC05:03
*** coolsvap_ is now known as coolsvap05:04
*** ddieterly has quit IRC05:04
*** jamespage has quit IRC05:05
*** gnuoy has quit IRC05:05
*** sdake_ has quit IRC05:05
*** gnuoy has joined #openstack-infra05:06
*** jamespage has joined #openstack-infra05:06
*** sdake_ has joined #openstack-infra05:06
*** penguinRaider has quit IRC05:09
*** sdake__ has quit IRC05:09
*** garyh has quit IRC05:09
*** penguinRaider has joined #openstack-infra05:10
*** asselin_ has quit IRC05:11
*** unicell has quit IRC05:15
*** Sukhdev has joined #openstack-infra05:15
*** jcoufal has joined #openstack-infra05:18
*** tiswanso has quit IRC05:19
*** carl_baldwin has joined #openstack-infra05:19
*** woodster_ has quit IRC05:20
*** BharatK has joined #openstack-infra05:20
*** cpallares has quit IRC05:24
*** otter768 has joined #openstack-infra05:32
*** otter768 has quit IRC05:37
*** yamamoto_ has joined #openstack-infra05:38
*** jcoufal has quit IRC05:39
*** jcoufal_ has joined #openstack-infra05:39
*** koolhead17 has joined #openstack-infra05:41
*** carl_baldwin has quit IRC05:42
*** che-arne has quit IRC05:42
*** jcoufal_ has quit IRC05:44
*** oomichi has quit IRC05:47
*** abhisheks has joined #openstack-infra05:47
*** che-arne has joined #openstack-infra05:49
*** achanda has quit IRC05:50
openstackgerritTimothy R. Chavez proposed openstack-infra/nodepool: Write the node UUID to /etc/nodepool/uuid  https://review.openstack.org/15933905:53
openstackgerritTimothy R. Chavez proposed openstack-infra/project-config: Print instance UUID to log  https://review.openstack.org/15625205:54
*** ParsectiX has quit IRC05:59
*** ParsectiX has joined #openstack-infra06:00
*** ddieterly has joined #openstack-infra06:01
*** sarob has quit IRC06:01
*** yfried_ has quit IRC06:01
*** sarob has joined #openstack-infra06:03
*** ParsectiX has quit IRC06:04
*** ParsectiX has joined #openstack-infra06:04
*** ddieterly has quit IRC06:05
*** ParsectiX has quit IRC06:05
*** woodster_ has joined #openstack-infra06:06
openstackgerritTimothy R. Chavez proposed openstack-infra/project-config: Print instance UUID to log  https://review.openstack.org/15625206:07
*** sarob has quit IRC06:08
*** ghostpl_ has joined #openstack-infra06:11
*** wenlock has quit IRC06:14
abhisheksi am installing master for master for my ci and updating vars.sh file, so do i have to change UPSTREAM_GERRIT_HOST_PUBLIC_KEY with my new gerrit key also??06:15
*** ghostpl_ has quit IRC06:15
*** tobiash has quit IRC06:18
*** achanda has joined #openstack-infra06:19
*** tobiash has joined #openstack-infra06:19
*** che-arne has quit IRC06:19
*** tobiash has quit IRC06:20
*** tobiash has joined #openstack-infra06:22
openstackgerritAndrey Pavlov proposed openstack-infra/project-config: Add rally job to ec2-api project  https://review.openstack.org/15913806:23
*** koolhead17 has quit IRC06:25
*** unicell has joined #openstack-infra06:25
*** fifieldt has joined #openstack-infra06:26
*** sdake_ has quit IRC06:29
*** sarob has joined #openstack-infra06:30
*** BharatK has quit IRC06:31
*** thingee has quit IRC06:35
*** sarob has quit IRC06:41
*** SumitNaiksatam has joined #openstack-infra06:43
*** BharatK has joined #openstack-infra06:44
*** cnesa has joined #openstack-infra06:49
*** mrunge has joined #openstack-infra06:51
*** megan_garcia has joined #openstack-infra06:51
megan_garciai hope someone online can help me out:/ Im getting tortured by JJB.. so apparently if i have a shell builder script, that uses awk '{print $3}'.. JJB is failing to create such job, because it thinks 'print $3' is a key mapping to it :((06:53
megan_garciaso annoying06:53
*** unicell1 has joined #openstack-infra06:54
megan_garciais there any workaround to it? thing is i have to use awk in my shell script ...but JJB wont let me create such job due to its innate {key} mapping thing ..06:54
megan_garciafor instance my shell builder script has this line ::              image=$(sudo docker images | awk '/$BUILD_ID/ {print $3}')06:55
megan_garcianow JJB fails because it thinks {print$3} is mapping to some key..06:55
*** unicell has quit IRC06:56
*** hdd has quit IRC06:58
*** abhisheks has quit IRC06:59
*** che-arne has joined #openstack-infra07:00
*** cnesa has quit IRC07:01
*** ddieterly has joined #openstack-infra07:01
crinklemegan_garcia: I think you need to use extra braces to escape them07:02
crinkleas in {{print $3}}07:02
crinklenot sure though07:02
megan_garciabut wouldnt that effect awk output?07:03
*** belmoreira has joined #openstack-infra07:03
*** camunoz has quit IRC07:03
megan_garciacrinkle: i have another job that uses awk too..it runs fine..because im not passing any key mappings to that job..07:03
*** Mithrandir has quit IRC07:04
megan_garciathe issue only happens if i hav a job to which some keys are mapped and must be passed..for that job , JJB takes awk {} as some key mapping ..07:04
crinklemegan_garcia: I'm reading from http://ci.openstack.org/jenkins-job-builder/configuration.html#job-template , potentially someone else here can clarify07:05
*** ddieterly has quit IRC07:06
*** Mithrandir has joined #openstack-infra07:07
*** jcoufal has joined #openstack-infra07:07
*** cnesa has joined #openstack-infra07:07
*** katyafervent is now known as katyafervent_awa07:08
megan_garciacrinkle: thanks...well i actually found a tempraory workaround to that...i created a separate builder macro for shell scripts that uses awk...and then inject those builder macros in my job that has key mapping :)07:10
megan_garciaits working now..07:10
*** scheuran has joined #openstack-infra07:10
megan_garciacrinkle: thanks for the late night help..lol..im glad someone is awake in this irc to help :> thanks again D07:10
crinklenp07:11
*** sarob has joined #openstack-infra07:13
*** megan_garcia has quit IRC07:15
*** Sukhdev has quit IRC07:15
*** katyafervent_awa is now known as katyafervent07:17
*** sarob has quit IRC07:18
*** vasu has joined #openstack-infra07:18
vasudo anyone know how to use sos-ci for creating OpenStack CI environment??07:19
*** ildikov has quit IRC07:21
*** HeOS has quit IRC07:24
jesusaurusvasu: i dont know what sos-ci is, but the openstack ci enviroment is configured using the puppet modules you can find at http://git.openstack.org/cgit/openstack-infra/07:26
*** yfried_ has joined #openstack-infra07:26
jesusaurushttp://git.openstack.org/cgit/openstack-infra/system-config is the main puppet entry-point and http://git.openstack.org/cgit/openstack-infra/project-config has a bunch of configuration files07:27
vasujesusaurus : using these takes a lot of time to install, also there is no surity that it works fine07:28
*** cnesa has quit IRC07:30
*** ghostpl_ has joined #openstack-infra07:31
*** yfried__ has joined #openstack-infra07:31
*** ghostpl_ has quit IRC07:31
*** yfried_ has quit IRC07:31
*** ghostpl_ has joined #openstack-infra07:31
vasujesusaurus : other than that is there any way that I can deploy a Ci environment easily??07:32
jesusauruswell, you know that it works because its being used to run all of openstack's tests07:32
*** otter768 has joined #openstack-infra07:33
*** armax has quit IRC07:33
*** tnovacik_ has joined #openstack-infra07:33
*** vasu has quit IRC07:36
*** otter768 has quit IRC07:37
*** vasu has joined #openstack-infra07:37
*** chlong_ has quit IRC07:42
*** chlong has quit IRC07:43
*** markvoelker has joined #openstack-infra07:44
*** cnesa has joined #openstack-infra07:44
*** vasu has quit IRC07:46
*** markvoelker has quit IRC07:50
*** afazekas_ has joined #openstack-infra07:50
openstackgerritTrinath Somanchi proposed openstack-infra/project-config: Adding project 'networking-freescale' to Stackforge.  https://review.openstack.org/15936207:54
*** trinaths has joined #openstack-infra07:54
*** ashleighfarnham has joined #openstack-infra07:58
*** ddieterly has joined #openstack-infra08:02
*** dims has joined #openstack-infra08:03
*** e0ne has joined #openstack-infra08:03
*** ashleighfarnham has quit IRC08:03
*** mpaolino has joined #openstack-infra08:05
*** ddieterly has quit IRC08:06
*** sergsh has joined #openstack-infra08:07
*** dims has quit IRC08:07
*** devvesa has joined #openstack-infra08:08
*** ildikov has joined #openstack-infra08:09
*** achanda has quit IRC08:09
*** woodster_ has quit IRC08:10
*** doude has joined #openstack-infra08:11
*** ihrachyshka has joined #openstack-infra08:12
*** matrohon has joined #openstack-infra08:14
*** mpavone has joined #openstack-infra08:20
*** dtantsur|afk is now known as dtantsur08:26
*** jcoufal has quit IRC08:28
*** jcoufal_ has joined #openstack-infra08:28
*** ihrachyshka has quit IRC08:36
*** salv-orlando has joined #openstack-infra08:39
*** armax has joined #openstack-infra08:40
*** HeOS has joined #openstack-infra08:43
*** mrunge is now known as mrunge_bbl08:43
*** andreykurilin_ has joined #openstack-infra08:49
*** skolekonov has joined #openstack-infra08:49
*** Ryan_Lane has quit IRC08:49
*** Ala has joined #openstack-infra08:49
*** jcoufal_ has quit IRC08:50
*** MaxV has joined #openstack-infra08:51
*** yamahata has quit IRC08:53
*** jistr has joined #openstack-infra08:53
*** stevemar has quit IRC08:57
*** SpamapS has quit IRC09:04
*** SpamapS has joined #openstack-infra09:05
*** SpamapS has quit IRC09:10
*** jlibosva has joined #openstack-infra09:10
*** teran__ has quit IRC09:14
*** sarob has joined #openstack-infra09:14
*** jgrimm is now known as zz_jgrimm09:15
*** derekh has joined #openstack-infra09:17
openstackgerritMichal Jastrzebski (inc0) proposed openstack/requirements: Oslo.versionedobjects requirement  https://review.openstack.org/15938009:18
*** abhishek_cb has joined #openstack-infra09:19
*** sarob has quit IRC09:19
*** abhishek_cb has quit IRC09:23
*** devvesa has quit IRC09:24
*** _shaps_ has joined #openstack-infra09:26
*** mpaolino has quit IRC09:27
*** teran has joined #openstack-infra09:28
*** zz_johnthetubagu is now known as johnthetubaguy09:28
*** bswartz has quit IRC09:30
*** jgallard has joined #openstack-infra09:30
*** jedimike has joined #openstack-infra09:32
*** SpamapS has joined #openstack-infra09:32
*** SpamapS has quit IRC09:32
*** SpamapS has joined #openstack-infra09:32
*** andreykurilin_ has quit IRC09:32
*** vdo has joined #openstack-infra09:32
*** otter768 has joined #openstack-infra09:33
*** markus_z has joined #openstack-infra09:36
*** jp_at_hp has joined #openstack-infra09:37
*** otter768 has quit IRC09:38
*** mpaolino has joined #openstack-infra09:41
*** BobBall_AWOL is now known as BobBall09:44
*** dizquierdo has joined #openstack-infra09:45
*** markvoelker has joined #openstack-infra09:47
*** teran has quit IRC09:48
*** yamamoto_ has quit IRC09:49
*** salv-orlando has quit IRC09:49
*** armax has quit IRC09:51
*** markvoelker has quit IRC09:52
*** markus_z has quit IRC09:53
*** hichihara has quit IRC09:57
*** cody-somerville has joined #openstack-infra09:57
*** achanda has joined #openstack-infra10:01
*** ddieterly has joined #openstack-infra10:04
*** armax has joined #openstack-infra10:06
*** sergsh has quit IRC10:07
openstackgerritLukas Bednar proposed openstack-infra/jenkins-job-builder: publishers: groovy-postbuild changes between 1.x & 2.x  https://review.openstack.org/13427610:07
*** ddieterly has quit IRC10:08
*** ssam2 has joined #openstack-infra10:10
*** EmilienM|afk is now known as EmilienM10:11
*** mrmartin has joined #openstack-infra10:12
*** koolhead17 has joined #openstack-infra10:13
*** sarob has joined #openstack-infra10:16
*** achanda has quit IRC10:17
*** sergsh has joined #openstack-infra10:19
openstackgerritDaniel Korn proposed openstack-infra/project-config: Use Neutron in Horizon integration tests job  https://review.openstack.org/15939710:19
*** teran has joined #openstack-infra10:21
*** sarob has quit IRC10:21
*** pelix has joined #openstack-infra10:21
*** mpaolino has quit IRC10:23
*** teran_ has joined #openstack-infra10:24
*** teran has quit IRC10:25
*** yamamoto has joined #openstack-infra10:26
*** tnovacik_ has quit IRC10:27
openstackgerritMerged openstack-infra/system-config: Additions to the Third-party CI doc "Creating a Service Account" section  https://review.openstack.org/15896910:36
*** mpaolino has joined #openstack-infra10:37
*** _nadya_ has joined #openstack-infra10:37
*** _nadya_ has joined #openstack-infra10:38
*** mrunge_bbl is now known as mrunge10:40
*** ihrachyshka has joined #openstack-infra10:41
*** SergK has joined #openstack-infra10:42
openstackgerritAleksey Ripinen proposed openstack-infra/infra-specs: Foreign keys support in StoryBoard  https://review.openstack.org/15940610:44
*** trinaths has left #openstack-infra10:45
openstackgerritAleksey Ripinen proposed openstack-infra/infra-specs: Foreign keys support in StoryBoard  https://review.openstack.org/15940610:46
openstackgerritAleksey Ripinen proposed openstack-infra/infra-specs: Foreign keys support in StoryBoard  https://review.openstack.org/15940610:48
*** eharney has quit IRC10:49
*** sergsh has quit IRC10:50
*** achanda has joined #openstack-infra10:50
*** dims has joined #openstack-infra10:52
openstackgerritAleksey Ripinen proposed openstack-infra/storyboard: Added branches to storyboard  https://review.openstack.org/15044710:55
*** dims_ has joined #openstack-infra10:55
openstackgerritNikita Konovalov proposed openstack-infra/storyboard: Added milestones  https://review.openstack.org/15250010:56
*** dims has quit IRC10:58
*** eharney has joined #openstack-infra11:02
deepakcsanteaya, clarkb ping11:02
openstackgerrityolanda.robla proposed openstack-infra/storyboard: Add tools to run integration tests  https://review.openstack.org/15597511:03
deepakcsclarkb, 104.239.136.99 and 15.126.235.20  are no longer needed, please delete them. Thanks11:03
*** sergsh has joined #openstack-infra11:04
*** ddieterly has joined #openstack-infra11:04
*** ssam2 has quit IRC11:06
*** asselin_ has joined #openstack-infra11:06
*** scheuran has quit IRC11:08
openstackgerritMerged openstack-infra/infra-manual: Fix two typos  https://review.openstack.org/15675311:09
*** ddieterly has quit IRC11:09
openstackgerritMerged openstack-infra/project-config: o.vo does not need old release testing  https://review.openstack.org/15887811:10
*** asselin_ has quit IRC11:11
*** amotoki has joined #openstack-infra11:12
*** rlandy has joined #openstack-infra11:13
*** tnovacik has joined #openstack-infra11:13
*** dtantsur is now known as dtantsur|bbl11:14
*** yamamoto has quit IRC11:15
*** salv-orlando has joined #openstack-infra11:15
*** achanda has quit IRC11:17
*** sarob has joined #openstack-infra11:17
*** yamamoto has joined #openstack-infra11:18
*** pcaruana has joined #openstack-infra11:19
*** ssam2 has joined #openstack-infra11:20
*** SergK has quit IRC11:21
*** sarob has quit IRC11:22
*** tnovacik has quit IRC11:24
*** tnovacik has joined #openstack-infra11:24
*** yfried__ is now known as yfried|afk11:30
*** jgallard has quit IRC11:31
ihrachyshkasdague, hey11:32
ihrachyshkasdague, there is something I don't know how to achieve for stable branches in project-config11:32
ihrachyshkasdague, we need to run pypy tests for stable/icehouse on an image with pypy installed (obviously), but it runs on base-precise which afaik misses it11:33
*** otter768 has joined #openstack-infra11:34
sdaguefor most projects, icehouse always runs on precise11:34
sdaguewhat is the project?11:35
*** ldnunes has joined #openstack-infra11:35
*** yfried|afk is now known as yfried__11:36
*** ashleighfarnham has joined #openstack-infra11:37
*** otter768 has quit IRC11:39
*** _nadya_ has quit IRC11:39
*** ashleighfarnham has quit IRC11:41
*** salv-orlando has quit IRC11:44
*** pc_m has joined #openstack-infra11:45
deepakcsjhesketh, https://review.openstack.org/159217 - can u give +A too ?11:45
*** dguerri is now known as dguerri`afk11:47
openstackgerritIhar Hrachyshka proposed openstack-infra/project-config: Don't run gate-trove-functional-dsvm-mysql for Icehouse  https://review.openstack.org/15942311:48
*** tosky has joined #openstack-infra11:49
*** markus_z has joined #openstack-infra11:49
toskyhi all, the recent update of firefox 36 broke selenium (https://code.google.com/p/selenium/issues/detail?id=8399) and the horizon integration tests11:50
toskysee for example: http://logs.openstack.org/60/157760/4/check/gate-horizon-dsvm-integration/fd3c268/console.html11:50
toskywhere should I open a ticket/bug? StoryBoard?11:50
toskythis is going to affect many reviews11:51
*** chlong has joined #openstack-infra11:51
*** chlong_ has joined #openstack-infra11:51
*** cdent has joined #openstack-infra11:54
openstackgerritMarton Kiss proposed openstack-infra/puppet-drupal: Get cron key variable from conf parameter  https://review.openstack.org/15942611:56
*** _nadya_ has joined #openstack-infra11:58
*** deepakcs has quit IRC11:59
ihrachyshkasdague, zaqar, they have pypy tox target12:00
sdagueI don't think that's really runable on the icehouse branch12:03
sdaguedo they really need that on their icehouse?12:04
*** mpaolino has quit IRC12:04
*** SergK has joined #openstack-infra12:04
*** aysyd has joined #openstack-infra12:04
flaper87sdague: we don't12:05
flaper87:)12:05
flaper87we can remove it, although it used to work12:05
*** ddieterly has joined #openstack-infra12:05
sdagueflaper87: are you maintaining the icehouse branch actively?12:05
sdaguethere is no real requirement on zaqar to maintain that, as it wasn't integrated at that time12:06
flaper87right, we've been maintaining it "passively" (whatever that means). As in, I've been tracking bot sycns (except for the one failing now) but no patches have been backported.12:07
sdagueyeh, that seems like more trouble than it's worth12:07
*** ddieterly has quit IRC12:09
*** erlon has joined #openstack-infra12:10
*** kgiusti has left #openstack-infra12:11
*** adalbas has joined #openstack-infra12:13
*** ssam2 has quit IRC12:14
*** dims_ has quit IRC12:14
*** dims has joined #openstack-infra12:14
openstackgerritNikolay Fedotov proposed openstack-infra/project-config: Functional test job for Kolla  https://review.openstack.org/15872612:17
*** SergK has quit IRC12:17
*** mwagner_lap has quit IRC12:22
*** ssam2 has joined #openstack-infra12:26
*** cody-somerville has quit IRC12:28
*** baoli has joined #openstack-infra12:30
*** dtantsur|bbl is now known as dtantsur12:31
openstackgerritLukas Bednar proposed openstack-infra/jenkins-job-builder: publishers: groovy-postbuild changes between 1.x & 2.x  https://review.openstack.org/13427612:32
*** mtanino has joined #openstack-infra12:33
*** annegentle has quit IRC12:33
*** rfolco has joined #openstack-infra12:35
*** ssam2_ has joined #openstack-infra12:35
*** dimtruck is now known as zz_dimtruck12:39
*** baoli_ has joined #openstack-infra12:41
*** tnovacik has quit IRC12:42
*** baoli has quit IRC12:44
flaper87sdague: hey, since I have (had?) you here, could you take a look at https://review.openstack.org/#/c/156982/ it's blocking one of our blueprints, which is basically ready to land12:49
flaper87jd__: ^12:49
*** markvoelker has joined #openstack-infra12:50
sdagueflaper87: min version?12:51
*** ZZelle has quit IRC12:52
*** ZZelle has joined #openstack-infra12:52
*** salv-orlando has joined #openstack-infra12:54
*** armax has quit IRC12:55
*** scheuran has joined #openstack-infra12:56
*** markvoelker has quit IRC12:56
*** rfolco has quit IRC12:56
flaper87sdague: lets add one, 2s. :)12:57
*** yfried__ has quit IRC12:59
*** yfried__ has joined #openstack-infra12:59
*** weshay has joined #openstack-infra13:00
AJaegerautobahn sounds cool ;)13:00
openstackgerritMichal Jastrzebski (inc0) proposed openstack/requirements: Oslo.versionedobjects requirement  https://review.openstack.org/15938013:01
*** tosky_ has joined #openstack-infra13:02
*** _nadya_ has quit IRC13:04
*** tosky has quit IRC13:04
*** jkraj has joined #openstack-infra13:05
*** afazekas_ has quit IRC13:05
*** ddieterly has joined #openstack-infra13:06
*** woodster_ has joined #openstack-infra13:07
*** esker has joined #openstack-infra13:07
flaper87sdague: could you point me to the docs (or examples) to create a devstack plugin?13:08
flaper87I'll start working on zaqar's13:08
*** BharatK has quit IRC13:09
*** zz_dimtruck is now known as dimtruck13:09
*** yfried__ has quit IRC13:09
*** mwagner_lap has joined #openstack-infra13:10
*** ddieterly has quit IRC13:11
*** kgiusti has joined #openstack-infra13:11
*** mbacchi has joined #openstack-infra13:13
openstackgerritVictoria Martínez de la Cruz proposed openstack/requirements: Adds Autobahn to global requirements  https://review.openstack.org/15698213:14
*** sputnik13 has quit IRC13:17
*** jgallard has joined #openstack-infra13:18
sdagueflaper87: there is an ML post on it13:19
sdaguewhich I think I linked to some examples13:19
*** tsg has quit IRC13:19
*** yfried__ has joined #openstack-infra13:19
*** dkranz has joined #openstack-infra13:19
*** cody-somerville has joined #openstack-infra13:19
*** chlong has quit IRC13:20
flaper87sdague: yeah, I was playing my lazyweb card but it didn't work13:20
*** chlong_ has quit IRC13:20
flaper87:)13:20
* flaper87 dives into os-dev13:20
*** sdake_ has joined #openstack-infra13:21
*** dprince has joined #openstack-infra13:23
*** sabeen has joined #openstack-infra13:27
openstackgerritMerged openstack-infra/shade: Don't try to add an IP if there is one  https://review.openstack.org/15608813:30
openstackgerritMerged openstack-infra/shade: Return extra information for debugging on failures  https://review.openstack.org/15641113:31
*** cody-somerville has quit IRC13:31
openstackgerritTimothy R. Chavez proposed openstack-infra/project-config: Print instance UUID to log  https://review.openstack.org/15625213:33
ihrachyshkasdague, https://review.openstack.org/15942313:34
*** dimtruck is now known as zz_dimtruck13:34
*** bcourt has quit IRC13:34
*** bcourt has joined #openstack-infra13:35
*** otter768 has joined #openstack-infra13:35
*** zz_dimtruck is now known as dimtruck13:36
*** ildikov has quit IRC13:38
*** yfried__ has quit IRC13:39
*** otter768 has quit IRC13:40
*** cody-somerville has joined #openstack-infra13:43
*** penguinRaider has quit IRC13:44
*** mtanino has quit IRC13:44
mkj_wkhmm, test failures in jenkins that don't reproduce locally. My favorite...13:45
*** signed8bit has joined #openstack-infra13:45
*** penguinRaider has joined #openstack-infra13:46
*** armax has joined #openstack-infra13:49
*** bswartz has joined #openstack-infra13:50
*** ildikov has joined #openstack-infra13:51
*** ildikov is now known as ildikov_afk13:51
*** ddieterly has joined #openstack-infra13:51
*** markvoelker has joined #openstack-infra13:53
*** BharatK has joined #openstack-infra13:54
mkj_wkzaitcev dinged me on Closes-Bug in review...13:55
mkj_wkBut his 'Maybe "Related-Story:" or something...' is hard to act on13:56
*** mjturek1 has quit IRC13:57
mkj_wkso a concrete statement of what it *should* say would be helpful. Is there yet a standard for referencing storyboard?13:57
*** ddieterly has quit IRC13:58
*** ddieterly has joined #openstack-infra13:58
openstackgerritNikita Konovalov proposed openstack-infra/python-storyboardclient: Added tags controller  https://review.openstack.org/15785713:59
*** mjturek1 has joined #openstack-infra13:59
*** markvoelker has quit IRC13:59
*** sigmavirus24_awa is now known as sigmavirus2413:59
*** eharney has quit IRC14:00
*** tsg has joined #openstack-infra14:00
*** tosky_ is now known as tosky14:01
*** bswartz has quit IRC14:02
*** annegentle has joined #openstack-infra14:02
openstackgerritMonty Taylor proposed openstack-infra/shade: Make image processing work for v2  https://review.openstack.org/15624714:02
*** flip214 has left #openstack-infra14:02
mordredI love it when gerrit can't merge a patch14:03
*** asselin_ has joined #openstack-infra14:03
*** viktors|afk is now known as viktors14:04
*** rfolco has joined #openstack-infra14:04
openstackgerritMerged openstack-infra/shade: Utilize dogpile.cache for caching  https://review.openstack.org/15890614:04
*** eharney has joined #openstack-infra14:06
*** eharney has quit IRC14:06
*** eharney has joined #openstack-infra14:06
*** dimtruck is now known as zz_dimtruck14:07
*** marun has joined #openstack-infra14:07
*** _nadya_ has joined #openstack-infra14:08
*** bhunter71 has joined #openstack-infra14:08
openstackgerritMerged openstack-infra/python-storyboardclient: Added subscriptions controller  https://review.openstack.org/15865914:09
openstackgerritMerged openstack-infra/python-storyboardclient: Added timeline controller  https://review.openstack.org/15865114:10
*** teran_ has quit IRC14:10
*** markvoelker has joined #openstack-infra14:10
*** marun has quit IRC14:12
*** doug-fish has joined #openstack-infra14:12
*** che-arne has quit IRC14:15
*** salv-orlando has quit IRC14:15
*** bknudson has quit IRC14:15
openstackgerritMonty Taylor proposed openstack-infra/shade: Configure dogpile.cache from clouds.yaml  https://review.openstack.org/15948014:16
*** changbl has quit IRC14:17
*** gordc has joined #openstack-infra14:17
*** bswartz has joined #openstack-infra14:17
*** asselin_ has quit IRC14:19
*** sarob has joined #openstack-infra14:20
*** sarob has quit IRC14:25
openstackgerritMichael Johnson proposed openstack-infra/git-review: Choose correct branch for rebase when submitting  https://review.openstack.org/15887714:25
*** tiswanso has joined #openstack-infra14:25
openstackgerritMerged openstack-infra/shade: Make image processing work for v2  https://review.openstack.org/15624714:26
*** cnesa has quit IRC14:27
*** yfried__ has joined #openstack-infra14:28
*** dkliban_afk is now known as dkliban14:30
*** afazekas_ has joined #openstack-infra14:31
openstackgerritMike Heald proposed openstack-infra/pypi-mirror: Mirror appropriate wheels in main mirror  https://review.openstack.org/15948514:32
*** yfried has joined #openstack-infra14:32
*** sabeen has quit IRC14:33
*** yfried__ has quit IRC14:33
*** radez_g0n3 is now known as radez14:34
*** rcarrillocruz has quit IRC14:35
*** bknudson has joined #openstack-infra14:36
SpamapSyolanda: \o/ we collaborated :-D14:37
SpamapSmordred: careful, there are collaborators on shade14:37
mordredSpamapS: I know. scary14:37
yolandait was nice to hear about dogpile, never heard about that before14:38
openstackgerritMike Heald proposed openstack-infra/pypi-mirror: Mirror appropriate wheels in main mirror  https://review.openstack.org/15948514:38
*** rcarrillocruz has joined #openstack-infra14:39
SpamapSyolanda: it's really quite good.14:39
SpamapSAnd we can add config to let people point their shade at a memcached now.14:39
*** xyang1 has joined #openstack-infra14:39
SpamapSlike a boss14:39
mordredSpamapS: check my latests shade patch, btw - and the corresponding os-client-config one14:40
mordredSpamapS: I _think_ that should allow for expressing cache config in clouds.yaml14:41
yolandathere are few places where custom caching is used, that could benefit from it14:41
SpamapShm did I just find a bug in gertty's search?14:41
SpamapSSyntax error at "os-client-config" in search string "os-client-config"14:41
SpamapSI can't find os-client-config :-P14:42
SpamapScorrection: I'm unwilling to focus hard enough to find it14:42
*** zz_jgrimm is now known as jgrimm14:43
*** deepakcs has joined #openstack-infra14:44
mordredSpamapS: stackforge/os-client-config14:44
mordredheh14:44
SpamapSlovely14:44
SpamapSseriously let's just rename everything unilaterally14:44
SpamapS;)14:44
mordredyolanda: yah - I was also thinking that if we plumb this all correctly, then the inventory module and the launching modules will work together very nicely14:44
mordredyolanda: especially if we're careful to do things like having creates and deletes push cache invalidation events14:45
mordredyolanda: that way, if you had all of your systems configured to use redis or memcached14:45
mordredyou'd get screaming-fast performance _and_ correct data all the time14:45
mordredbecause the only thing updating your cloud resources would be you14:45
yolandamm, yes, cache invalidation is a todo for shade14:45
mordredit'll take us a few cycles to get that fully right I'm sure :)14:45
yolandamm, nodepool has also custom caching for server, floating ips14:46
yolandathis could benefit from it14:46
SpamapS:-D14:47
*** sarob has joined #openstack-infra14:47
yolandalet's make everything use dogpile :)14:48
SpamapSmordred: I like the direction. Will hold off reviewing until you get your CI s*** together. ;)14:48
*** cnesa has joined #openstack-infra14:48
*** mriedem has joined #openstack-infra14:49
*** carl_baldwin has joined #openstack-infra14:49
mordredSpamapS: yes. I think I'd like to not land any additional feature patches for shade now until we have the testing story solid14:49
mordredbecause I'm about to tell ansible upstream that the first PR that uses this can land14:49
mordredwhich means, at this point, breaking it is very bad form14:49
*** jgrimm is now known as zz_jgrimm14:51
*** sdake_ has quit IRC14:52
*** sarob has quit IRC14:52
*** cnesa1 has joined #openstack-infra14:53
*** cnesa has quit IRC14:53
openstackgerritMichael Johnson proposed openstack-infra/git-review: Choose correct branch for rebase when submitting  https://review.openstack.org/15887714:58
*** wenlock has joined #openstack-infra14:59
*** bswartz has quit IRC14:59
*** prad has joined #openstack-infra15:00
*** bswartz has joined #openstack-infra15:00
* mkj_wk injects intentional failure into review hoping to get new information about why tests are failing in Jenkins in ways that don't repeat locally15:00
SpamapSmordred: I may be weird, but I enjoy writing tests like that.15:00
*** sabeen has joined #openstack-infra15:01
SpamapSmordred: that may stem from my firm belief that, done right, they enable higher velocity15:01
zxiiropelix: Can we go ahead with this patch? It's high priority for us so I'm hoping to get it merged soon. I pushed up a separate patch already for the consolidation work. https://review.openstack.org/155865/15:02
*** zz_dimtruck is now known as dimtruck15:03
*** sabeen2 has joined #openstack-infra15:03
*** signed8b_ has joined #openstack-infra15:03
*** sabeen has quit IRC15:05
mordredSpamapS: I totally agree, actually - for me the main thing is that writing them is context switch and I keep not having quite enough brainspace to context switch in15:05
SpamapSmordred: yeah it is.15:06
*** signed8bit has quit IRC15:06
SpamapSmordred: I'm going to try and reorient behind rcarrillocruz and his functional tests since I think that will actually be more important.15:06
mordredSpamapS: ++15:07
SpamapSI mean, when I'm not making puppet things.. because I need to start making puppet things. :)15:07
mordredwell, honestly, this shade + nodepool stuff is part of getting nodepool-dib finished15:07
SpamapSalso I think a tiny bit of effort on dib testing is in order.15:07
*** nelsnelson has joined #openstack-infra15:07
rcarrillocruzSpamapS: yeah, so i talked to clarkb and mtreinish, i really should base shade func tests the same way python-novaclient and other clients do15:07
mordredwhich is a priority effort aroudn here15:07
rcarrillocruzi'm looking into that, i'm hoping to have it something functional this weekend15:07
mordredI'd also like to see a patch to devstack that outputs a clouds.yaml so that we can delete the "read tempest config" parts out of rcarrillocruz's patch15:08
mordredbut I haven't gotten to that yet15:08
rcarrillocruzmordred: i'm getting rid of that, will instead load up openrc in a post hook funciton and read creds with os.environ on tests15:09
rcarrillocruzthat read tempest config was FUGLY15:09
mordredawesome15:09
*** e0ne is now known as e0ne_15:09
mordredthat will actually test that the OS_ config stuff for shade works too15:09
mordred:)15:09
*** dmsimard_away is now known as dmsimard15:10
mordredSpamapS: so - another task that someone who has good dib foo might want to tackle15:10
yolandaricardocarrillocruz, you finally got the tests working?15:11
*** mattfarina has joined #openstack-infra15:11
*** yamahata has joined #openstack-infra15:11
mordredSpamapS: is a dib element that installs cloud-init into a venv - but still installs all of the systemd/upstart/sysvinint stuff properly15:11
mordred(which is to say, the normal files, probably some symlins, etc)15:11
mordredbut ...15:11
openstackgerritJohn Griffith proposed openstack-infra/elastic-recheck: Add query for cinder bug 1425310  https://review.openstack.org/15949515:11
openstackbug 1425310 in Cinder "tgtadm: this access control rule does not exist - results in timeout waiting for volume to attach (stuck in detach) or detach" [High,Confirmed] https://launchpad.net/bugs/142531015:11
mordredSpamapS: to make it more fun - it needs to be cloud-init with these patches applied: https://github.com/racker/cloud-init-debian-pkg/tree/master/debian/patches15:11
*** BharatK has quit IRC15:11
*** sputnik13 has joined #openstack-infra15:12
mordredSpamapS: and it also might be friendly to JayF if someone wanted to take on getting those upstreamed into cloud-init and getting scott to cut another release to make our element easier15:12
mordredNg: you're still around for a little while before paternity, yeah? does that ^^ sound interesting and achievable before break?15:13
*** salv-orlando has joined #openstack-infra15:14
Ngmordred: I am. The upstreaming bits? I'd be happy to talk to ScottyMo about that :)15:15
*** ghostpl_ has quit IRC15:15
*** annegentle has quit IRC15:16
*** acfleury has joined #openstack-infra15:16
*** ashleighfarnham has joined #openstack-infra15:16
openstackgerritMichael Johnson proposed openstack-infra/git-review: Choose correct branch for rebase when submitting  https://review.openstack.org/15887715:17
*** adalbas has quit IRC15:17
mordredNg: woot. I'm not sure what JayF has done in terms of submitting those patches yet15:19
*** e0ne_ has quit IRC15:19
mordredNg: but the three patches in their debian packaging branch above are needed to read the networking info in the vendor_data in configdrive15:20
Ngmordred: I was just checking into that. 1338614, 1316475 and 1303986 are all listed as Fix Released for cloud-init 0.7.615:20
mordredNg: fwiw, they're running those modified cloud-init's in production for their onmetal service15:20
rcarrillocruzso folks15:20
rcarrillocruzjust so i understand this new hp internal cloud15:20
mordredcool - so then it may just be the configdrive-onmetal patch that needs upstream15:20
rcarrillocruzyou plan to put ironic powers to nodepool15:20
rcarrillocruz?15:20
mordredrcarrillocruz: that may or may not be a topic for this channel?15:20
Ngmordred: yeah, just checking over that to see if it's been submitted anywhere15:21
mordredrcarrillocruz: oh - you mean the new infra cloud15:21
AJaegersdague, mordred I have one patch up to start less jobs for documentation jobs, could you review https://review.openstack.org/#/c/158406 this week, please?15:21
rcarrillocruzre: jeblair email15:21
mordredno - no ironic in nodepool15:21
mordredNg: woot15:21
*** markvoelker has quit IRC15:21
jrollNg: mordred: we're waiting for the nova changes to land before submitting the cloud-init patches... which has been booted to L15:21
*** ashleighfarnham has quit IRC15:21
Ngjroll: aha15:21
mordredrcarrillocruz: we'll have an ironic cloud that our manual "launch_node" process can use15:21
*** markvoelker has joined #openstack-infra15:21
mordredjroll: any specific reason to not submit them anyway? I mean, that patch is now live in the wild, which means that cloud-init should be able to figure out how to read it, no?15:22
rcarrillocruzah, ok...15:22
*** matrohon has quit IRC15:22
jrollmordred: because vendor-data vs metadata15:22
mordredrcarrillocruz: and we'll use that to spin up servers like we do othe rthings like gerrit and whatnot15:22
mordredjroll: hrm15:22
jrollmordred: we want nova to standardize on the location15:22
jrollotherwise we'll just need to fix it in cloud-init again15:22
jrolland do backwards compat and stuff15:22
mordred*headdesk*15:22
jrollya15:23
jrolland it wasn't good enough for FFE15:23
jrollso ya15:23
jrollgood times15:23
mordredjroll: FFE?15:23
jrollfeature freeze exception15:23
mordredjroll: it's been up and rolling for months?15:23
mordredI mean, I know what an FFE is15:23
*** sergsh has quit IRC15:23
jrollmordred: in our cloud, patch is still in flight to nova15:23
mordredI'm just saying- config-drive is epically and completely useless without this15:23
Ngwe're not in feature freeze yet15:23
jrollnova is15:23
mordredNg: nova is15:23
Ngoh, weeeelll, if we're going to bring facts into this15:24
jrollhttps://review.openstack.org/#/c/153097/15:24
jrollthe original patch has only been up since Jun 25, 2014 https://review.openstack.org/#/c/102649/2015:25
Ngjroll: any objections if I at least open a bug against cloud-init that references that change and the extant cloud-init patch?15:25
jrollNg: I don't care, but I'd rather a cloud-init release didn't rely on downstream patches :/15:25
*** markvoelker has quit IRC15:26
deepakcsclarkb, Can you bless https://review.openstack.org/158392 - thanks!15:26
jrollNg: and I'd rather not carry that patch downstream if this ever goes upstream15:26
deepakcsclarkb, wait, thats wrong one15:26
deepakcsclarkb, https://review.openstack.org/#/c/159217/ - this is the one15:26
*** jgallard has quit IRC15:28
Ngjroll: hrm, ok, I'm happy to wait until the nova situation clears up before we start bothering Scott about landing things in cloud-init15:29
jrollNg: yeah, JayF and moser talked about this already15:29
jrollNg: but that means waiting for liberty and going back through the spec process and whatnot15:30
mordredcarrying it as a local patch to cloud-init in our element shouldn't be terible15:30
mordredjroll: everything about what you're talkinga bout just makes me sad15:30
jrollmordred: you have no idea.15:30
Ngugh15:31
jrollmordred: are you using our cloud-init builder? :)15:31
mordredjroll: no15:31
mordredjroll: we do not want to install it from packages15:31
mordredvery explicitly15:31
mordredjroll: beecause our nodes are test nodes for openstack, and cloud-init depends on the phonebook15:31
*** sarob has joined #openstack-infra15:32
mordredso installing it pollutes the node in question15:32
jrollmordred: ah, I see15:32
jrollya15:32
mordredjroll: so - since you're putting things in vendor data, but eventually they're going to be in network_data.json15:33
openstackgerritDoug Hellmann proposed openstack-infra/project-config: Add requirements jobs to heat  https://review.openstack.org/15950815:33
jrollmordred: yeah15:33
mordredjroll: this means that infra will ultimate need a version of cloud-init that can handle both, right?15:33
mordredNg: ^^15:33
mordredNg: I think I've just dubbed you cloud-init king :)15:34
jrollmordred: I'm thinking we could carry a patch for a bit to provide both15:34
jrollbut if you want to be the one to handle that more power to you :)15:34
Ngmordred: I'm going to immediately set about having more babies so I can be on paternity forever15:34
mordredNg: good call15:34
jrolllol15:34
rcarrillocruz  hehe15:34
mordredjroll: I mean, we don't really need to worry about this until december, right?15:34
mordredjroll: or will rackspace roll out the change once the nova patch lands?15:35
*** jerryz has quit IRC15:35
*** zz_jgrimm is now known as jgrimm15:35
jrollmordred: depends how good we are at CD when that lands :)15:35
mordredawesome15:36
mordredso - the _safest_ thing for infra to do, in all seriousness (no offense intended) is to modify our cloud-init patch to look in both places15:36
*** otter768 has joined #openstack-infra15:36
mordredso that whenever rax decides to roll out the official patch, our stuff still works15:36
mordredand then the only change we have to make in the future is to stop patching once it's all upstream15:37
jrollmordred: agree, though even if y'all are ready, I'd like to give that some deprecation time15:37
jrollin case someone else decided to use it15:37
mordredoh totoally15:37
mordredI think rax should definitely do double-duty with deprecation15:37
mordredand if you guys deprecate, and we handle both15:37
mordredthen we _should_ have a seamless experience15:38
jrollyou'll be good to go, yeah15:38
jrollmordred: mind pointing us at that cloud-init patch that handles both when it's done? :)15:38
jrollso onmetal can be ready too :P15:38
mordredjroll: absolutely!15:38
jrollawesome15:38
jrollJayF: ^^ you'll enjoy this discussion, for some definition of enjoy15:38
*** e0ne has joined #openstack-infra15:39
mordredNg: so - for context, in case you don't have it - one of the goals in dib-nodepool is for all of our nodepool nodes to be built with dib and be uploaded to glance15:39
mordredNg: of course, our clouds are different - one uses dhcp for network config, one needs either a xen agent or a modified version of cloud-init to get network info15:40
mordredNg: additionally, we don't _Really_ want cloud-init installed, because of all of the depends and the way they step on test payloads dependency chains15:40
mordredNg: so I've got this: https://review.openstack.org/#/c/140840/15:40
openstackgerritMerged openstack-infra/project-config: Run tempest-dsvm-full-glusterfs job on ubuntu  https://review.openstack.org/15921715:40
*** otter768 has quit IRC15:41
mordredNg: which was the addition of some elements before I gave in to using cloud-init in a venv15:41
mordredNg: oh - wait - that's the wrong patch - one sec15:41
mordredNg: this one: https://review.openstack.org/#/c/154132/15:41
mordredread-vendor-json is the attempt to not use cloud-init15:42
mordredbut really should be replaced with an element that patches and installs cloud-init into a venv, installs systemd units and upstart whatnots and whatnot15:42
mordredNg: make sense?15:42
Ngmordred: reading the changeset atm15:43
mordredNg: most of that patch will make you very very sad15:43
*** kirshil has joined #openstack-infra15:43
*** sarob has quit IRC15:44
HenryGHi folks. I heard it is possible, using some devstack variables, to get the gate to test an unmerged tempest change with an unmerged project (neutron in my case) change. Can you point me to instructions or an example?15:44
openstackgerritMichael Johnson proposed openstack-infra/git-review: Choose correct branch for rebase when submitting  https://review.openstack.org/15887715:45
mordredHenryG: you just need to mark one of them as "Depends-On: $change-id" on the other15:45
*** skolekonov has quit IRC15:45
openstackgerritAndreas Jaeger proposed openstack-infra/project-config: Run specs site index generation only if needed  https://review.openstack.org/15951215:46
HenryGmordred: very cool, thanks! Let me try that15:46
AJaegerHenryG, documented at https://review.openstack.org/15885515:46
kirshilmordred: what if tempest change depends on few patchsets in neutron?15:47
AJaegerkirshil, see my document above and search for multiple changes15:47
openstackgerritJames E. Blair proposed openstack-infra/storyboard: Remove API paging from comments  https://review.openstack.org/15951515:47
HenryGkirshil: I believe I know what they mean, I'll help you out15:48
*** AJaeger has quit IRC15:48
*** reed has joined #openstack-infra15:48
openstackgerritJohn Griffith proposed openstack-infra/elastic-recheck: Add query for cinder bug 1425310  https://review.openstack.org/15949515:48
openstackbug 1425310 in Cinder "tgtadm: this access control rule does not exist - results in timeout waiting for volume to attach (stuck in detach) or detach" [High,Confirmed] https://launchpad.net/bugs/142531015:48
JayFjroll: for no definition of enjoy15:49
jrollJayF: at least there's a game plan :|15:49
*** MarkAtwood has joined #openstack-infra15:50
JayFjust like we had one for K15:50
jroll-.- yeah15:51
*** markvoelker has joined #openstack-infra15:51
pelixzxiiro: will try to get to it today, course you may want to ping zaro as well as it'll need two +2's for approval. If you find that you'd benefit from landing patches locally and automate their removal when syncing with upstream you may be interested in https://github.com/stackforge/git-upstream/15:52
* pelix tries to look innocent after shamelessly plugging a project he authored15:52
kirshilAjaeger: thk, indeed simple!15:52
mordredsdague: ^^ is there anything I can personally do to help convey the level of importance of the above spec/patch to infra?15:52
mordredsdague: I mean, I know that's not how nova makes decisions, but _perhaps_ if they knew that this was a showstopped-bad problem for infra that we're having to work around by doing evil, _maybe_ it would elevate the importance?15:53
mordredsdague: or should I just keep my fat mouth shut?15:53
*** che-arne has joined #openstack-infra15:53
JayFmordred: I put the fact Infra was using these patches in the FFE request, it seemed to make no difference.15:53
mordredJayF: awesome15:53
* jroll wishes for the millionth time that we used dhcp15:54
JayFmordred: it's okay, lets fix it with another set of 4 dozen emails about process to the mailing list /s15:54
mordredJayF: sweet! that sounds great15:55
mordredjroll: well, yes. to be fair, this _is_ only needed because of lack of dhcp15:56
jrollmordred: yeah, well aware15:56
mordredso - I mean, I actually think it's a crazy feature to need15:56
mordredBUT15:56
mordredthe real world being what it is15:56
jrollmordred: we decided not to use a rootkit^W nova-agent and so we've been through the same pain15:56
jeblairwhere can i see the history of this?15:57
sdaguemordred: sorry, which patch?15:57
mordredjeblair: https://etherpad.openstack.org/p/kilo-nova-ffe-requests15:57
JayFjeblair: in gerrit on the patches, the FFE was on the list ... we would've needed an FFE for the spec too because we couldn't get reviews, but someone landed it in the nova meeting talking about spec ffes15:57
mordredjeblair: search for "proxy neutron configuration"15:57
sdaguethere is much scrollback15:57
mordredsdague: morning sdague !15:57
sdaguemordred: can you point me directly at whatever you want me to look at15:57
mordredsdague: tl;dr - adding network info to config-drive did not make the kilo cutoff, this makes infra sad15:58
jeblairJayF: one of the confusing things for me is that this has been worked on since june, right?15:58
JayFit's been worked on since juno cycle15:58
jrolland months of feature freezes make me sad15:58
JayFhell we've been running it downstream since before juno15:58
jrolljeblair: initial patch was june 25, it's been in production at rackspace since july15:58
*** jlibosva has quit IRC15:59
openstackgerritMichael Johnson proposed openstack-infra/git-review: Choose correct branch for rebase when submitting  https://review.openstack.org/15887715:59
sdagueok, so can someone link actual patches implicated here?15:59
zxiiropelix: ok thanks15:59
jeblairsdague: http://lists.openstack.org/pipermail/openstack-dev/2015-February/056782.html  seems important15:59
sdaguethere are 650 outstanding patches for nova, so I don't have a map of them all in my head15:59
mordredsdague: https://review.openstack.org/#/c/153097/15:59
*** yamahata has quit IRC16:00
jeblairso patches for an approved kilo spec still need an ffe?16:00
*** dangers_away is now known as dangers16:00
mordredsdague: I believe that https://review.openstack.org/#/c/143755 and https://review.openstack.org/#/c/102649/ are also relevant16:01
*** ildikov_afk is now known as ildikov16:01
jeblairhttp://specs.openstack.org/openstack/nova-specs/specs/kilo/approved/metadata-service-network-info.html16:01
jrolljeblair: yeah, there's a spec freeze and then a code freeze later16:01
jrolljeblair: and ffe's for both16:02
mrmartinmorning16:02
jeblairi know where the release schedule is... where's the _nova_ release schedule?16:03
*** mikedillion has joined #openstack-infra16:03
*** jlibosva has joined #openstack-infra16:03
mrmartinmay I ask somebody to review / approve this mega-patch: https://review.openstack.org/#/c/159426/ I successfully closed out cron-scheduled tasks yesterday from groups and groups-dev portals.16:03
jeblairoh, it's notated on the regular release schedule: https://wiki.openstack.org/wiki/Kilo_Release_Schedule16:04
flaper87hey guys, any chance I can get access to the vm that run this job: http://logs.openstack.org/29/151529/6/check/check-tempest-dsvm-redis-zaqar/269ea01/ ?16:05
flaper87or run that job again and get access16:05
flaper87I've literally no idea what's going on there and it happens just with that patch16:05
flaper87and just for redis16:05
*** marun has joined #openstack-infra16:05
flaper87I tried it locally (in a vm, etc) and it seems to work16:05
flaper87so, no clue :(16:05
*** sdake_ has joined #openstack-infra16:06
jrolljeblair: yeah, it's laid out more clearly here as well http://lists.openstack.org/pipermail/openstack-dev/2015-February/057614.html16:06
clarkbflaper87 would need to rerun that job. the VM for a finished job is long deleted16:06
jeblairjroll, sdague: i guess i got confused by comments from john like "this spec is not approved for kilo".  but the spec _is_ approved for kilo.  i guess what he really ment to say was that the patches were not eligible to land because we've passed feature freeze and this is not a priority feature?16:06
clarkbmrmartin I will review as soon as I am awake16:07
mrmartinclarkb, it is enough to solve until tomorrow16:07
mrmartinso feel free to drink your morning cofee16:07
mrmartinthnx.16:07
sdaguejeblair: yeh, that's what he meant16:07
jrollsdague: yeah, I guess so16:07
*** sigmavirus24 is now known as Slackwarebot16:07
jroller, jeblair16:07
*** Slackwarebot is now known as sigmavirus2416:07
*** penguinRaider has quit IRC16:08
clarkbflaper87 if you kick the job off again point us to the jenkins page for it, can be found on zuul's status page16:08
*** penguinRaider has joined #openstack-infra16:08
*** changbl has joined #openstack-infra16:09
*** tosky_ has joined #openstack-infra16:10
*** mikedillion has quit IRC16:11
*** penguinRaider has quit IRC16:12
*** penguinRaider has joined #openstack-infra16:12
sdagueyeh, so, I think we're pretty late in the process, but i'm not a nova-driver, so poking at johnthetubaguy or mikal might be more appropriate. The patch wasn't resurrected until christmas. And was in a fail test state a number of times before the freeze consideration.16:13
sdaguethis is the first time I've seen the patch, but again, there are 650 outstanding nova patches, and I've only probably seen 200 in the last 30 days16:14
*** erikwilson has joined #openstack-infra16:14
*** erikmwilson is now known as Guest1779816:14
*** erikwilson is now known as erikmwilson16:14
*** erikwilson has joined #openstack-infra16:14
*** tosky has quit IRC16:14
*** tnovacik has joined #openstack-infra16:15
*** mrmartin has quit IRC16:15
*** annegentle has joined #openstack-infra16:16
JayFsdague: mordred: fwiw we did poke those people; ended up with a no anyway. I was already to the point of acceptance :/16:17
clarkbgah mrmartin just left... I left some comments on that change16:17
*** yamamoto has quit IRC16:18
flaper87clarkb: doing now, thanks16:19
sdagueyeh, don't know what to say. If I'd seen the patch before freeze I could have evaluated it. I didn't even realize such an effort was happening16:19
JayFsdague: I'll remember that; I think this time we're going to try to make it be the first damn patch to land in L16:20
JayFheh16:20
sdagueyeh, I'm happy to help with that.16:21
JayFsdague: also, feel free to use it on Rackspace Cloud (except it's in vendor_data.json['network'] downstream)16:21
*** klrmn has joined #openstack-infra16:21
*** annegentle has quit IRC16:22
klrmncan anyone remind me where the subunit2html filter lives?16:22
*** dprince has quit IRC16:22
clarkbklrmn: openstack-infra/project-config/jenkins/scripts iirc16:22
*** mrunge has quit IRC16:23
*** sdake__ has joined #openstack-infra16:23
*** gyee has joined #openstack-infra16:24
*** dannywilson has joined #openstack-infra16:25
sdaguealso, in the future, 'important to infra' should probably be replaces with a mordred rant of the terribleness of work around. It might have more emotional impact.16:25
*** sabeen2 has quit IRC16:26
*** spzala has joined #openstack-infra16:27
*** deepakcs has quit IRC16:27
jeblairsdague: indeed :)  also, when we get a cloud up and running, oh the rants there will be...16:27
*** spzala has quit IRC16:27
*** sdake_ has quit IRC16:28
JayFjeblair: part of me wants Ironic somewhere in that cloud so we can hear the ranty feedback too16:28
JayFjeblair: I'm always a fan of people finding broken stuff before our public cloud does :P16:29
jrollJayF: there will be, if I read correctly16:29
sdaguesure, though, again a big part of the problem is the 650 patches in the backlog. Because stuff gets lost in that pile all the time because the pile is so big.16:29
jrollI'm sure we'll get all the rants16:29
jrollcan't wait.16:29
jeblairJayF: i think the back of the etherpad plan atm is an all-in-one ironic cloud that supports the workload kvm cloud16:29
JayFjeblair: sweet. Where do I send the bribes to get you guys to use agent_* driver? :P16:30
*** tonytan4ever has joined #openstack-infra16:31
marcusvrnanteaya: clarkb: ping16:31
jeblairJayF: that reminds me, we still need to add bitcoin processing to zuul for patch prioritization...16:32
jeblairoh, forgot ;)16:32
sdaguemight I also recommend that the infra team proposes an 'infra top wants from nova' session at the next design summit. I think openning up that communication path more would be really handy.16:34
clarkbmarcusvrn: hello16:34
*** viktors is now known as viktors|afk16:34
*** ashleighfarnham has joined #openstack-infra16:34
jeblairsdague: yes, in all seriousness, opening up a feedback loop to a highly visible cloud op is something i'm really excited about.  that's a good idea (next summit will be early days still, but hopefully we'll have something to talk about by then).16:35
sdagueyou'll at least have this feature to talk about16:35
sdagueI honestly don't think anyone really understood the pain not having this has caused16:36
*** david-lyle_afk is now known as david-lyle16:36
sdagueand, while in an ideal world, a working patch up in gerrit gets evaluated on it's own merrits, we're at queue overflow in that model, so stuff needs to be advocated for more to get eyes.16:36
*** jlibosva has quit IRC16:37
marcusvrnclarkb: I switched my CI to "check" mode...and it's working well. I see that stackalytics has a page that show which drivers has or not CI (http://stackalytics.com/report/driverlog?vendor=Hitachi)16:37
ttxjeblair: so.. irc-meetings is getting close to being usable. I'd like to pick your brains on where you think we should publish (the ICS file and the human-readable list of meetings)16:37
ttxjeblair: original plan was to post to wiki16:37
jeblairi'll speak for mordred and say that he's always willing to provide rant-based patch prioritization :)16:37
ttxjeblair: but then I want the wiki to die16:37
anteayamarcusvrn: what is the question?16:38
jeblairttx: oh killing the wiki seems a tall order :)16:38
marcusvrnclarkb: how can I change the "CI exists" column?16:38
ttxa specific website sounds a bit too much though16:38
marcusvrnanteaya: ^16:38
johnthetubaguysdague: can I help at all? sorry only half busy?16:38
anteayamarcusvrn: change the jason file in the stackalytics repo16:38
ttxjeblair: correction: I want the wiki to die as a repository of reference information16:38
anteayamarcusvrn: submit a patch that changes the json file16:38
jeblairttx: ++16:38
*** amitgandhinz has joined #openstack-infra16:38
ttxjeblair: if it's to post a protected page with a autogenerated list of meetings, bit dumb16:39
*** ashleighfarnham has quit IRC16:39
ttxanyway, open to suggestions16:39
sdaguejohnthetubaguy: it's a nova FFE question, the infra team is having to do a lot of work to work around the lack of - http://lists.openstack.org/pipermail/openstack-dev/2015-February/056782.html16:40
*** jp_at_hp has quit IRC16:40
marcusvrnanteaya: just it?16:40
clarkbasselin: re 149032, trying to evaluate what we need to do to be copying jaypipes content and follow the terms of his license16:40
* asselin looking16:40
jeblairttx: so until we have better integration with the website, www.o.o/meetings is tricky.... which is why we have tarballs, ci, logs, docs-draft, security, governace, specs, summit all hosted on static.o.o.  we also have status.o.o.16:41
clarkbasselin: I see to recall it was a CC license of some sort but I am not finding specifics on his blog16:41
anteayamarcusvrn: yes, which is why I give the stackalytics driver page no credence whatsoever16:41
jeblairttx: the closest hostname that sounds reasonable is governance.o.o.  so the only things i can come up with off the top of my head are governance.o.o/meetings, or just go ahead and make meetings.o.o16:41
anteayasince changing the json file by the ci operator doesn't seem the best way to evaluate which cis exist16:41
mordredjeblair: I will always rant16:42
anteayamarcusvrn: but to the best of my knowledge that is how it is done16:42
jeblairmordred: whew, i got that right16:42
ttxjeblair: technically we could do static.o.o/meetings, I guess16:42
mordredttx: I also want the wiki to die as a repository of reference informatino16:42
jeblairttx: oh, i don't think we advertise static.o.o as a hostname for a url, it's just the server name that holds all those vhosts16:43
ttxok16:43
asselinclarkb, yea, I don't see anything there either.16:43
ttxstatus.o.o ?16:43
* ttx tries to get creative16:43
openstackgerritStefano Maffulli 'reed' proposed openstack-infra/system-config: Removing inactive admins from root access  https://review.openstack.org/15886216:43
*** amotoki has quit IRC16:43
jeblairttx: we could see if there's a way to integrate it with eavesdrop...16:43
ttxah.hm16:44
*** stevemar has joined #openstack-infra16:44
jeblairttx: as in, a single hostname for meeting schedules and logs16:44
ttxThat.. would make sense16:44
clarkbasselin: is the comment about copying still valid? if so I think we should at least run it by jaypipes (I don't expect any issues with it but ya)16:44
*** tsg has quit IRC16:44
ttxirc.openstack.org sounds like an IRC server though. meetings.o.o doesn't sound like where you would store channel logs... Argh, naming is hard16:45
mordredJayF: I believe we're planning on using the agent driver16:45
clarkbflaper87: https://jenkins03.openstack.org/job/check-tempest-dsvm-redis-zaqar/61/ is that the right job to hold?16:45
mordredJayF: basically, whatever it is that deva said was the "right" way to do it these days16:45
ttxjeblair: but then I have enough for the next step, which is I should probably publish RST rather than WIKIMOIN syntax16:45
mordredJayF: also - we'll be running that like we run everythign else aroudn here - so we'll highly welcome your input and expertise in reviewing those patches16:45
ttxor mediawiki16:45
asselinclarkb, I need to double check. I did rewrite quite a bit. He did give an intial -1 early on that did imply approval of this change.16:45
reedoh, not another domain name :)16:45
ttxreed: I tried not to :)16:45
jeblairttx: yeah, we've got several bad choices now!  we can either pick one, or maybe come up with a good while you're working on it :)16:46
ttxreed: arguably, we would remove eavesdrop.o.o16:46
ttxso we don't ADD one :)16:46
jeblairright16:46
flaper87clarkb: yes16:46
reedrotfl, math jujitsu :)16:46
*** ashleighfarnham has joined #openstack-infra16:47
* mordred also agrees on removing eavesdrop.o.o16:47
flaper87clarkb: I swear that just started, I've been staring at the freaking zuul for 15mins and switched windows for 2 secs16:47
ttxI can't wait to enable it though -- keeping the YAML in sync is as unfun as keeping the wiki page in sync16:47
flaper87:P16:47
reedfrankly, that sounds like two changes not 016:47
clarkbflaper87: np :) I have held that node, point me at an ssh public key and I can set that up for you16:47
*** koolhead17 has quit IRC16:48
reedwhat's wrong with publishing on the wiki? is that impossible or too hard?16:48
marcusvrnanteaya: ok! I will do it! Do you know what is the file that I have to change?16:48
ttxreed: it's wrong -- you would publish a protected page to a WIKI16:48
* reed ending up defending mediawiki... that's wrong too16:48
ttxit feels like desperate workaournd because we can't publish anywhere elese16:49
flaper87clarkb: https://github.com/flaper87.keys16:49
flaper87`/bu3616:49
flaper87ops16:49
asselinclarkb, this sentence is copy and pasted: "When code is pushed to Gerrit, a series of jobs are triggered that run a series32416:49
asselinof tests against the proposed code."16:49
ttxit's clearly not the right place for an autogenerated document16:49
*** tmcpeak has joined #openstack-infra16:49
reedthe advantage is that the wiki is already a destination for many and it has a search engine16:49
ttxI need to publish an ICS file too16:50
*** pcaruana has quit IRC16:50
Ng+11000 for ICS files for meetings16:50
ttxwiki attachment is not fun16:50
clarkbflaper87: jenkins@104.239.132.167 I would wait for the job to finish running before you do too much otherwise you will be fighting each other16:50
jeblairttx, reed: we could always publish to "www2.openstack.org" and move all our random hostnames under there :)16:50
reedbut I feel a piece of me dying trying to defend the undefendable junk :)16:50
*** sabeen1 has joined #openstack-infra16:50
*** rwsu-afk is now known as rwsu16:50
Ngttx: are we going to get a separate ICS feed for each project? :)16:50
clarkbI am going to delete the nodes we held for deepakcs and BharatK now16:50
* jeblair remembers when every major company had sites on like "www14.example.com"...16:51
jeblair(to be clear, i'm kidding about www2)16:51
ttxjeblair: now it's just HP16:51
*** tsg_ has joined #openstack-infra16:51
flaper87clarkb: yup, thanks a lot16:51
jeblairttx: that's right.  keeping it old school.16:51
reedlol16:51
ttxNg: well... yaml2ical (the lib that converts YAML to ICS) can output a single file or one per YAML description16:51
ttxso technically we /could/ publish an all.ics and a nova-meeting.ics up to you to get them all16:52
ttxand we could iterated on that and create a category concept16:52
*** doude has quit IRC16:52
*** carl_baldwin has quit IRC16:52
jeblairttx: i think all + nova, etc would be very popular16:52
ttxNg: feel free to help on https://git.openstack.org/cgit/openstack-infra/yaml2ical/ -- I feel a bit lonely there16:53
ttxself-reviewing is half as fun16:53
anteayamarcusvrn: no, I would have to look the same as you16:53
*** salv-orlando has quit IRC16:53
clarkball other outstanding held nodes have been marked for deletion now16:53
*** tnovacik has quit IRC16:53
*** koolhead17 has joined #openstack-infra16:53
ttxjeblair: I'm not even done and I have feature requests already16:53
redrobothi infra!  Does anyone know why a dsvm gate would vote NOT_REGISTERED?16:53
*** amotoki has joined #openstack-infra16:54
clarkbredrobot: that indicates that gearman does not have that job registered in its list of jobs it can run16:54
*** unicell has joined #openstack-infra16:54
*** unicell1 has quit IRC16:54
marcusvrnanteaya: ok! I will figure it out. thanks!16:54
*** jlibosva has joined #openstack-infra16:54
anteayamarcusvrn: welcome and good luck16:54
clarkbredrobot: I would check that jenkins has nodes available to run that job and that your jenkins gearman config is working16:54
* SpamapS reads backscroll about cloud-init stuff :-P16:55
redrobotclarkb gearman?16:55
SpamapSjroll: question about the nova patch that is bumped to L: is the _spec_ at least approved?16:56
*** tmcpeak has quit IRC16:56
*** amitgandhinz has quit IRC16:56
clarkbredrobot: ya, so the way zuul triggers jobs is via a gearman bus. zuul talks to a gearman server as a client to say "run job X" and jenkins talks to gearman as a bunch of workers saying "I can run job X"16:56
*** tmcpeak has joined #openstack-infra16:56
NgSpamapS: it seems to be, yeah16:56
clarkbredrobot: so you need the jenkins half to advertise to the gearman server that it is capable of running jobs16:56
jrollSpamapS: yes16:57
openstackgerritMatthew Treinish proposed openstack-infra/devstack-gate: DO NOT MERGE COPY ALL OF .TESTREPOSITORY to log server  https://review.openstack.org/14971916:57
jrollSpamapS: http://specs.openstack.org/openstack/nova-specs/specs/kilo/approved/metadata-service-network-info.html16:57
clarkbredrobot: we do this using a jenkins gearman plugin, it needs to be configured in the main global configuration of jenkins16:57
redrobotclarkb I see. Is this in project-config as well?16:57
Ngttx: what sort of stuff needs doing on it?16:58
clarkbredrobot: typically it is pointed at tcp://ip_addr_of_zuul:473016:58
ttxNg: hmm.. by-category ics generation ?16:58
clarkbredrobot: the plugin install should be, but we are not able to manage our global jenkins config with puppet right now16:58
clarkbredrobot: because jenkins is special in what it puts in the global config16:58
ttxNg: I guess I'll roll it out without that feature16:58
ttxSpent a lot of time turning that university project into something barely usable16:59
*** yfried has quit IRC16:59
ttxprobably easier to complete that first16:59
*** tmcpeak has quit IRC16:59
redrobotclarkb I wonder if this is because I asked for a Fedora 21 node? http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/barbican.yaml#n4316:59
*** tmcpeak has joined #openstack-infra16:59
clarkbredrobot: gah for some reason I thought you were asking about a third part ci, I have been hip deep in third party ci stuff too long :)17:00
*** afazekas_ is now known as afazekas17:00
redrobotclarkb ah, no problem :)17:00
jgriffithjeblair: hey, I've wanted to ask you something for a while and youre email just reminded me....17:00
clarkbredrobot: yes, that would be why. because we run jobs on different nodes based on branch there is one other place where you need to update zuul to make that job run on fedora2117:00
clarkbredrobot: I will get a link17:00
*** ihrachyshka has quit IRC17:00
jgriffithjeblair: why nodepool?  Why not just provision/build on demand?17:00
jgriffithjeblair: other than the obvious "time"17:01
clarkbredrobot: https://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/openstack_functions.py#n115 so you can just append -f21 to your job name or update https://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/openstack_functions.py#n39 to special case your job a bit more17:01
*** tmcpeak has quit IRC17:01
*** tmcpeak has joined #openstack-infra17:01
*** amotoki has quit IRC17:02
clarkbredrobot: I have a preference for the -f21 string insertion as the python function is known to work in that case so chance for bugs is low :)17:02
SpamapSjroll: so I'd think if the spec is approved, cloud-init should accept the patches based on the spec.17:02
redrobotclarkb I see. I'll send a patch to change the gate to -f21. Thank you so much!17:02
*** jistr has quit IRC17:02
clarkbredrobot: no problem, and sorry for the initial confusion.17:03
jrollSpamapS: except bumping to next cycle automatically de-approves the spec, AIUI17:03
*** bdpayne has joined #openstack-infra17:03
*** che-arne has quit IRC17:03
*** AJaeger has joined #openstack-infra17:03
SpamapSjroll: that has to change.17:03
*** vdo has quit IRC17:04
SpamapSjroll: like seriously.17:04
jrollSpamapS: sure, but I can't change that.17:04
SpamapSjroll: I understand dropping it from planning.. but the content should be kept as ratified and able to be depended on "once we get to the implementation"17:04
jrollSpamapS: I'm not going to risk letting spec changes break cloud-init, so we wait.17:04
SpamapSOtherwise we basically tie all our futures to one bottleneck.17:05
openstackgerritJohn Griffith proposed openstack-infra/elastic-recheck: Add query for cinder bug 1425310  https://review.openstack.org/15949517:05
openstackbug 1425310 in Cinder "tgtadm: this access control rule does not exist - results in timeout waiting for volume to attach (stuck in detach) or detach" [High,Confirmed] https://launchpad.net/bugs/142531017:05
jrollSpamapS: I'm painfully aware of how broken our process is :/17:05
mordredSpamapS: I sadly agree with jroll on this one - I have no faith in a future-looking process on this one - and a cloud-init that gives me broken network is worse17:05
jroll++17:05
SpamapSjroll: I like mordred's idea. Make it work fine both ways, and let cloud-init adapt if the final is indeed different.17:05
SpamapSbut thats a discussion to have with cloud-init. :-P17:06
*** AJaeger has quit IRC17:06
*** unicell has quit IRC17:06
jrollSpamapS: we've had this discussion with cloud-init and came to the conclusion of "wait until the code lands in nova"17:06
*** hdd has joined #openstack-infra17:06
SpamapSmordred: the internet RFC process has worked this way and I heard that some things based on the RFC process do actually work.17:06
*** unicell has joined #openstack-infra17:06
SpamapSjroll: Yeah, I think they made the wrong call, but maybe there are devils in the details?17:07
SpamapSLike, is it not even possible to cascade?17:07
SpamapSLike, try the way the spec changes show, and if that fails, try the way that actually works now.. ?17:07
jrollSpamapS: I don't trust nova's process to not change the format, and I'm not going to break cloud-init if it does.17:08
jrollI don't want to write compatiibility code for things that aren't even upstream17:08
*** tsg_ has quit IRC17:08
*** Ala has quit IRC17:08
openstackgerritRicardo Carrillo Cruz proposed openstack-infra/shade: Reorder envlist to avoid the rm -fr .testrepository when running tox -epy34  https://review.openstack.org/15954517:09
SpamapSjroll: well, arguably, rax is the upstream for cloud-init17:09
SpamapSjroll: the same way AWS is the upstream for cloud-init17:09
jroll:/17:09
mordredyeah - to a degree, cloud-init shoudl be able to handle every possible random vendor thing that might or might not exist17:10
mordredwhether those things are official or spec or not17:10
mordredthat's really its value17:10
SpamapSYeah, that's basically its job.17:10
mordredif specs really worked, we probably wouldn't need much of cloud-init17:10
SpamapSLet me handle my OS the same way, hide vendor details.17:11
SpamapSmordred: indeed17:11
zaromorning17:11
SpamapSalso, and I don't mean to beat a dead horse but my golf swing is about to tee off using Flicka's eye socket.. DHCP17:11
*** tmcpeak has quit IRC17:12
*** tmcpeak1 has joined #openstack-infra17:12
*** dtantsur is now known as dtantsur|afk17:13
*** tmcpeak has joined #openstack-infra17:15
*** unicell has quit IRC17:16
openstackgerritDouglas Mendizábal proposed openstack-infra/project-config: Rename barbican dogtag devstack-gate job  https://review.openstack.org/15955017:16
*** tonytan4ever has quit IRC17:16
*** scheuran has quit IRC17:17
*** tmcpeak1 has quit IRC17:18
*** annegentle has joined #openstack-infra17:18
*** yamamoto has joined #openstack-infra17:19
*** andreaf_ has joined #openstack-infra17:19
clarkbmordred: has 157512 received any testing?17:20
*** andreaf_ has quit IRC17:20
clarkbmordred: I am particularly worried that if I merge that and restart nodepool that nothing nova related will work17:20
mordredclarkb: /me looks17:20
openstackgerritMichael Johnson proposed openstack-infra/git-review: Choose correct branch for rebase when submitting  https://review.openstack.org/15887717:21
mordredclarkb: no. we should do more testing. I mean, nodepool doesnt have version pins already, so it's not _particularly_ blocked from being broken on new python-novaclient - other than we just don't upgrade all the time17:21
clarkbmordred: right17:22
mordredclarkb: but I would not just glibly approve without care, because of all of the reasons you stated17:22
clarkbbut I know the current version works :)17:22
mordredclarkb: the import change itself is tested to work17:22
mordredclarkb: I think we want to merge this when our logs start filling with deprecation warnings17:22
mordredclarkb: which will mean that we will have upgraded novaclient for some other reason17:22
clarkband likely sooner, but probably not today17:23
mordredyup17:23
*** MaxV has quit IRC17:23
*** annegentle has quit IRC17:23
*** belmoreira has quit IRC17:23
*** yamamoto has quit IRC17:24
*** ihrachyshka has joined #openstack-infra17:24
tosky_reping again,  the recent update of firefox 36 broke selenium (https://code.google.com/p/selenium/issues/detail?id=8399) and the horizon integration tests17:26
tosky_so basically _all_ horizon tests are failing17:27
tosky_which is not nice17:27
tosky_I collected the findings here: https://storyboard.openstack.org/#!/story/200017917:27
tosky_could you raise this in your priority list? This is going to be worse, at least until selenium devs release a new version (they wrote "soon", but...)17:28
*** che-arne has joined #openstack-infra17:28
anteayatosky_: what is it you need us to do?17:29
david-lylesee non-integrated release cycles work17:29
david-lyle:P17:29
*** tmcpeak has quit IRC17:29
*** yamahata has joined #openstack-infra17:29
anteayatosky_: if horizon tests are failing, horizon devs need to fix the tests and submit patches17:29
anteayadavid-lyle: did you need more from us on that?17:29
david-lyletosky_: horizon can propose the test changes17:29
david-lyleanteaya: no17:29
tosky_anteaya: did you read my storyboard tickets?17:29
anteayadavid-lyle: okay thanks17:29
openstackgerritRyan Hsu proposed openstack-infra/os-loganalyze: Update SUPPORTS_SEV regex to support new devstack logs  https://review.openstack.org/15928617:29
david-lyletosky_: we can propose any patch, qa will just have to review it17:30
anteayatosky_: no I did not, perhaps talking with david-lyle is your best direction17:30
johnthetubaguysdague: hmm, gotcha. Looking to make it higher priority in Liberty seems the best we can offer at this point. Its feels like the first I heard it would help infra so much, otherwise it might have got more attention during kilo, we need to get better and understanding around those sorts of impact :(17:30
tosky_the problem is that a test dependency which can't be easily killed or replaced (selenium) is broken thanks to a change in firefox 36, which just landed in Trusty17:31
david-lylethe only thing infra could do is try building a custom image without the newest firefox, which I don't think they want or should do17:31
klrmnclarkb: thanks so much!17:31
david-lyletosky_: we'll fix in horizon17:31
tosky_now, I'm not sure it's possible to disable trusty updates when installing firefox, or use pinning so that the "base" firefox is installed17:31
david-lyletosky_: that was not a serious plan17:32
*** tsg_ has joined #openstack-infra17:32
tosky_david-lyle: that's why I was asking, what is the serious one?17:32
david-lylebut that's what would be in the realm of infra17:32
david-lylemoving to openstack-horizon, you can find me there17:32
*** mriedem is now known as mriedem_away17:34
*** HeOS has quit IRC17:35
*** otter768 has joined #openstack-infra17:37
*** jlvillal|afk is now known as jlvillal17:39
*** SumitNaiksatam has quit IRC17:40
*** derekh has quit IRC17:40
*** otter768 has quit IRC17:41
*** zul has quit IRC17:43
*** che-arne has quit IRC17:43
*** ayoung-mtg is now known as ayoung17:44
jeblairjgriffith: sorry, was one phone... short answer: time :) ... longer answer:17:44
jgriffithjeblair: haha... well17:44
jeblairjgriffith: when nodepool demand is low, (which it used to be generally) the answer is time -- nodes are usually ready to start tests immediately17:44
jgriffithjeblair: shorts fine, just seekign knowledge :)17:44
jeblairjgriffith: when it's under high demand, it essentially degrades to spinning up nodes on demand and we lose the time benefit17:45
jgriffithjeblair: ahhh..17:45
*** carl_baldwin has joined #openstack-infra17:45
openstackgerritClark Boylan proposed openstack-infra/project-config: Upload files concurrently to swift.  https://review.openstack.org/15678817:45
clarkbzaro: ^ thanks a bunch for the review, was helpful17:46
jeblairjgriffith: but even so, we still need _something_ to spin up and destroy those nodes... jenkins is unable to do that because the node spinup time may be counted in the job runtime, and node boot errors are counted as job failures...17:46
jeblairjgriffith: and we have delays and errors constantly :(17:46
SpamapSjeblair: which is basically the same as every thread pool ever implemented. :)17:46
jgriffithjeblair: that makes sense17:46
clarkbzaro: I think I have reviewed the gerrit related changes too, let me know if I have missed any17:46
jgriffithjeblair: one more quick question :)17:46
jeblairSpamapS: indeed!17:46
jgriffithjeblair: from what I've looked at we don't bake anything into the node-pool instances do we?17:46
*** unicell has joined #openstack-infra17:46
jeblairjgriffith: there's also some logic in the current nodepool around getting sets of nodes together in the same AZ, etc, for multinode; i'm proposing to make that a bit more useful and flexible in the future...17:47
jeblairjgriffith: ah, yes, that is another important aspect of nodepool.  we do create our own images.17:47
jgriffithjeblair: hmm... ok, that is key17:47
jeblairjgriffith: currently most images are made by booting a provider image, running some stuff on it, then snapshotting that and booting the real worker instances from the snapshot17:47
jeblairjgriffith: we are building a few images using diskimage-builder and uploading them to glance17:48
jgriffithjeblair: is that code in the nodepool repo and I just missed it?17:48
jgriffithahh17:48
jeblairjgriffith: we're moving toward doing everything in dib so that we can have consistent images in all providers (and also only have to build each image once)17:48
*** mwagner_lap has quit IRC17:48
jeblairjgriffith: turns out talking to glance was harder than expected, so that's a big, in-progress operation17:49
jgriffithjeblair: I've fought image building for a number of reasons the past year17:49
*** salv-orlando has joined #openstack-infra17:49
jgriffithjeblair: TBH glance wasn't even my biggest headache, but sure17:49
jgriffith:)17:49
jeblairjgriffith: the code to actually run the setup scripts and do the image snapshots (and also run dib and upload to glance) is in nodepool; the actual scripts it runs are in the project-config repo in the nodepool/ directory17:50
*** shashankhegde has joined #openstack-infra17:50
jgriffithjeblair: great, I'll have a look17:50
jeblairjgriffith: yeah, we're also having to do things like carry patched cloud-init, etc.  it's all very, er, lovely.17:50
jgriffithjeblair: so one of the things I've been looking at lately is trying to make sure I coorelate possible changes in the image itself17:50
*** cody-somerville has quit IRC17:50
jgriffithjeblair: when looking at changes in behavior etc17:51
* mkj_wk feels better understanding why tests were failing in jenkins and succeeding locally ☺17:51
clarkbjeblair: don't forget its an important piece in being multi cloud residents17:51
clarkbmkj_wk: yay17:51
jgriffithjeblair: yeah, cloud-init has been my headache most recently17:51
jgriffithjeblair: great tool, but install/config of it seems to have some *tricks*17:51
jgriffithfor me at least17:51
jeblairyeah, we tried hard not to need it, but i think mordred was finally convinced it wasn't going to happen :)17:52
*** che-arne has joined #openstack-infra17:52
flaper87clarkb: I'm done with the VM, thank you!17:52
* flaper87 found the issue17:52
jgriffithjeblair: seems to be the way it goes, but at least we try and avoid it :)17:52
clarkbflaper87: no problem will delete it shortly17:52
jgriffithjeblair: ok, that helped alot, thanks for the tutorial17:52
mordredjgriffith: yeah - I hate cloud-init17:53
*** johnthetubaguy is now known as zz_johnthetubagu17:53
mordredjgriffith: but we have at least one cloud that does not dhcp17:53
mordredjgriffith: if all of your clouds dhcp, and you do image building17:53
mordredI think you can skip cloud-init very safely17:53
clarkbgreghaynes: I am going to give a go testing your nodepool ssh key change with dib now17:53
jgriffithmordred: well... what about dynamic partitioning?17:53
mordredshrug17:53
mordreddon't care17:53
clarkbgreghaynes: my dib cache is going to be so stale17:53
jgriffithmordred: LOL17:54
jgriffithmordred: sadly that's a big one for me17:54
mordredjgriffith: I mean, I undersatnd your point17:54
jgriffithmordred: and I yours17:54
mordredit's not a worry for me :)17:54
jgriffithmordred: right.. I got ya17:54
jgriffithmordred: honestly maybe I should "change" me worries :)17:54
mordredjgriffith: well, hopefully we'll have a solid cloud-init dib element that can be re-used soonish17:54
mordredhahahaha17:54
clarkbwe have our jobs do the dynamic partitioning at run time17:55
jgriffithmordred: yeah, I need to look at dib, TBH I haven't paid much attention to the OOO stuff17:55
clarkbit doesn't take long17:55
*** klrmn has left #openstack-infra17:55
jgriffithmordred: I probably should17:55
jgriffithclarkb: oh?  Just using the disk layout input to Nova?17:55
*** dizquierdo has quit IRC17:55
*** cpallares has joined #openstack-infra17:57
mordredjgriffith: nope. we try to minimize the number of things we ask clouds to do for us17:57
mordred:)17:57
mordredjgriffith: also, definitely learn about dib - it's a pretty standalone piece and we're starting to use it more and more extensively in infra17:57
jgriffithmordred: will do, thanks17:58
jgriffithmordred: done monkeying with veewee, packer etc17:58
clarkbjgriffith: no devstack-gate says "I don't have swap and I need a volume for cinder" then it carves up the ephemeral drive into multiple pieces17:58
jgriffithclarkb: ahh.. yeah, I see17:59
jgriffithhmm... I fear I've now just given myself a tangent to distract me for the rest of the day18:00
clarkbhttp://logs.openstack.org/16/155516/1/check/check-tempest-dsvm-full/eaaaae2/logs/devstack-gate-setup-workspace-new.txt.gz#_2015-02-19_23_48_00_62418:00
*** _shaps_ has quit IRC18:00
clarkbthe only expensive bit there is copying old /opt/git onto newly mounted /opt but thats to put a cache in place so is worth the trouble (takes 2 minutes but saves many many more)18:00
jgriffithyeah, and just go direct to parted and do it yourself... cool18:00
jgriffithclarkb: tricky18:01
jgriffithclarkb: or I should say "clever"18:01
openstackgerritRicardo Carrillo Cruz proposed openstack-infra/shade: Add initial compute functional tests to Shade  https://review.openstack.org/15846318:01
*** SumitNaiksatam has joined #openstack-infra18:01
clarkbgreghaynes: confirmed my cache is very cold18:02
*** ivar-lazzaro has joined #openstack-infra18:03
*** emagana has joined #openstack-infra18:03
*** tonytan4ever has joined #openstack-infra18:04
*** Ryan_Lane has joined #openstack-infra18:04
*** AJaeger has joined #openstack-infra18:05
mordredjeblair, clarkb: fyi - I have told ansible upstream it's ok to merge the first set of shade-based modules - so at this point we should consider any potential breaking changes to shade carefully18:05
clarkbmordred: ok18:05
clarkbmordred: maybe tag a 1.0?18:05
*** doude has joined #openstack-infra18:05
mordredwell, there is one change to the current api that is not being consumed anywhere that wants to be changed18:05
mordredthat TheJulia is working on18:06
mordred(related to ironic)18:06
mordredbut as soon as that's in, yes18:06
clarkbrgr18:06
*** SumitNaiksatam has quit IRC18:06
clarkbI should setup a weekly/daily cron to just update my dib cache18:06
*** shashankhegde has quit IRC18:07
*** emagana has quit IRC18:07
*** SumitNaiksatam has joined #openstack-infra18:07
openstackgerritRafael Folco proposed openstack-infra/system-config: Updates to running-your-own CI docs: Initial setup section  https://review.openstack.org/15586418:07
*** zul has joined #openstack-infra18:09
openstackgerritRicardo Carrillo Cruz proposed openstack-infra/shade: Add initial compute functional tests to Shade  https://review.openstack.org/15846318:09
*** mpavone has quit IRC18:11
jeblairkrotscheck: you were asking about apscheduler and sqlalchemy yesterday; i believe apscheduler is unrelated -- other changes are failing api.auth.test_oauth.TestOAuthAccessToken.test_valid_access_token_time18:12
*** doude_ has joined #openstack-infra18:13
krotscheckjeblair: The original issue I found ended up being in pyMySQL 0.6.4 (rather than 0.6.3).18:13
*** vilobhmm has joined #openstack-infra18:13
krotscheckjeblair: What you're seeing is a new thing that I believe is the storyboard DB code not playing well with the apscheduler db code.18:13
jeblairkrotscheck: but apscheduler has not merged, right?18:13
krotscheckjeblair: Nope.18:14
jeblairkrotscheck: there are changes other than apscheduler that are failing that test18:14
krotscheckHrm.18:14
*** tmcpeak has joined #openstack-infra18:14
* krotscheck certifies it as 'works on my machine'.18:14
*** vilobhmm1 has joined #openstack-infra18:14
krotscheckSo what's the diff betwen what I have and infra I wonder.18:14
*** doude has quit IRC18:15
clarkbkrotscheck: have you updated our tox envs recently?18:15
jeblairkrotscheck: 159135 and 159515 are failing.  only one of the jobs in 159135 failed, bouth in 159515.  might be nondeterministic, or 159135 might have just caught the change as it was happening18:15
krotscheckclarkb: Last night, yes.18:15
clarkbkrotscheck: infra gate tests always start with fresh tox envs18:15
clarkbkk18:15
krotscheckclarkb: That's how I found the pymysql thing.18:15
clarkbgotcha18:15
*** tmcpeak has quit IRC18:16
jeblairkrotscheck: did you land a pin for pymsql?18:16
jeblairkrotscheck: or was there a fix in storyboard to support both versions?18:16
krotscheckjeblair: I was going to include that in my apscheduler patch, but I'm pulling it out right now.18:16
anteayajeblair: it is going to take me another couple of passes to fully digest your zuul email18:16
*** afazekas has quit IRC18:16
anteayajeblair: for now can I get a better understanding of how introducing the concept of tenant into the config with change things?18:17
jeblairkrotscheck: oh, so the pymisql thing might be affecting 159135 and 15951518:17
anteayawill18:17
krotscheckjeblair: Maybe?18:17
*** ChuckC has quit IRC18:17
*** vilobhmm has quit IRC18:18
krotscheckAlso, jeblair, I would strongly prefer that you update the js ui to show all the comments, rather than removing things from the API.18:18
jeblairanteaya: it will allow us to adjust how much zuul configuration is shared between projects.  right now, the only answer is "fully shared".  tenants let us set that parameter anywhere between "fully" and "none at all".  we might be somewhere in the middle.18:19
*** annegentle has joined #openstack-infra18:19
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Exclude pyMySQL v0.6.4  https://review.openstack.org/15956418:19
krotschecklet's see if htat works.18:20
anteayahmmm okay, this will be allow the same template to be reused to two different projects with project specific consequences?18:20
jeblairkrotscheck: i'm not sure why the api ever had paging for comments, it does not seem useful to me.  regardless, perhaps we should only put that sort of thing in if we are actually have both sides of the equation prepared.18:21
openstackgerritJames E. Blair proposed openstack-infra/storyboard: Remove API paging from comments  https://review.openstack.org/15951518:22
jeblairstacked ^18:22
*** jlibosva has quit IRC18:22
krotscheckjeblair: That's a valid statement for future development. In this case, however, I strongly prefer moving forward, rather than back, especially given that StoryBoard isn't the most resource intensive project out there.18:22
krotschecki.e. resource == people18:22
greghaynesPaging is useful if you encounter a js error :)18:23
greghaynesIf it is what im thinking18:23
*** tosky_ is now known as tosky18:23
jeblairkrotscheck: is there a reason to have paging for comments?  otherwise, i think that having less code is desirable all around.18:24
*** annegentle has quit IRC18:24
krotscheckAPI sanity?18:25
jeblairgreghaynes: in this case, paging is only returning the first N comments, and requiring a second request for the next N (except our client doesn't do that second part yet)18:25
krotscheckbrowser queued call latency? Same-domain call count restrictions?18:25
greghaynesah, so its purely an API thing18:25
krotscheckProviding a useful experience that loads in a reasonable time on stories that are crazy contentious?18:26
*** bswartz has quit IRC18:26
krotscheckThat comment list isnt' actually a comment APi. it's a list of all events that happened to a story. That is likely to get large.18:27
*** doude_ has quit IRC18:27
krotscheckAnd I apologize if I haven't been able to put love towards the UI since, oh, I dunno, July.18:27
jeblairkrotscheck: fair points.  perhaps we can find something that we could do quickly to make storyboard usable again?  how about we comment out the limit without changing the API signature and throw in a FIXME: add this back when someone gets around to writing the javascript bit?18:28
krotscheckAnd if I had more people willing to contribute to the python side of things the story would DEFINITELY be different.18:28
*** shashankhegde has joined #openstack-infra18:28
openstackgerritJames E. Blair proposed openstack-infra/storyboard: Remove API paging from comments  https://review.openstack.org/15951518:29
*** patrickeast has joined #openstack-infra18:29
openstackgerritJames E. Blair proposed openstack-infra/storyboard: Exclude pyMySQL v0.6.4  https://review.openstack.org/15956418:29
*** cody-somerville has joined #openstack-infra18:29
*** hdd has quit IRC18:30
*** doude has joined #openstack-infra18:30
openstackgerritJames E. Blair proposed openstack-infra/storyboard: Remove API paging from comments  https://review.openstack.org/15951518:30
openstackgerritJames E. Blair proposed openstack-infra/storyboard: Exclude pyMySQL v0.6.4  https://review.openstack.org/15956418:30
* krotscheck is going to shut up now.18:30
SpamapSkrotscheck: regarding paging for comments.. you could copy what launchpad has done for that. You can have "the first and last 10" or "all"18:30
*** tmcpeak has joined #openstack-infra18:31
greghaynesYou know, itd be really neat to hack paging into wsme18:31
*** tmcpeak has quit IRC18:32
greghaynesits seems reasonable to it to just support it when returning collections18:32
*** tmcpeak has joined #openstack-infra18:32
* greghaynes stops derailing the actual issue18:32
*** jlibosva has joined #openstack-infra18:33
*** cdent has quit IRC18:34
*** cody-somerville has quit IRC18:36
openstackgerritMerged openstack-infra/python-storyboardclient: Added user_tokens controller  https://review.openstack.org/15865218:36
*** ihrachyshka has quit IRC18:37
clarkbgreghaynes: your dib element change doesn't appear to have broken existing builds. As soon as files are done being copied I will see if the ssh key update works as well18:37
*** tmcpeak has quit IRC18:37
greghaynes\O/18:38
SpamapSgreghaynes: paging is hard for databases.18:38
greghaynesyea, you would do the lazy eval on collection thing18:38
greghaynesit just would take care of the api clue18:38
greghaynesglue18:38
*** mfink_ has joined #openstack-infra18:38
*** erikmwilson has quit IRC18:38
*** erikwilson is now known as erikmwilson18:38
*** mayurig has joined #openstack-infra18:38
*** _nadya_ has quit IRC18:38
jeblairgreghaynes: but paging is racy without server state18:39
jeblairbecause the second time you execute the query of comments in reverse order, you might be off by one.18:39
*** doude_ has joined #openstack-infra18:39
*** tmcpeak has joined #openstack-infra18:39
greghaynesYep. Thats kind of an orthogonal deal, youd just have to pass through your seed from the previous request18:39
*** doude has quit IRC18:40
*** mtanino has joined #openstack-infra18:40
greghaynesbut yea, I think were all on the same page - its a nontrivial feature18:40
*** ChuckC has joined #openstack-infra18:40
greghaynesI simply mean that wsme would be nice in that it would make sure youre using consistent api bits for doing that, which is kind of the whole idea18:41
*** dteselkin_ has quit IRC18:41
marunanteaya: I have a project-config change that could use your blessing :)18:41
marunhttps://review.openstack.org/#/c/15923218:41
*** tmcpeak has quit IRC18:42
*** cpallares has quit IRC18:42
* anteaya clicks18:42
marunjogo: the skip-if fix has landed: https://review.openstack.org/#/c/156716/18:42
jeblairmarun: but is not in production18:42
*** bswartz has joined #openstack-infra18:43
marunjeblair: ah.  I should have trusted my instinct and said 'appears to have landed'18:43
jeblairmarun: we'll need a zuul restart for that18:43
marunjeblair: I'm guessing that's expensive18:43
*** grantbow has quit IRC18:44
*** ssam2_ has quit IRC18:45
*** ssam2 has quit IRC18:46
jeblairmarun: moderately... i'd like to get a couple of other small fixes in first if we can18:46
*** grantbow has joined #openstack-infra18:46
*** ChuckC has quit IRC18:47
marunjeblair: is state preserved across restarts?18:47
jeblairmarun: only with some manual effort18:47
marunjeblair: or does it effectively mean that ci jobs have to be stopped and restarted?18:47
jeblairmarun: we snapshot it, restart, and reload the state of the queues.  it's a bit silly that's not handled automatically, but the last time i touched the restart code, there were 6 projects and "wait for jobs to complete" was a viable restart strategy.18:48
jeblairmarun: it isn't really anymore :)18:48
openstackgerritMerged openstack-infra/storyboard: Added branches to storyboard  https://review.openstack.org/15044718:48
*** coolsvap is now known as coolsvap_18:48
*** gyee has quit IRC18:48
marunjeblair: fair enough.  is there an open issue in storyboard?18:48
*** bswartz1 has joined #openstack-infra18:48
jeblairmarun: i don't think so18:49
*** dteselkin_ has joined #openstack-infra18:49
*** ChuckC has joined #openstack-infra18:49
*** bswartz has quit IRC18:49
*** bswartz1 is now known as bswartz18:49
*** melwitt has joined #openstack-infra18:49
jeblairmarun: i think what it should do is abort all running jobs, save the queue state, exit, then on startup, load a saved queue state if it exists.18:49
jeblair(the abort all running jobs will help nodepool cleanup faster)18:49
clarkbin fairness the auto restart safely code does work, it just queues everything up for a while18:50
*** ihrachyshka has joined #openstack-infra18:50
*** jlibosva has quit IRC18:51
*** dims_ has joined #openstack-infra18:51
jeblairclarkb: yeah, so restarts would queue work for probably at least an hour or two, possibly many hours if we're having a bad day18:52
*** dims_ has quit IRC18:53
*** dims_ has joined #openstack-infra18:54
openstackgerritDavid Lyle proposed openstack-infra/project-config: Pinning firefox version temporarily  https://review.openstack.org/15957418:54
*** dims has quit IRC18:55
clarkbdavid-lyle: ^ that macro update will only affect the horizon integration test. Is the normal selenium test also affected?18:56
david-lyleclarkb: no, the process for installing selenium is different, for some reason18:56
david-lylenot entirely sure why at this point18:57
david-lylebut the base selenium job uses the version built into the image18:57
david-lyleof firefox18:57
openstackgerritMerged openstack-infra/project-config: Use neutron's api job to validate func job change  https://review.openstack.org/15923218:57
david-lyleI will work on tracking down the need for a separate install18:58
jeblairclarkb, marun: i would really like to get 155160 157080 in a restart soon as they address operational issues we have seen.18:58
openstackgerritAndreas Jaeger proposed openstack-infra/project-config: Run some project-config jobs only if needed  https://review.openstack.org/15951218:58
clarkbdavid-lyle: right, we use the unittest nodes for the selenium unittests. You can probably get away with adding that firefox install macro to the normal selenium job right now18:58
clarkbdavid-lyle: in fact I will -1 so that you can do it in that change18:59
clarkbdavid-lyle: that should fix selenium across the board18:59
clarkbjeblair: ok will look18:59
david-lyleclarkb: it hasn't been breaking in the other selenium suite yet18:59
clarkbdavid-lyle: probably because our image builds haven't pciked up new firefox yet, but should tomorrow19:00
clarkbdavid-lyle: or maybe this isn't a 100% fail? I am happy to wait for it to break before fixing it if you want19:00
*** e0ne is now known as e0ne_19:00
*** markus_z has quit IRC19:00
toskydavid-lyle: thanks; just note that =35 wouldn't work, as it's not in any repository anymore; the version that could be installed is the base one, 28.0(+build2)19:00
david-lyletosky: oh, that's super19:01
*** prad has quit IRC19:01
openstackgerritgreghaynes proposed openstack-infra/project-config: Support custom static nameserver during build  https://review.openstack.org/15755419:01
AJaegerjeblair: you gave a -1 to 159512, is this nothing you want - or do you want me to improve it?19:01
david-lyleclarkb, I'll WIP that patch for a bit19:01
AJaegerjeblair: ah, saw your comments coming in...19:01
AJaegerjeblair: will update - thanks for the quick review19:01
jeblairAJaeger: yeah, we were racing each other there, sorry :)19:02
AJaegerjeblair: no problem ;)19:03
AJaegerjeblair: I did something similar for documentation jobs, could you review that one as well, please? https://review.openstack.org/15840619:03
openstackgerritAndreas Jaeger proposed openstack-infra/project-config: Run some project-config jobs only if needed  https://review.openstack.org/15951219:04
AJaegerjeblair: updated the project-config patch19:04
openstackgerritMerged openstack-infra/zuul: Don't report on non-live changes dequeueing  https://review.openstack.org/15708019:06
*** hdd has joined #openstack-infra19:07
AJaegerclarkb, jeblair: Should every IRC channel that is in gerritbot/channels.yaml also be setup with accessbot via accessbot/channels.yaml?19:07
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Reduce the number of timezones under test.  https://review.openstack.org/15957619:08
clarkbAJaeger: yes, at the very least they need to have the correct perms and accessbot is used to enforce that so transitively yes19:08
AJaegerclarkb: There're a couple of channels where this is not the case - I'll review better in the future and will update this now.19:09
*** bswartz has quit IRC19:09
*** koolhead17 has quit IRC19:09
*** bswartz has joined #openstack-infra19:09
clarkbgreghaynes: comment on 15755419:10
clarkbAJaeger: ok, we may also need to make the docs more explicit19:10
*** dimtruck is now known as zz_dimtruck19:10
clarkbgreghaynes: but didn't vote because unsure if it matters for you usecase19:10
greghaynesclarkb: yea, I considered it, decided to name the variable NODEPOOL_BUILD instead but not for any good reason19:11
*** ZZelle_ has joined #openstack-infra19:11
greghayneswell, my specific use I only care about build - my local mirror requires me to use a local dns19:11
clarkbgreghaynes: I just wonder if you don't want to use google during build do you want to use it after your node boots?19:11
AJaegerclarkb: infra manual does not mention setting up new channels19:11
greghaynesI am inclined to do your suggestion actually, because why not19:11
clarkbAJaeger: http://ci.openstack.org/irc.html does, maybe we need to move some of that content19:12
AJaegerclarkb: or at least reference the page from infra/manual19:13
AJaegerclarkb: is there a way to figure out whether openstackinfra has access to a channel?19:14
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Replaced python-crontab with apscheduler  https://review.openstack.org/15895519:14
*** erikwilson has joined #openstack-infra19:14
*** erikmwilson is now known as Guest5515519:14
*** erikwilson is now known as erikmwilson19:14
*** erikwilson has joined #openstack-infra19:14
clarkbAJaeger: yup, one sec19:14
*** erikmwilson has quit IRC19:15
*** erikwilson is now known as erikmwilson19:15
clarkbAJaeger: /msg chanserv access #channel list19:15
AJaegerthanks, clarkb !19:15
*** ihrachyshka has quit IRC19:16
openstackgerritTom Cameron proposed openstack-infra/project-config: Improve gate-check logging  https://review.openstack.org/15957819:16
*** sabeen1 has quit IRC19:16
greghaynesclarkb: is there really a desire to keep the puppet and nodepool-base elements independent?19:17
clarkbgreghaynes: no, but we have to until we can dib everything19:18
clarkbgreghaynes: there isn't a 1:1 because dib makes a few things weird like the dns stuff19:18
greghaynesoh? so the elements/puppet element is used for other things ATM?19:18
clarkboh maybe I misread, we have to keep nodepool/elements and nodepool/scripts separate19:19
greghaynesbasically, if I put a elements-deps on nodepool-base in the pupet element how bad would that be19:19
clarkbI think nodepool-base and puppet elements are separate because puppet is independently useful eg when we start to dib our long lived servers19:19
clarkbya we should avoid that19:19
greghaynesok19:19
*** armax has quit IRC19:20
clarkbgreghaynes: OH! hrm ya we may be coupling the finalise 99-unbound script there a bit too closely to puppet element19:20
clarkbgreghaynes: its possible we need to uncouple that a bit more19:20
*** bkero has quit IRC19:21
*** coolsvap_ is now known as coolsvap19:21
openstackgerritAdam Sheldon proposed openstack-infra/gitdm: Add Adam Sheldon to email mapp file.  https://review.openstack.org/15927419:21
*** bkero has joined #openstack-infra19:21
*** zz_dimtruck is now known as dimtruck19:22
AJaegerclarkb, I reviewed the missing channels and commented on http://paste.openstack.org/show/182614/19:22
AJaegerI think it's fine to add those to accessbot where the setup is fine19:23
AJaegerFor the no access and not registered ones, we just ignore them - correct?19:23
clarkbAJaeger: and the master list you used was for gerritbot's channels?19:23
AJaegerclarkb: correct19:23
clarkbAJaeger: I think those with no access and not registered we hop on the channel and ask for them to do that and remove gerritbot at some date in the futuerif the don't19:23
clarkbsay give them a couple weeks19:24
AJaegerclarkb: I looked for all channels in gerritbot and then search accessbot19:24
clarkbAJaeger: gotcha19:24
*** sabeen has joined #openstack-infra19:24
AJaegerclarkb: then let me do a first patch and jump into the channels...19:24
*** HeOS has joined #openstack-infra19:25
jogomarun: thanks!!19:26
*** doude_ has quit IRC19:27
marunjogo: it's not live yet, but hopefully soon.19:27
AJaegerclarkb: #dragonflow has only #openstackgerrit in it ;(19:27
clarkbAJaeger: I think we can just remove the bot from that channel then19:27
clarkbAJaeger: just note in the change the channel was empty and can be recreated later with proper registration and access if desired19:28
AJaegerok19:28
*** doude has joined #openstack-infra19:28
jogomarun: saw, I will post the nova patch again, in preparation though19:28
openstackgerritJoe Gordon proposed openstack-infra/project-config: Revert "Revert "Add conditional skips for nova jobs""  https://review.openstack.org/15958219:31
*** pelix has quit IRC19:31
*** vkmc has joined #openstack-infra19:31
*** doude has quit IRC19:31
*** matrohon has joined #openstack-infra19:31
vkmcAJaeger, hi there! I wanted to ask you about this change https://review.openstack.org/#/c/159222/119:32
*** tmcpeak has joined #openstack-infra19:32
jogomarun: posted and added a comment explaining the current status19:32
AJaegervkmc: yes19:32
vkmcAJaeger, could you let me know how this change https://review.openstack.org/#/c/156328/ affects on having those tox targets?19:33
vkmcAJaeger, if you have a moment, of course19:33
AJaegervkmc: what's did you see the two previous changes? https://review.openstack.org/#/c/156328/ and https://review.openstack.org/#/c/156315/ ?19:33
AJaegerthis patch is just a cleanup.19:33
AJaegervkmc: 156328 removed the tox environments from the gates after 156315 made them nops - and now it's cleanup19:34
*** MaxV has joined #openstack-infra19:34
*** purp has joined #openstack-infra19:37
vkmcAJaeger, oh I see, thanks :)19:37
*** koolhead17 has joined #openstack-infra19:37
*** koolhead17 has joined #openstack-infra19:37
marcusvrnanteaya: is there a way to change the ID of the CI account?19:37
*** otter768 has joined #openstack-infra19:38
*** mwagner_lap has joined #openstack-infra19:39
*** tmcpeak has quit IRC19:39
openstackgerritAndreas Jaeger proposed openstack-infra/project-config: Cleanup accessbot and gerritbot  https://review.openstack.org/15958519:40
AJaegerclarkb: here's the first cleanup as discussed ^19:41
AJaegerclarkb: thanks for your help!19:41
*** otter768 has quit IRC19:43
* AJaeger waves good-bye!19:43
*** AJaeger has quit IRC19:43
*** tmcpeak has joined #openstack-infra19:44
*** achanda has joined #openstack-infra19:45
openstackgerritSteve Martinelli proposed openstack-infra/project-config: Revert "Remove lib-forward-testing from openstackclient"  https://review.openstack.org/15959019:45
openstackgerritSteve Martinelli proposed openstack-infra/project-config: Revert "Remove lib-forward-testing from openstackclient"  https://review.openstack.org/15959019:45
SpamapSjeblair: btw, I just reported this: https://storyboard.openstack.org/#!/story/2000181 ... any ideas before I jump into the rabbit hole a little?19:47
*** ddieterly has quit IRC19:47
*** ddieterly has joined #openstack-infra19:47
*** vkmc has left #openstack-infra19:48
greghaynesclarkb: yep, im thinking I might break out unbound / static nameserver setting into its own element19:48
openstackgerritSpencer Krum proposed openstack-infra/system-config: Update docs for puppet 3  https://review.openstack.org/15627219:48
clarkbsdague: want to weigh in on 156456?19:48
clarkbgreghaynes: that would probably be best19:49
nibalizerclarkb: pleia2 care to review https://review.openstack.org/#/c/159085/ ? im actually using that now, and it breaks without that change19:50
anteayamarcusvrn: the gerrit id? no, that number is the unique id for the account, you would have to create another account19:50
*** penguinRaider has quit IRC19:50
anteayamarcusvrn: what are you trying to do that changing the account id seems like the solution?19:50
openstackgerritTom Cameron proposed openstack-infra/project-config: Improve gate-check logging  https://review.openstack.org/15957819:51
clarkbjeblair: for the hidden flag in gertty if a new comment comes in it gets unhidden but if I hit reviewed flag it won't show up against until a new patchset arrives?19:51
*** e0ne_ is now known as e0ne19:52
marcusvrnanteaya: no, actually the id created was "hds-hnas-ci", but this account has all CIs for HDS (hnas and hbsd). So if I can change the name from "hds-hnas-ci" to "hds-ci", it would be better. But it's not a problem...just to keep more clear19:53
*** tmcpeak has quit IRC19:55
marcusvrnanteaya: and about the json...if you want to know where can we add the CI on stackalytics, is in default_data into driverlog project :)19:55
*** jkraj has quit IRC19:55
anteayamarcusvrn: ah that is the gerrit username19:56
anteayaso if your account was created by infra, no you can't change the gerrit username yourself19:56
viglesiasjeblair: i heard you had some musings on zuul v3, do you have a pointer to that?19:56
anteayaif the account was created by yourself using a launchpad account, well you would have had to change the username when you first signed into the account using your launchpad credentials19:57
viglesiasor anyone for that matter, im guessing its on a list19:57
anteayaonce the gerrit username is established it can't be changed19:57
marcusvrnanteaya: yeah, it was created by infra team (we sent an email with all info and infra team has created)19:57
anteayaokay19:57
anteayamarcusvrn: if you want to make changes to your ci gerrit account follow these instructions: https://wiki.openstack.org/wiki/OldtoNewGerritCIAccount19:58
anteayait tells you how to create a new account using launchpad19:58
*** bdpayne_ has joined #openstack-infra19:59
anteayaremember that the first time you sign into gerrit with the launchpad credentials you can set your gerrit username19:59
openstackgerritMerged openstack-infra/project-config: Respect NODEPOOL_SSH_KEY in prepare-node for dib  https://review.openstack.org/15720519:59
anteayaafter it is set it can't be changed19:59
anteayamarcusvrn: and thanks for the feedback on stackalytics19:59
*** tosky has quit IRC20:00
anteayaI said change the gerrit username when you first sign in and that is incorrect, I mean set the gerrit username when you first signin20:00
*** bdpayne has quit IRC20:01
*** bdpayne_ has quit IRC20:02
anteayaviglesias: http://lists.openstack.org/pipermail/openstack-infra/2015-February/002471.html20:02
viglesiasanteaya: thank you!!!!20:04
sdagueclarkb: so... isn't 156456 in the wrong direction?20:05
anteayaviglesias: welcome20:05
marcusvrnanteaya: nice! thanks and you are welcome!20:05
clarkbsdague: that was jogo's assertion and you wrote the proposed spec so figured you could weigh in20:05
anteayamarcusvrn: sure20:05
clarkbsdague: but in general yes I think so. keystonecleitn and OSC should both have their own functional tests that enforce their functional contracts20:05
openstackgerritJohn Griffith proposed openstack-infra/elastic-recheck: Add query for cinder bug 1425310  https://review.openstack.org/15949520:05
openstackbug 1425310 in Cinder "tgtadm: this access control rule does not exist - results in timeout waiting for volume to attach (stuck in detach) or detach" [High,Confirmed] https://launchpad.net/bugs/142531020:06
openstackgerritTom Cameron proposed openstack-infra/project-config: Improve gate-check logging  https://review.openstack.org/15957820:06
sdagueright, and more importantly, you'd want to test keystone client patches by the consumer, not randomly pick keystone client head for osc patches, if you cared about that linkage20:06
*** vilobhmm1 has quit IRC20:07
*** vilobhmm has joined #openstack-infra20:07
clarkbSpamapS: around? 158148, why the silent queue?20:11
clarkbsdague: ya20:11
sdagueok, comments left20:12
clarkbsdague: thank you20:12
pleia2nibalizer: good catch, approved20:13
*** MaxV_ has joined #openstack-infra20:15
*** shashankhegde has quit IRC20:15
*** tonytan4ever has quit IRC20:16
SpamapSclarkb: because the patches to fix py34 haven't landed yet20:16
SpamapSclarkb: if they land, I'll flip it. :)20:16
SpamapSclarkb: https://review.openstack.org/#/c/158147/20:17
clarkbSpamapS: why not experimental or non voting check?20:17
clarkbSpamapS: my concern with silent is we use that for trending data collection and isn't really helpful in the make $X thing work20:17
SpamapSclarkb: because I don't understand what each pipeline does.20:17
SpamapSand I didn't see anything that said "non voting"20:17
SpamapShttp://ci.openstack.org/test-infra-requirements.html#test-run-styles <-- does not show "non voting"20:18
*** yfried has joined #openstack-infra20:19
*** MaxV has quit IRC20:19
clarkbSpamapS: voting can be a pipeline config (what to do on success/failure) or a job option voting: false20:19
*** mayurig has quit IRC20:19
clarkbso you can put the job in check and toggle the job's voting flag to false to get non voting jobs in check20:19
SpamapShah you know what, I looked for 'vote'20:20
SpamapSnot 'voting'20:20
openstackgerritMerged openstack-infra/pypi-mirror: Mirror appropriate wheels in main mirror  https://review.openstack.org/15948520:21
*** koolhead17 has quit IRC20:22
openstackgerritMerged openstack-infra/puppet-httpd: Remove a2mod definion  https://review.openstack.org/15908520:22
*** yamamoto has joined #openstack-infra20:22
*** bdpayne has joined #openstack-infra20:22
SpamapSclarkb: you guys are so much better at writing docs than basically anything I've worked with before.. I really should just start at docs and not the cargo cult :-P20:22
SpamapShttp://ci.openstack.org/zuul/zuul.html#configuration has the answers20:22
clarkbSpamapS: they aren't perfect but we try :)20:23
SpamapSthey are20:23
SpamapSI mean20:23
SpamapSthey are as in, they exist20:23
SpamapSwhich is more than I'm used to :)20:23
SpamapSclarkb: how about we just merge the py34 fixes and then I can just let it vote. :)20:23
clarkbSpamapS: that sounds good too20:23
openstackgerritMerged openstack-infra/elastic-recheck: Add query for cinder bug 1425310  https://review.openstack.org/15949520:24
openstackbug 1425310 in Cinder "tgtadm: this access control rule does not exist - results in timeout waiting for volume to attach (stuck in detach) or detach" [High,Confirmed] https://launchpad.net/bugs/142531020:24
pleia2hm, what's the name of our backups server?20:24
clarkbpleia2: http://ci.openstack.org/sysadmin.html#backups20:25
pleia2clarkb: oh lovely, documented :)20:25
* clarkb reenforces what SpamapS has discovered >_>20:25
pleia2clarkb: ok to add to cacti?20:25
clarkbpleia2: maybe? I know jeblair intentionally kept that node in something of a bubble so that the outside wouldn't affect it a whole lot20:26
clarkbbut cacti is probably safe20:26
clarkbeg its not puppet making config changes20:26
*** yamamoto has quit IRC20:26
pleia2clarkb: I worry that one has a full disk20:27
clarkbpleia2: its a good thing to worry about20:27
pleia2based on root messages from last night20:27
pleia2ci-backup-rs-ord.openstack.org20:28
clarkbconfirmed20:28
clarkb:/20:28
*** penguinRaider has joined #openstack-infra20:28
clarkbjeblair: ^ what was your plan for dealing with that when we have stricktly append only backups?20:28
*** mayurig has joined #openstack-infra20:29
* clarkb imagines building a new node, moving DNS and starting a new set of append only backups20:29
*** tmcpeak has joined #openstack-infra20:29
clarkbthen delete old node once we don't need that set anymore20:29
clarkb(if ever)20:29
*** pc_m has quit IRC20:30
mordredclarkb: I just accidentally learned a new thing about ansible inventories from scanning the GCE guide20:30
mordred"As with all dynamic inventory plugins in Ansible, you can configure the inventory path in ansible.cfg. The recommended way to use the inventory is to create an inventory directory, and place both the gce.py script and a file containing localhost in it. This can allow for cloud inventory to be used alongside local inventory (such as a physical datacenter) or machines running in different providers."20:30
mordredwhich is to say - I will make a patch for us to do that - because that's very nice and clean and easy to manage with the way we manage things20:30
mordredI mean, we won't put gce.py in there20:30
*** sigmavirus24 is now known as sigmavirus24_awa20:32
*** tmcpeak has quit IRC20:33
*** bdpayne has quit IRC20:34
clarkbpleia2: I am considering other options but I think with the way bup works the simplest thing is a new node. Hopefully jeblair can chime in too20:34
jeblairSpamapS: i left a comment on your story -- i want gertty to use gerrit's search syntax, and that also doesn't work in gerrit.  of course, it could, er, have a better error message :)20:34
jeblairclarkb: get more space?20:35
*** bdpayne has joined #openstack-infra20:35
clarkbjeblair: I thought fungi said we can't do resizes anymore but I may have misunderstood20:35
jeblairclarkb: that node predates cinder20:35
mordredthat node is very old20:35
jeblairclarkb: so i think we should make a new node with big cinder volumes and move the data over20:35
clarkbjeblair: I see20:36
clarkbany idea if our cinder quota is region specific?20:36
clarkbI don't think it is so that should work in ord20:36
jeblairpleia2: want to spin up the replacement? :)20:37
pleia2I can't get ci-backup-hp-az1.openstack.org to respond (nc on ssh or ping), but could be just an hpcloud thing?20:37
pleia2jeblair: well sure20:38
jeblairpleia2: it's an hpcloud deleted that node thing :)20:38
mordred*headdesk*20:38
jeblairthat happened long ago20:38
pleia2jeblair: hah, so time to update docs to stop mentioning it?20:38
jeblairpleia2: if you want to spin up one there too, that would be cool20:38
pleia2or that, sure20:38
jeblairpleia2: no, i think we really do want a second one20:38
* pleia2 nods20:38
pleia2following instructions here? https://git.openstack.org/cgit/openstack-infra/system-config/tree/launch/README20:39
clarkb+1 to the second one. The non jenkins account in hpcloud should work now after jeblair unsuspended it20:39
jeblairi'm going to remove the jenkin-dev backups on rs-ord20:40
clarkbpleia2: ya, that doc is mostly correct. I would double check flavor and image before running the launch script though20:40
clarkbjeblair: +120:40
jeblairpleia2: since we'll move to cinder, we can have a smaller flavor20:40
jeblair(current flavor is big just to get a big local disk)20:40
clarkbpleia2: `nova flavor-list` and `nova image-list` after sourcing that file will show you what options we have20:40
* pleia2 nods20:40
mordredooh. it's pleia2's first node launch!20:41
pleia2the week of pleia2's first so-many-things :)20:41
mordredpleia2: :)20:41
jeblairpleia2: would probably do something between 2 and 8 G; hard to say withut cacti; i'd make sure it has a few cpus20:42
*** purp has quit IRC20:42
clarkbalso we can use the performance flavors20:42
clarkbI am going to get some lunch now, but will be around for other questioins if they come up20:43
*** penguinRaider has quit IRC20:44
pleia2same FQDN as the server we're replacing?20:45
jeblairpleia2: yeah20:45
jeblairpleia2: just don't set the forward dns yet20:45
* pleia2 nods20:45
jeblairpleia2: (reverse is okay now, we'll switch forward once we've moved the data)20:45
jeblairpleia2: that's part of why the launch script prints out those commands instead of just doing it20:46
morganfainbergjroll, ping that fix for the catalog stuff: https://review.openstack.org/#/c/159315/ though we should have a stern talking to people for defining the same service name as two totally different things.20:46
*** mriedem_away is now known as mriedem20:47
jrollmorganfainberg: agree, but I have no idea who to yell at about that20:47
jrollmorganfainberg: you should have mordred test that for you :)20:47
morganfainbergjroll, already asked him to when he has a chance.20:47
mordredit's on my tdl20:47
mordredjroll: probably pvo20:48
morganfainbergjroll, this was a ping so we could try and figure out who we should have the stern talking to.20:48
*** mayurig has quit IRC20:48
jrolllol20:48
jrollhopefully we'll just retire first-gen and then we can stop20:48
morganfainbergjroll, also poking at keystone folks. but this one deserves explaining to people why this is a baaaaad idea.20:48
morganfainbergin how every many words it takes.20:49
openstackgerritErno Kuvaja proposed openstack-dev/hacking: Add own block in imports for six overwrites  https://review.openstack.org/15919620:49
morganfainbergjroll, we should still explain why this is still bad so thery don't do it again20:49
jogoso https://launchpad.net/openstack-api-site is not part of openstack group on launchpad20:49
jrollmorganfainberg: ya, I just have no idea what email address to give you for that20:49
jrollmorganfainberg: except maybe pvo but idk20:49
*** tonytan4ever has joined #openstack-infra20:50
*** penguinRaider has joined #openstack-infra20:50
jogofungi: who else besides anne can fix that?20:51
SpamapSjeblair: thanks! :)20:52
jeblairjogo: i can20:52
*** purp has joined #openstack-infra20:52
jogojeblair: thanks!20:52
jeblairjogo: oh wait, i can not20:53
jeblairjogo: only anne.  we need her te make https://launchpad.net/~openstack-doc-core the maintainer20:53
*** dustins has joined #openstack-infra20:54
jogotrying to use project membership to openstack group to test e-r queries20:54
jogoturns out to be more work then I expected20:54
*** penguinRaider has quit IRC20:56
*** morganfainberg is now known as needscoffee20:57
mtreinishjogo: heh, I guess that another reason it's good my old test wasn't gating on new queries20:58
ctlaughI am getting errors trying to run nodepool where it's throwing an exception: NotFound: No more floating ips available.  I have a pool named 'public' with 200+ IPs in it, and I have allocated several of them to the tenant.  Any suggesions what I've missed?21:03
jogomtreinish: since they hit the network the tests are fairly slow too 160 seconds or so on my connection21:03
*** rfolco has quit IRC21:03
clarkbyou may need to specify the pool name21:03
clarkbctlaugh pool: name under your provider in the yaml21:03
*** kgiusti has left #openstack-infra21:06
*** mrmartin has joined #openstack-infra21:06
pleia2having some trouble, can a core take a look at ~root/ci-launch/openstackci-rs-nova.sh and make sure we have the right dfw endpoint?21:08
openstackgerritJulia Kreger proposed openstack-infra/shade: Add behavior to enable ironic noauth mode  https://review.openstack.org/15960921:08
pleia2can't even do nova flavor-list21:08
pleia2I haven't worked much with rax before21:09
pleia2s/core/root21:09
clarkboh wait21:09
TheJuliamordred: The above is not what we discussed a while back, but I realized that the class is inherited, and given that ironic's use is fairly limited, I could see keeping it in an operator class.  Ultimately it is a really fast and simple change if there is a desire to do that, but working with it I see why it was done21:09
clarkbwe dont want that file because we want ord21:09
clarkbpleia2 ^21:10
pleia2k21:10
clarkbI will look as soon as cookies get out of oven21:10
pleia2cookies++21:10
mordredTheJulia: cool21:11
pleia2maybe create openstackci-rs-nova.sh for regions21:11
mordredTheJulia: if it really does make sense, then awesome21:11
*** purp has quit IRC21:11
clarkbpleia2 I think that may exist already21:11
clarkbbut nova may be broken due to the catalog thing21:11
clarkbmordred ^21:11
pleia2they exist for different hp regions, and jenkinsy things21:12
mordredhrm21:12
mordrednova flavor-list shoudl work - the nova client special cases things21:12
pleia2it just happens to be the typical, harmless command I use to test creds21:13
mordredTheJulia: I have an evil suggestion...21:13
TheJuliaEvil, we like evil21:13
*** cnesa1 has quit IRC21:13
openstackgerritJoe Gordon proposed openstack-infra/elastic-recheck: Add launchpad test to test_queries  https://review.openstack.org/15513821:13
*** tmcpeak has joined #openstack-infra21:14
mordredTheJulia: what if you override the auth_token property in OperatorCloud and do "if self._auth_plugin in ('', 'None', 'ironic'): return self._auth_token else: return super(OperatorCloud, self).auth_token()21:14
mordredTheJulia: so that the ironic token name hack is localized to places where ironic is likely to be used21:15
*** coolsvap is now known as coolsvap_21:15
*** ldnunes has quit IRC21:15
*** FallenPegasus has joined #openstack-infra21:17
*** markmcclain has quit IRC21:17
SpamapSclarkb: I was just reviewing https://review.openstack.org/#/c/154304/ and it made me think of https://www.youtube.com/watch?v=Q4XCZfkGF8k ... ;-)21:17
* SpamapS writes this on a chalk board 20 times: "I will not be a pedant. I will not be a pedant..."21:18
TheJuliamordred: thats not a bad idea21:19
jeblairSpamapS: oh yeah, prepositions are not to end sentences with.21:19
clarkbSpamapS: that change will prbaoby get abandoned anyways21:19
* TheJulia gives it a try21:19
*** markmcclain has joined #openstack-infra21:19
clarkbmordred is going at it in a different way21:19
pleia2jeblair: hahaha21:19
clarkbpleia2: ERROR (EndpointNotFound): publicURL endpoint for compute service in DFW region not found <- is that what you get?21:20
pleia2clarkb: yep21:20
clarkbalso SecurityWarning: Certificate has no `subjectAltName`21:20
clarkbI think this is the requests thing?21:20
clarkbmordred: ^21:20
pleia2that too, with the bug report21:20
*** sabeen has quit IRC21:20
*** sabeen has joined #openstack-infra21:21
*** ihrachyshka has joined #openstack-infra21:21
mordredyes. the securitywarning is normal21:21
*** shashankhegde has joined #openstack-infra21:21
mordredclarkb, pleia2 which region are you trying?21:21
*** changbl has quit IRC21:21
clarkbmordred: dfw21:21
pleia2it's what is hard coded in our ~root/ci-launch/openstack-rs-nova.sh21:22
mordredof the rackspaceci account/21:22
mordred?21:22
pleia2yep21:22
clarkbya the one in the launch readme21:22
clarkb(we need to not use it for this particular use case but figuring out the problem would be good)21:22
mordredyeah - investigating  - might be that issue21:23
mordredone sec I'll hav ea workaroudn21:23
*** tmcpeak has quit IRC21:23
clarkbord is the same21:23
clarkbpleia2: for ORD just export OS_REGION_NAME=ORD and export os_region_name=ord after sourcing that file21:23
clarkbalso I ate too many hot cookies in a short period of time21:24
*** tmcpeak has joined #openstack-infra21:24
anteayaclarkb: oh nice21:24
*** tmcpeak has quit IRC21:24
jeblairclarkb: i wish i had your problems21:24
mordrednova --service-name=cloudServersOpenStack flavor-list21:25
mordredworks21:25
pleia2clarkb: yep, will do21:25
mordredpleia2: ^^ clarkb ^^21:25
pleia2mordred: oh, obviously (what)21:25
clarkbmordred: we probably need to update the launch script too?21:25
mordredso - the problem is in the wild21:25
*** dimtruck is now known as zz_dimtruck21:25
mordredneedscoffee: ^^ fun and join21:25
mordredjoy21:25
mordredclarkb: yes21:25
mordredI can make a patch - one sec21:25
mordredalso - I suppose it's tie for me to port launch_node.py to shade, since shade already has the fix for this21:25
mordredbut I'll just make a launch_node.py fix for now21:26
needscoffeeReading.21:26
mordredthis is the two-services-named-compute21:26
mordredneedscoffee: this is what it looks like21:26
needscoffeeYep21:26
needscoffeeSure does.21:26
*** jamesmcarthur has joined #openstack-infra21:26
needscoffeeWow. Fun times.21:26
needscoffeeYeah I'll get that fix for keystoneclient rolled up and released soon.21:27
pleia2\o/21:27
clarkbSlickNik: fyi 159423 is a thing21:28
clarkbSlickNik: I assume you are ok with that since the functional job is newer than icehouse21:28
openstackgerritMarton Kiss proposed openstack-infra/puppet-drupal: Get cron key variable from conf parameter  https://review.openstack.org/15942621:29
pleia2so I'm a pretty strict adherent to sudo, but in the case of spinups this seems to be a think we go to root shell for?21:29
pleia2s/think/thing21:29
mordredclarkb, pleia2: while I'm woring on this, could you add export OS_SERVICE_NAME=cloudServersOpenStack to the rackspace credentials files?21:29
mrmartinhi21:30
clarkbpleia2: no it should all work from your user21:30
*** sabeen has quit IRC21:30
clarkbpleia2: you do need to be in the puppet group thouhg iirc21:30
mrmartinclarkb: https://review.openstack.org/#/c/159426/ I made changes you requested here21:30
pleia2clarkb: ah, yes, I can read that file as user, thanks :)21:30
openstackgerritMonty Taylor proposed openstack-infra/system-config: Pass in service name to nova constructor.  https://review.openstack.org/15961521:30
clarkbpleia2: ya, add yourself to the puppet group then everything should work21:31
pleia2yep, in it, thanks21:31
jheskethMorning21:31
pleia2g'day jhesketh21:31
*** ghostpl_ has joined #openstack-infra21:31
clarkbjhesketh: I pushed a new patchset of the upload logs to swift concurrently change after some feedback from zaro, any chance you want to diff that and see if th eedits make sense to you21:31
*** mayurig has joined #openstack-infra21:32
needscoffeemordred: I should be able to release that fix to the wild today. Just approved the change.21:32
mordredclarkb, pleia2: ^^ see if that works - also - btw - OS_SERVICE_NAME does not propagate down into nova client - so flavor-list will have to be --service-name=$OS_SERVICE_NAME21:32
mordredneedscoffee: what's the number? let me see if it fixes it for me21:32
jheskethclarkb:  cool, yep21:32
openstackgerritJulia Kreger proposed openstack-infra/shade: Add behavior to enable ironic noauth mode  https://review.openstack.org/15960921:32
needscoffeehttps://review.openstack.org/#/c/159315/21:32
pleia2mordred: amusingly (or something), we have this commented out in our current file: #export OS_SERVICE_NAME=cloudServersOpenStack21:33
mordredpleia2: :)21:33
mordredneedscoffee: yes. that fixes it for us21:33
needscoffeeCool21:33
mordredpleia2, clarkb ^^ new release of keystoneclient will also fix this21:34
clarkbpleia2: were you going to update those files as suggested by mordred?21:34
*** achanda has quit IRC21:34
needscoffeeI'll release it after lunch.21:34
jeblairjhesketh: also, can you take a look at 155160? :)21:34
mordredneedscoffee: awesome21:34
pleia2clarkb: since I think we won't get this change trickled down to us until The Future, I think I will go ahead and uncomment the OS_SERVICE_NAME for now21:34
anteayasdague: oh have you gone over the nova-net to neutron migration spec yet?21:35
anteayasdague: so you are all prepped for ops meetup?21:35
*** achanda has joined #openstack-infra21:35
clarkbpleia2: sounds good21:35
sdagueanteaya: I have not, monday, I promiss21:36
anteayaI'll keep bugging21:36
anteayaand thanks21:36
clarkbanteaya: where is that spec? in nova or neutron or openstack specs?21:36
anteayaneutron21:37
* anteaya digs21:37
anteayaclarkb: http://specs.openstack.org/openstack/neutron-specs/specs/kilo/migration-from-nova-net.html21:37
anteayasdague: ^^21:37
clarkbanteaya: thank you21:37
anteayathis is part 1 of a two parter21:37
*** abhirc has joined #openstack-infra21:37
anteayawelcome21:37
*** tnovacik has joined #openstack-infra21:37
anteayathe second part, with implementation details won't get merged until L now21:38
anteayathis just sets audience expectations21:38
sdagueanteaya: yeh, this week I was actually making progress on some code, so I punted on my non code responsibilities.21:38
anteayaI understand21:38
anteayaI have mars in my 12th house right now and have been punting on quite a bit21:38
anteayacomeon mars in taurus21:38
clarkbfirst thng that jumps out to me is that dvr is not ocmpatible with nova net multihost and that this won't be supported21:39
*** otter768 has joined #openstack-infra21:39
clarkbbut most people run nova net multihost aiui21:39
clarkbso this is like saying we have a migration path for only the few people that improperly deployed nova net21:39
anteayathat has been a position that has been raised21:40
anteayaalso cells are not supported21:40
jheskethclarkb: reviewed along with the concurrent patch21:40
anteayaas you pointed out this is the third anniversary of trying to get anywhere with this work21:40
anteayaso to make any kind of progress we had to agree to the simplest of deployments21:41
clarkbanteaya: yup, its just if you prioritize work you pick the most common deployment whihc is multihost nova net21:41
anteayaand invite operators with other use cases to build on what is provided21:41
anteayawell I didn't pick or prioritze anything21:42
clarkbanyways I will comment on the multinode testing21:42
anteayait came from the group21:42
anteayasure and thanks21:42
jeblairfwiw, i think the prevailing sentiment with infra-cloud is that we will probably regularly destroy it and recreate it rather than attempt any kind of migration.21:42
anteayajeblair: yay21:42
jeblairthis is a path that is only available to people who do not run important things in their clouds21:42
openstackgerritMerged openstack-infra/project-config: Don't run gate-trove-functional-dsvm-mysql for Icehouse  https://review.openstack.org/15942321:42
clarkboh the spec is merged already?21:43
jeblairbut that's us, so there we have it.21:43
* clarkb isn't sure where to comment21:43
*** otter768 has quit IRC21:43
anteayathere is part two of the spec21:44
anteayain the neutron specs repo21:44
clarkbjeblair: it didn't come up yet but we probably need to figure out what we want to do for network nodes21:44
openstackgerritMerged openstack-infra/zuul: Remove job parameters from status.json  https://review.openstack.org/15516021:44
clarkbjeblair: I really like nova network multihost because its simple and distributed21:44
pleia2fwiw, for rax it's NOVA_SERVICE_NAME we wanted, not OS_SERVICE_NAME21:44
clarkbanteaya: kk will look thorugh the list a bit more21:44
openstackgerritMerged openstack-infra/project-config: Rename barbican dogtag devstack-gate job  https://review.openstack.org/15955021:45
*** abhirc has quit IRC21:45
*** claudiub has joined #openstack-infra21:45
mordredclarkb, anteaya: I knowI have no voice in this decision21:45
clarkbpleia2: can you comment on mordred's change with that info?21:45
anteayahttps://review.openstack.org/#/c/142456/21:45
mordredbut I think that a migration path is a crazy idea - and I think that nova holding neutron hostage on this point is crazy21:45
pleia2clarkb: yep21:45
mordredbecause I think zero people are going to use it21:45
clarkbanteaya: https://review.openstack.org/#/c/142456/ is that the second bit?21:45
mordredfor the record21:45
anteayayes21:45
mordredthere is NO WAY it will EVER be a good idea21:45
anteayaokay so I appreciate all your opinions21:45
mordredthat's all21:45
sdagueclarkb: so there was also the theory in this spec of "write down what's possible; provide tools on the stuff that can be automated; make it better over time"21:46
anteayaand for the record I don't even know what I'm doing21:46
clarkbmordred: thats a bit like saying "well you guys refused to do this thing that you were required to do for long enough that I don't care anymore"21:46
clarkbmordred: which is sad21:46
anteayawhich I am sure you all know21:46
mordredclarkb: no.21:46
anteayaI am just trying to hold meetings21:46
mordredthat's not what I thin it is at all21:46
anteayaand get people to talk to each other21:46
mordredI think it's that the two have copmletely differnt world views21:46
anteayaand find out where there is agreement21:46
anteayaand do something with that21:46
mordredit's like saying "write me a migration mysql to postgres"21:46
mordredwhat does that even mean?21:46
pleia2clarkb: ah, I think mordred's change is correct actually21:46
anteayaask me to name more than two network managers? I can't do that21:46
clarkbpleia2: oh good21:46
anteayabut i do know how to hold a meeting21:46
clarkbmordred: except that one of the requirements for neutron was to not do that aiui21:46
anteayaso I am doing that21:46
clarkbmordred: I remember way back in san diego21:47
mordredclarkb: yeah. shrug. life has happened since then21:47
mordredlet's move on21:47
*** tonytan4ever has quit IRC21:47
clarkbmordred: around that time frame a lot of the discussion was around lets make this not so terrible21:47
clarkbbecause the right thing to do is to have at least semi consistent world views21:47
mordredthere is literally one person running nova network21:47
mordred2 if you count piston21:47
mordredbut I don't really care about them21:47
clarkbwalmart and cern use it apparently21:47
mordredyes. cern21:47
*** Sukhdev has joined #openstack-infra21:47
mordredwalmart JUST deployed it21:47
sdaguewell if you use cells you use it, but that's just crazy for other reasons :)21:48
mordredso that's on them21:48
mordredno sympathy for walmart21:48
anteayacern shows up to meetings21:48
mordredI care about tim21:48
mordredand cern21:48
mordredand if it were me, I'd say "write a README for Tim"21:48
anteayacern is trying to support this process21:48
sdaguemordred: right, which is basically what this is21:48
sdagueplus some tools where they can fit21:48
anteayaI'm not surprised you hate it21:50
anteayaeveryone hates it21:50
pleia2clarkb, mordred going to approve 159615, hopefully not breaking all the things21:50
anteayabut this is the small piece that everyone hates the least21:50
sdagueyou know what I hate more, having 2 network stacks in openstack :)21:50
anteayasdague: there we go21:50
jeblairclarkb: yeah, before we do anything with infra-cloud (well before), we need to have stuff we talked about written up so we can review it; i think that will make things clear and we can revisit any decisions that were made too lightly21:50
*** markvoelker has quit IRC21:51
*** markvoelker has joined #openstack-infra21:51
jeblairso that's like step 2; step 1 is assign someone to do step 2.21:52
*** dmellado has quit IRC21:52
mkj_wkZZelle_: question about your review, if you are around... "The "interactive" part is out of change scope. It should be in a specific change." ← a falsifying test is hard to come by because the problem it solves is buried in the feature I'm implementing, so I'm not sure how to write a meaningful test against it to submit it separately21:52
mkj_wkbasically, it's handling test suite fragility that doesn't show up without making changes in the code per se21:53
ZZelle_mkj_wk, hi21:53
mkj_wkI wanted to pull it out but couldn't see how to falsify it21:54
mkj_wkand I hate commits without at least some falsifying tests21:54
*** radez is now known as radez_g0n321:54
mkj_wkat least, when I can help it21:54
ZZelle_mkj_wk, mixing 2 things in the same change is a bad idea (typically if we need to revert)21:54
*** e0ne has quit IRC21:54
ZZelle_mkj_wk, a specific change helps to understand each "thing" perimeter and allows to describe more the problem in its change21:55
mkj_wkwell, in this case, if you have to revert my patch, you might as well revert the interactive bit too :)21:55
*** markvoelker has quit IRC21:56
ZZelle_mkj_wk, fixing a bug does not imply new unittests but when nontrivial explanations for sure :)21:56
mkj_wkjust to be clear about intent, I'm a stickler for seprate commits for separate things myself21:56
ctlaughclarkb: Thank you -- I'll try that21:56
*** markvoelker has joined #openstack-infra21:57
mkj_wkZZelle_: sure, I had an explanation of why in a previous review :)21:57
*** mbacchi has quit IRC21:57
ZZelle_mkj_wk, such explanation should be in the commit message as looking at all reviews is slow21:59
mkj_wkZZelle_: wrt unit test; I thought that unit tests were in test_unit.py and story tests in test_git_review.py21:59
mkj_wkZZelle_: (right, just wasn't trying to bury it; I'll add it to the commit message)22:00
openstackgerritMerged openstack-infra/system-config: Pass in service name to nova constructor.  https://review.openstack.org/15961522:00
mkj_wkZZelle_: the reason that I didn't break up the story is that I would have to replicate almost all the setup from the first half to test the second half, and it wasn't obvious that was worth a 10-20 second penalty on each run22:01
ZZelle_mkj_wk, oups, indeed unittests are in test_unit but there is no story tests22:01
timrcWow the custom search for Kibana has a pretty neat default in the From  textbox: 0NaN-NaN-NaN NaN:NaN:NaN -- Almost makes me want to put "batman" in the To textbox.22:02
mkj_wkwell, I don't know what to call test_git_review.py — I'm flexible about test types22:02
ZZelle_test_git_review contains functional tests for testing the integration between git-review and the system and gerrit22:02
mkj_wkbut the ones in test_git_review.py take tens of seconds and test_unit are instantanous22:02
mkj_wk(more or less, at least the ones I looked at)22:02
mkj_wkis it worth 20 seconds to do all the initial branchy setup twice?22:03
mkj_wk(I figured everyone runs tox before submitting and it's already pretty slow...)22:03
ZZelle_test_git_review focuses on integration tests with system/gerrit, where test_unit focuses on unittests (without external calls)22:03
*** gordc has quit IRC22:04
*** yfried is now known as yfried|afk22:04
ZZelle_mkj_wk, to speed up things, you can run tox -epy27 git_review.tests.test_git_review.$CLASS.$METHOD22:04
ZZelle_mkj_wk, i don't understand your bug as query_reviews is called with the same information when testing git_review -d22:05
*** e0ne has joined #openstack-infra22:05
openstackgerritK Jonathan Harker proposed openstack-infra/nodepool: Reset the client object before creating the image  https://review.openstack.org/7494322:05
*** rwsu has quit IRC22:06
clarkbanteaya: left a note about multinode testing on the spec22:06
clarkbpleia2: how goes image building?22:06
pleia2clarkb: so I don't have comparable specs for the system we're using, how many vcpus and ram do we actually want?22:07
anteayaclarkb: thank you22:08
clarkbpleia2: earlier jeblair said between 2GB and 8GB of ram and multiple cpus22:08
pleia2thinking either performance1-2 or 1-4?22:08
pleia2yeah, multiple, 2 or 4?22:08
clarkbmaybe 4? it is doing git stuff which include compression22:09
clarkbprobably handy to have a few CPUs22:09
clarkbespecially when reading in several incoming backup streams22:09
clarkbjeblair: ^ how does that sound?22:09
*** grue_pm is now known as needsmoarcoffee22:10
openstackgerritTimothy R. Chavez proposed openstack-infra/project-config: Print instance UUID to log  https://review.openstack.org/15625222:11
timrcclarkb, Okay searched Kibana for a connection reset by peer error and chose a job upstream where the uuid might be useful -- really if this lands upstream it's going all over the place downstream :)22:12
timrc(as soon as I cherry-pick it)22:12
mkj_wkZZelle_: sorry, it's really not when calling git review -d; it's when calling git review *after* calling git-review -d22:12
*** FallenPegasus has quit IRC22:12
*** needscoffee is now known as needsmostcoffee22:12
ZZelle_mkj_wk, i mean the trouble comes from query_reviews?22:12
mkj_wkZZelle_: basically yes, git review -d gets the review, which has the branch, but it doesn't do anything with the branch information22:13
ZZelle_mkj_wk, and git review -d already calls query_reviews22:13
*** needsmostcoffee is now known as morganfainberg22:13
*** stevemar is now known as thinksmorganneed22:13
ZZelle_mkj_wk, so the trouble is not query_reviews but multiple calls/concurrency?22:13
*** abhirc has joined #openstack-infra22:13
mkj_wkoh, it only shows up when you have an *interactive* ssh command that runs within git review, after running a non-interactive ssh command from within git-review, and that only within the test suite22:13
mkj_wkit doesn't break in real live22:13
*** thinksmorganneed is now known as stevemar22:13
clarkbtimrc: thanks, we can expand it to more jobs upstream too if that helps reconsumption22:14
mkj_wkI know that git review -d calls query_reviews22:14
mkj_wkbut it throws away the branch information22:14
mkj_wkand then when you git commit --amend afterward, and resubmit, it will rebase against the wrong branch22:14
*** needsmoarcoffee is now known as grue_pm22:14
clarkbtimrc: if that ends up working well we likely want it on most of our jobs22:15
mkj_wk(lacking a .gitreview file, and managing tens of thousands of those without getting anything wrong is fragile, which is how I got here in the first place)22:15
*** claudiub has quit IRC22:16
*** cody-somerville has joined #openstack-infra22:17
*** mrmartin has quit IRC22:17
*** jedimike has quit IRC22:18
jeblairclarkb, pleia2: yeah 4vcpu sounds like a good guess22:19
pleia2jeblair: thanks, on it22:19
*** jgrimm is now known as zz_jgrimm22:20
*** erlon is now known as erlon_away22:21
*** vilobhmm has quit IRC22:22
openstackgerritClark Boylan proposed openstack-infra/project-config: Upload files concurrently to swift.  https://review.openstack.org/15678822:22
clarkbjhesketh: ^ how is that? I am going to write a followup change to start moving all the logging to python logging22:23
clarkbjhesketh: makes dealing with exceptions easier22:23
*** yamamoto has joined #openstack-infra22:24
ZZelle_mkj_wk, you don't need to define run_command_status_int(interactive, ...) just to rename **env in **kwargs and set interactive = kwargs.pop('interactive', True)22:24
*** yfried|afk is now known as yfried22:24
jeblairclarkb: what is max_file_count?22:25
clarkbjeblair: it comes from zuul's config (either ini or layout depending) but its a number of files to upload together at once?22:26
jeblair    The maximum number of separate files to allow22:26
jeblairThe max_file_count attribute must be included and indicates the22:27
jeblairmaximum number of files that can be uploaded with the form.22:27
jeblairfrom swift ^22:27
jeblairclarkb: so it seems like that's per each POST request22:27
jeblairclarkb: i think zuul defaults it to 10...22:28
*** yamamoto has quit IRC22:28
clarkbgotcha so we are overloading that there, we could just ignore that value22:28
jeblairclarkb: but the use of that to set the number of threads is arbitrary, yeah?22:28
clarkbyes its arbitrary22:28
pleia2ok, at the generate cert step for puppet, and instructions say "If you are launching a replacement server, you may skip the generate step and specify the name of an existing puppet cert (as long as the private key is on this host)." but I don't see certs for the backup servers when I go to `puppet cert list` them - clarkb you mentioned earlier that it's not puppet making the config changes?22:28
pleia2but we do have a backup_server.pp22:28
jeblairclarkb: so we could leave it unlimited and end up with 200 threads doing 200 POSTs at once, or keep it at 10, because, hey, that sounds reasonable22:29
clarkbjeblair: jhesketh actually tested several numbers of threads and 10 seems to e a reasonable number22:29
clarkbiirc 200 didn't make a difference22:29
*** rwsu has joined #openstack-infra22:30
clarkbpleia2: ya so puppet around this is tricky, we don't want the cert to be there generally to avoid puppet breaking the node22:30
jeblairclarkb: ok.  i'm fine with overloading it for this purpose.  maybe we can make that tidier later.  mostly wanted to understand22:30
clarkbjeblair: cool22:30
clarkbpleia2: I think what we may want to do is generate the cert now, build the host, then delete the cert22:30
clarkbjeblair: ^22:30
*** tnovacik has quit IRC22:31
*** vilobhmm has joined #openstack-infra22:32
jeblairclarkb, pleia2: yeah, this may be a little wonky too until we're done with the puppet apply transition.22:32
*** vilobhmm has quit IRC22:32
pleia2ok, thanks22:33
*** vilobhmm has joined #openstack-infra22:33
*** dkliban is now known as dkliban_afk22:33
ZZelle_clarkb, why not stopping puppet-agent using puppet cookbook?22:33
mkj_wkZZelle_: I didn't know that kwargs weren't considered sacred; I'll change to kwargs.pop22:34
mkj_wkZZelle_: that's much easier22:34
clarkbZZelle_: we already don't run a puppet agent daemon, instead we have ansible iterate our nodes and run puppet agent --test. The problem with that is ansible iterates using the puppet cert list :)22:34
ZZelle_clarkb, outch22:34
*** yfried is now known as yfried|afk22:34
*** jamesmcarthur has quit IRC22:35
mkj_wkZZelle_: refreshing my memory, it was because it was **env that I thought I shouldn't reuse it22:37
mkj_wkZZelle_: if it had been **kwargs I would have22:37
*** sdake__ has quit IRC22:38
ZZelle_mkj_wk, that's why you rename it as env commonly refers to an environment dict where kwargs is a dict :)22:38
clarkbthis reminds me to go back and review the ansible related changes22:38
mkj_wkZZelle_: OK, wilco22:38
*** gyee has joined #openstack-infra22:38
*** sdake_ has joined #openstack-infra22:38
clarkbactually I think I have reviewed all the ansible chnages \o/22:38
pleia2OSError: [Errno 2] No such file or directory: './../make_swap.sh'22:39
pleia2I never stop having fun22:39
*** MaxV has joined #openstack-infra22:39
* pleia2 looks into this one now22:39
clarkbpleia2: you need to run the launch node script from the launch/ dir22:39
pleia2derp22:40
mkj_wkZZelle_: I'm called to dinner, but I think I've got most of the changes, and I've slimmed down the test case, and commenting why it makes sense together, including renaming it22:40
ZZelle_mkj_wk, thx22:40
*** gyee has quit IRC22:40
*** camunoz has joined #openstack-infra22:40
*** mayurig has quit IRC22:40
*** MaxV_ has quit IRC22:41
jeblairclarkb, mordred: were you talking about this earlier?  i'm reading 158176 and don't know how we get the code that change removes22:41
mordredclarkb: you have!22:41
mordredjeblair: yeah - so - the sequence is this22:41
mordredcurrently, puppet installs a copy of that code in /etc/ansible/roles22:41
jeblairmordred: right,  the code removed in the next change (158177)22:42
*** gyee has joined #openstack-infra22:42
mordredthen, we land a change that stops installing code, but does not delete the dir - leaving the stale code in place22:42
jeblairmordred: what change is that?22:42
mordredone sec22:42
jheskethclarkb, jeblair: I didn't do any rigorous testing but there was no improvement from 10 to 20 or even 100 threads. I believe 10 is enough to saturate the bandwidth22:42
openstackgerritjose-idar proposed openstack-infra/project-config: Adding pep8 gate check to Opencafe  https://review.openstack.org/15964222:43
ctlaughclarkb: That did it... it's running now, building an image.  Thank you.22:43
jeblairjhesketh: cool, i think that's good enough :)22:43
clarkbctlaugh: no problem22:43
mordredjeblair: yeah - that's 15817722:43
mordredjeblair: it does not delete it - it just stops installing it22:43
*** hdd has quit IRC22:44
jeblairmordred: it's parent deletes it though22:44
mordredjeblair: then 158175 starts installing it again to the same location using ansible-galaxy22:44
openstackgerritElizabeth K. Joseph proposed openstack-infra/system-config: Add Wildfly Puppet module  https://review.openstack.org/14351222:44
*** tiswanso has quit IRC22:45
*** aysyd has quit IRC22:45
pleia2elizabeth@ci-backup-rs-ord:~$22:45
pleia2yay22:45
openstackgerritClark Boylan proposed openstack-infra/system-config: Also serve zuul status.json via HTTPS  https://review.openstack.org/15547422:45
clarkbrebased ^ for fungi22:45
clarkbpleia2: woot22:45
mordredjeblair: no - the parent moves the library code from /etc/ansible/library to be embedded in /etc/ansible/roles/puppet/library22:46
mordredjeblair: and removes the library22:46
mordredturns out, we don't need a global _library_ - just a global roles path into which libraries can be put22:46
anteayapleia2: nice work22:47
mordredjeblair: might be helpful if I wrote up the sequence with links to the changes for clarity?22:47
openstackgerritClark Boylan proposed openstack-infra/puppet-zuul: Option to allow HTTPS for the proxy  https://review.openstack.org/15547122:47
clarkband ^ that one as well22:47
*** cody-somerville has quit IRC22:47
jeblairmordred: 158176 has an ensure => absent for /etc/ansible/library, but you are suggesting we merge it before 158175 which will install the code that is currently there into /usr from the galaxy...22:48
jeblairmordred: i think that would be good22:48
clarkbpleia2: I think next step is going to be getting at least once cinder volume and LVMing it22:48
jeblairmordred: an etherpad for the writeup, that is22:48
pleia2new server is up at 23.253.166.59/2001:4801:7825:103:be76:4eff:fe11:205d time to commence migrating old backup server data? (I can't log in to the old one, a result of the disk being filled is my user is not functional)22:48
clarkbpleia2: also disabling puppet on that node by deleting its cert on the puppet master22:48
mordredjeblair: on it22:49
clarkbpleia2: if you want to work on those two things I can work to get you a user on the old host22:49
pleia2ok, looking at http://ci.openstack.org/sysadmin.html#cinder-volume-management22:49
pleia2and deleting the cert from puppetmaster now22:49
openstackgerritSteve Baker proposed openstack-infra/project-config: Elements for building a heat test image  https://review.openstack.org/15964522:49
pleia2s/delete/revoke22:50
stevebakerI'm more than happy to discuss anything about that change ^^22:50
*** matrohon has quit IRC22:51
pleia2mm "Note that the puppet master needs to be restarted after revoking certificates."22:51
clarkbpleia2: fun22:51
*** achanda has quit IRC22:51
clarkbpleia2: if you issue an apache restart on that host it iwll restart the master proceses22:52
pleia2goodie, shall I?22:52
clarkbyup shouldnt hurt anything22:52
jeblairstevebaker: i think the way to get an image onto tarballs.o.o is to have a job that runs in the post pipeline that builds the image and uploads it.22:52
jeblairstevebaker: (that's how we get tarballs there)22:52
pleia2k, all done22:53
*** andreykurilin_ has joined #openstack-infra22:53
stevebakerjeblair: the post-pipeline of the project-config repo?22:53
jeblairstevebaker: probably the elements should be in a heat-managed repo (heat itself, another heat-related repo, or maybe a new repo); and then it can be a post job on that repo that runs whenever those files change22:54
jeblairstevebaker: (if you put them it heat, you can still craft the job so that it only runs on updates to the element files)22:54
jeblairput then _in_ heat that is22:54
stevebakerinteresting, maybe heat-templates is where this can go then22:55
jeblairstevebaker: sounds reasonable22:55
stevebakerjeblair: can you point me at something which uploads to tarballs, only triggering on updates to certain files?22:56
jeblairclarkb, pleia2: i removed the jenkins-dev bup for space on current server22:56
jeblairstevebaker: will do22:57
mordredjeblair: https://etherpad.openstack.org/p/StandaloneAnsibleRoles22:57
stevebakerjeblair: cool, thanks22:57
*** cody-somerville has joined #openstack-infra22:57
morganfainbergmordred, pleia2, clarkb, about to release 1.2.0 of keystoneclient which will include that fix.22:57
morganfainbergfyi.22:57
pleia2jeblair: thanks22:57
jeblairstevebaker: http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/ironic-python-agent-jobs.yaml22:58
clarkbI almost have pleia2's user on that host manually created22:58
openstackgerritElizabeth K. Joseph proposed openstack-infra/system-config: Add changing to launch/ directory to a README  https://review.openstack.org/15964722:58
clarkbjust need to dlea with groups now22:58
jeblairstevebaker: that's the build/upload to tarballs part of it...22:58
jeblairstevebaker: that has its own repo though, so doesn't have file filtering for the job22:59
jeblairstevebaker: http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/layout.yaml#n247022:59
jeblairstevebaker: that's the job invocation for that...22:59
jeblairstevebaker: but there are other jobs that use the "files:" match, like this: http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/layout.yaml#n70122:59
clarkbpleia2: ok you should be able to login and sudo and things now23:00
jeblairstevebaker: reference for the zuul syntax: http://ci.openstack.org/zuul/zuul.html#jobs23:00
pleia2clarkb: logged in, thanks23:00
stevebakerjeblair: ok, I shall start digesting23:00
*** bkero has quit IRC23:01
*** bkero has joined #openstack-infra23:01
pleia2whoops, I have a meeting, will be back in a bit23:01
pleia2then I'll get to cinder, et al23:01
clarkbok23:02
clarkbthe old host was 500GB, I think if we start with a single 1TB LVM'd volume that is a reasonable place to begin?23:02
clarkbjeblair: ^ do you think we should just go ahead and make it much larger to aviid resizing things later?23:02
anteayaI'm going to do art tonight, so I am off to draw something things that noone else would want to look at23:02
*** achanda has joined #openstack-infra23:02
*** cnesa has joined #openstack-infra23:02
clarkbanteaya: enjoy23:02
jeblairclarkb: i think starting with 1tb is reasonable23:03
anteayaclarkb: thanks23:03
jeblairmordred: okay, i'm still stuck at step 1.23:04
jeblairmordred: "    Reorg the puppet role by putting the library file into the puppet role https://review.openstack.org/#/c/158176"23:04
jeblairmordred: that removes /etc/ansible/library.23:04
*** vilobhmm has quit IRC23:04
jeblairmordred: but i don't see that whatever is currently there will have ended up anywhere else by this point.23:05
*** vilobhmm has joined #openstack-infra23:05
openstackgerritK Jonathan Harker proposed openstack-infra/nodepool: Reset the client object after proxy timeouts  https://review.openstack.org/7494323:05
*** vilobhmm has quit IRC23:05
*** vilobhmm has joined #openstack-infra23:06
openstackgerritClark Boylan proposed openstack-infra/puppet-zuul: Option to allow HTTPS for the proxy  https://review.openstack.org/15547123:06
*** Ryan_Lane has quit IRC23:07
*** Ryan_Lane has joined #openstack-infra23:07
*** nelsnelson has quit IRC23:07
*** Sukhdev_ has joined #openstack-infra23:08
*** Ryan_Lane has quit IRC23:08
*** Sukhdev has quit IRC23:09
*** bknudson has quit IRC23:09
*** Ryan_Lane has joined #openstack-infra23:10
jeblairmordred: nm, i got it.23:12
morganfainbergpython-keystoneclient 1.2.0 should now be out in the wild (on pypi)23:13
jeblairmordred: i would prefer the file be called roles.yaml.23:14
*** signed8b_ has quit IRC23:15
*** andreykurilin_ has quit IRC23:15
*** andreykurilin_ has joined #openstack-infra23:15
openstackgerritJames E. Blair proposed openstack-infra/system-config: Rename roles.yml to roles.yaml  https://review.openstack.org/15966123:17
*** yamamoto has joined #openstack-infra23:17
clarkbjeblair: supposedly some parts of ansible care about that but not this one23:17
clarkbso should be fine to rename it23:17
jeblairclarkb: cool.  hopefully the rest of ansible can grow indifference as well23:18
dansmithclarkb: did the requirements dependency thing get fixed already?23:18
dansmithclarkb: https://review.openstack.org/#/c/159577/23:18
*** mattfarina has quit IRC23:18
dansmithclarkb: oh nevermind, that one isn't reported the same way, I see now23:18
clarkbok :)23:19
jeblairclarkb, mordred: what's the status of 154304?23:19
*** stevemar has quit IRC23:19
clarkbjeblair: I believe I need to abandon it, but was hoping mordred would chime in23:19
clarkbjeblair: iirc the way mordred wants to do thsi is have the ansible play query git directly and save that as an ansible var23:19
*** ihrachyshka has quit IRC23:19
clarkbjeblair: instead of figuring it out and passing through to ansible23:20
mordredjeblair: so - the reason I called it that is that their _docs_ say it matters23:21
mordredimperical testing shows taht it does not actually matter23:21
mordredbut I would hate to get bitten by someone "fixing a bug" - since we would not be able to complain since it was documented23:21
mordredotoh - I could ask upstream and perhaps fix the docs?23:21
jeblairmordred: perhaps we can fix the ansible documentation then?23:22
mordredjeblair: I'll look in to it23:22
jeblairyeah, that would be cool; it seems strange for them to want to be contrary widespread usage23:22
mordredjeblair: http://docs.ansible.com/galaxy.html#advanced-control-over-role-requirements-files23:22
mordredis where it says it23:22
david-lyleare there still devstack-precise nodes23:22
mordredit might just mean "needs to have a YAML extension"23:22
mordredas opposed to a txt extension23:22
mordredand tehy said that by saying .yml since that's their standard23:23
clarkbdavid-lyle: yes, until icehouse is EOL23:23
* mordred goes to read the source23:23
jeblairmordred: yeah, makes sense23:23
david-lyleclarkb: ok, thanks23:23
openstackgerritJJ Asghar proposed openstack-infra/project-config: Updated the Chef repo description  https://review.openstack.org/15966423:23
*** tiswanso has joined #openstack-infra23:24
*** dmellado has joined #openstack-infra23:24
mordredjeblair:         if role_file.endswith('.yaml') or role_file.endswith('.yml'):23:25
*** andreykurilin_ has quit IRC23:26
*** andreykurilin__ has joined #openstack-infra23:26
*** tiswanso has quit IRC23:26
*** tiswanso has joined #openstack-infra23:27
jeblairmordred: the code even says 'yaml' is preferred :)23:31
*** yamamoto has quit IRC23:31
* clarkb steps out for a bit23:32
mordredjeblair: :)23:32
*** mayurig has joined #openstack-infra23:33
*** yamamoto has joined #openstack-infra23:35
*** mriedem has quit IRC23:36
marunjeblair: what do you think of having adding the neutron functional job to devstack-gate's pipelines?23:37
marunjeblair: I've tested the change to override PROJECTS for the job, and it does provide a modest speedup, but I'm concerned that a d-g change could break us if we don't gate.23:38
marunjeblair: I ask only because the neutron tempest jobs are already part of those pipelines, and the functional job is likely to be more reliable than those.23:39
* SpamapS mumbles something something toml something23:39
jeblairSpamapS: you're now on the same probation mordred is23:40
*** otter768 has joined #openstack-infra23:40
mordredjeblair: no. he doesn't hate it enough for that to be binding - he might actually find it funny23:40
mordredSpamapS: I will kick ban you23:41
jeblairmarun: i think i might be okay with a nonvoting job23:41
jeblairspeaking of which...23:42
*** yamamoto has quit IRC23:42
jeblairsomeone was looking into how all those non-voting jobs became voting...23:42
marunjeblair: why non-voting?  We're not as trustworthy as tempest?23:42
marunjeblair: you mean the metajob bug?23:42
*** mfink_ has quit IRC23:43
*** amotoki has joined #openstack-infra23:43
jeblairmarun: no, because to maintain sanity, devstack-gate only has jobs voting on the master branch; for some reason (i believe an accident), stable branch jobs started voting, and now nothing can merge.23:43
SpamapSmordred: I actually hate it just as much, but I would find _your reaction_ funny enough to consider it.23:44
marunjeblair: oh no.  is this related to my change?23:44
jeblairmarun: i don't think so23:44
*** amotoki has quit IRC23:44
*** otter768 has quit IRC23:44
jeblairmarun: i don't want devstack-gate to break neutron, but i also don't want to have a voting job from every project on it because it will increase the chance of us hitting an unrelated breakage...23:45
jeblairmarun: i actually like what we (used to) have with the stable branches, which is that we have a lot of jobs, but they are not voting.  lots of failures there mean there is almost certainly an issue with a devstack change.  a few failures probably just means icehouse is broken.23:45
marunjeblair: fair enough. The main danger thing I wanted to avoid is d-g adding dependency on other repos for the wrapper portion of its execution.23:45
marunIf that happens we'll break, and hopefully it will be visible enough to encourage more cautious changes.23:46
jeblairmarun: like what other repos?23:46
jeblairmarun: like nova?23:46
jeblairmarun: something in the default PROJECTS?23:46
marunjeblair: anything in PROJECTS that the functional job excludes23:46
jeblairmarun: yeah, seems unlikely, but a good check.23:46
*** shashankhegde has quit IRC23:47
openstackgerritMaru Newby proposed openstack-infra/project-config: Limit target projects for the neutron func job  https://review.openstack.org/15595823:48
*** dmsimard is now known as dmsimard_away23:48
marunjeblair: it checks out - the workspace prep is much less now23:49
*** EmilienM is now known as EmilienM|afk23:49
jeblairmarun: i think the longer evolution of d-g has the default list getting very small or nonexistent, and then probably something like jobs say they need the "compute layer" or whatever23:50
marunjeblair: makes sense.  the matrix could drive that pretty easily23:51
jeblair++23:51
*** ddieterly has quit IRC23:53
*** chlong has joined #openstack-infra23:53
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email base plugin.  https://review.openstack.org/15194823:54
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email SMTP Client  https://review.openstack.org/14979223:54
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email Send Cron Worker  https://review.openstack.org/15141623:54
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email Working Directory Utility Method  https://review.openstack.org/15194723:54
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email Outbox  https://review.openstack.org/15141323:54
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Redirect Browsers from API to Client  https://review.openstack.org/15555123:54
*** dims_ has quit IRC23:55
*** MarkAtwood has quit IRC23:55
marunjeblair: I'm confused - is it possible to set a job non-voting just for one project?23:56
*** rlandy has quit IRC23:56
jeblairmarun: no, but a lot of the devstack-gate jobs have it templated so that they create '-nv' variants, then you can add that version to a project.23:56
marunjeblair: ah, ok.23:57
jeblairmarun: (eg, gate-tempest-something and gate-tempest-something-nv)23:57
jeblairmarun: (and there's already a metajob that sets '^.*-nv$' jobs to nonvoting)23:57
marunjeblair: I'm still a bit fuzzy on the templating stuff, but I'll figure it out.  thanks!23:57
*** salv-orlando has quit IRC23:57
jeblairmarun: (this oversight is addressed in my zuulv3 proposal: http://lists.openstack.org/pipermail/openstack-infra/2015-February/002471.html )23:58
marunjeblair: timely123:58
marun!23:58
openstackgerritDavid Lyle proposed openstack-infra/project-config: temporarily making selenium tests non-voting  https://review.openstack.org/15957423:58
*** andreykurilin__ has quit IRC23:59
« Wednesday, 2015-02-25 Index Friday, 2015-02-27 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!