| *** sacharya1 has joined #openstack-dev | 00:01 | |
| *** sacharya has quit IRC | 00:01 | |
| *** sacharya has joined #openstack-dev | 00:01 | |
| *** dtroyer_zzz is now known as dtroyer | 00:06 | |
| *** sacharya has quit IRC | 00:10 | |
| *** eglynn has quit IRC | 00:11 | |
| *** justinsb has joined #openstack-dev | 00:13 | |
| *** rkukura has quit IRC | 00:15 | |
| *** sacharya has joined #openstack-dev | 00:26 | |
| *** Ryan_Lane has joined #openstack-dev | 00:44 | |
| *** dtroyer is now known as dtroyer_zzz | 00:54 | |
| *** thingee has joined #openstack-dev | 01:06 | |
| *** martine has quit IRC | 01:13 | |
| *** agonellaa has quit IRC | 01:32 | |
| *** martine has joined #openstack-dev | 01:36 | |
| *** martine has quit IRC | 01:38 | |
| *** darkhorse has quit IRC | 01:54 | |
| *** bhuvan_ has quit IRC | 02:08 | |
| *** sniperd has joined #openstack-dev | 02:12 | |
| *** dachary1 has joined #openstack-dev | 02:13 | |
| *** dachary has quit IRC | 02:13 | |
| *** littleidea has joined #openstack-dev | 02:15 | |
| *** sdake has joined #openstack-dev | 02:20 | |
| *** ywu has quit IRC | 02:36 | |
| *** hugokuo has quit IRC | 02:37 | |
| *** sniperd_ has joined #openstack-dev | 02:39 | |
| *** sniperd has quit IRC | 02:39 | |
| *** dtroyer_zzz is now known as dtroyer | 02:39 | |
| *** johnpur has joined #openstack-dev | 02:44 | |
| *** ChanServ sets mode: +v johnpur | 02:44 | |
| *** bhuvan_ has joined #openstack-dev | 02:52 | |
| *** mdomsch has joined #openstack-dev | 02:57 | |
| *** Mandell has quit IRC | 03:10 | |
| *** Mandell has joined #openstack-dev | 03:17 | |
| *** hugokuo has joined #openstack-dev | 03:21 | |
| *** ncode has joined #openstack-dev | 03:37 | |
| *** ncode has joined #openstack-dev | 03:37 | |
| *** asalkeld has quit IRC | 03:38 | |
| *** ncode has quit IRC | 03:50 | |
| *** asalkeld has joined #openstack-dev | 03:52 | |
| *** sdake has quit IRC | 04:13 | |
| *** sandywalsh_ has quit IRC | 04:37 | |
| *** garyk has quit IRC | 04:48 | |
| *** troytoman is now known as troytoman-away | 04:50 | |
| *** littleidea has quit IRC | 04:51 | |
| *** sniperd_ has quit IRC | 04:54 | |
| *** rohitk has joined #openstack-dev | 05:05 | |
| *** nvez has quit IRC | 05:06 | |
| *** jakedahn_zz is now known as jakedahn | 05:24 | |
| *** Mandell_ has joined #openstack-dev | 05:28 | |
| *** Mandell has quit IRC | 05:28 | |
| *** rohitk has quit IRC | 05:28 | |
| *** dtroyer is now known as dtroyer_zzz | 05:28 | |
| *** rohitk has joined #openstack-dev | 05:46 | |
| *** Mandell_ has quit IRC | 05:48 | |
| *** Mandell has joined #openstack-dev | 05:52 | |
| *** jakedahn is now known as jakedahn_zz | 05:52 | |
| *** garyk has joined #openstack-dev | 05:59 | |
| *** rohitk has quit IRC | 06:00 | |
| *** hattwick has quit IRC | 06:03 | |
| *** aa__ has joined #openstack-dev | 06:06 | |
| *** andresambrois has quit IRC | 06:07 | |
| *** dachary1 has quit IRC | 06:13 | |
| *** shang_ has quit IRC | 06:15 | |
| *** davidha has quit IRC | 06:15 | |
| *** davidha has joined #openstack-dev | 06:18 | |
| *** GheAway is now known as GheRivero | 06:25 | |
| *** apevec has joined #openstack-dev | 06:39 | |
| *** davidha has quit IRC | 06:49 | |
| *** GheRivero has quit IRC | 06:51 | |
| *** Stackops-Jorge has joined #openstack-dev | 07:04 | |
| *** GheRivero has joined #openstack-dev | 07:06 | |
| *** journeeman has joined #openstack-dev | 07:10 | |
| *** azbarcea has joined #openstack-dev | 07:14 | |
| *** anderstj has quit IRC | 07:16 | |
| *** dachary has joined #openstack-dev | 07:22 | |
| *** bhuvan_ has quit IRC | 07:26 | |
| *** pmezard has joined #openstack-dev | 07:34 | |
| zykes- | hmm, why isn't openstack doing in the db code Base = declarative_base(cls=BaseModel) instead of doing multiple inheritance on models ? | 07:46 |
|---|---|---|
| *** danwent has quit IRC | 07:55 | |
| *** aa__ has quit IRC | 07:57 | |
| *** hattwick has joined #openstack-dev | 07:59 | |
| *** danwent has joined #openstack-dev | 07:59 | |
| *** davidha has joined #openstack-dev | 08:06 | |
| *** danwent has quit IRC | 08:08 | |
| *** lianhao has joined #openstack-dev | 08:13 | |
| *** lianhao has left #openstack-dev | 08:13 | |
| *** apevec has quit IRC | 08:16 | |
| journeeman | Could someone please paste the link to the Quantum API 2.0 draft? Couldn't find it. | 08:48 |
| *** davidha has quit IRC | 08:55 | |
| *** davidha has joined #openstack-dev | 09:04 | |
| *** Ryan_Lane has quit IRC | 09:18 | |
| zykes- | journeeman: you work on quantum ? | 09:38 |
| *** eglynn has joined #openstack-dev | 09:45 | |
| *** ipl31_ has quit IRC | 09:53 | |
| *** anniec72 has joined #openstack-dev | 10:08 | |
| *** rkukura has joined #openstack-dev | 10:21 | |
| *** joesavak has quit IRC | 10:22 | |
| *** lts has joined #openstack-dev | 10:26 | |
| dachary | sileht: are you around today ? | 10:30 |
| dachary | sileht: I was going to suggest that you talk about your work on the munin plugins to #openstack-infra . I don't know if they are using munin at all but they will probably be most interested anyway ;-) | 10:31 |
| *** mjfork has joined #openstack-dev | 10:32 | |
| journeeman | zykes-: Sorry for the late reply. Yes, just beginning. :) | 10:35 |
| zykes- | journeeman: care to look at a bug ? | 10:36 |
| dachary | bbl | 10:38 |
| journeeman | zykes-: Yes. Wish to fix this one (https://bugs.launchpad.net/quantum/+bug/842273). Seems like a good starting bug? | 10:39 |
| uvirtbot | Launchpad bug 842273 in quantum "make OVS plugin use net-ports" [Low,New] | 10:39 |
| zykes- | was more looking for https://bugs.launchpad.net/quantum/+bug/994471 ;) | 10:41 |
| uvirtbot | Launchpad bug 994471 in quantum "dnsmasq config is not updated upon instance creation" [Undecided,New] | 10:41 |
| *** joesavak has joined #openstack-dev | 10:42 | |
| journeeman | zykes-: Looking into it. | 10:43 |
| davidha | notmyname: testing 123 | 10:52 |
| davidha | what is the use/purpose of WSGIContext class in wsgi.py | 10:53 |
| davidha | ? | 10:53 |
| *** Mandell has quit IRC | 11:00 | |
| *** flaviamissi has quit IRC | 11:11 | |
| *** rkukura has quit IRC | 11:17 | |
| *** renier has joined #openstack-dev | 11:34 | |
| *** bsza has joined #openstack-dev | 11:49 | |
| *** journeeman has quit IRC | 11:52 | |
| *** ayoung_ has joined #openstack-dev | 11:52 | |
| *** salgado has joined #openstack-dev | 11:55 | |
| *** adalbas has joined #openstack-dev | 11:56 | |
| *** anniec72 has quit IRC | 11:58 | |
| *** davidha has quit IRC | 12:07 | |
| *** davidha has joined #openstack-dev | 12:08 | |
| *** markvoelker has joined #openstack-dev | 12:10 | |
| *** alaski has joined #openstack-dev | 12:10 | |
| *** rkukura has joined #openstack-dev | 12:17 | |
| *** sdake has joined #openstack-dev | 12:28 | |
| *** dtroyer_zzz is now known as dtroyer | 12:41 | |
| *** crobinso has joined #openstack-dev | 12:45 | |
| jaypipes | davidkranz: to verify, you do now have +2 ability on Tempest, right? | 12:55 |
| jaypipes | mtaylor, jeblair: ping | 12:56 |
| davidkranz | jaypipes: Yes, though I am not sure of the protocol of when to use +2 vs +1 with regard to our review policy. | 12:56 |
| *** Susanne-Balle has joined #openstack-dev | 12:56 | |
| *** davidha has quit IRC | 12:56 | |
| jaypipes | davidkranz: you should use +2 to indicate you approve of the patchset and if another core reviewer +2s, they can/should +1 Approve it to send it to the test pit | 12:57 |
| jaypipes | davidkranz: if you have some reservations but won't hold up the review for them, +1 is better... | 12:57 |
| jaypipes | davidkranz: +1 in the review, not the Approve :) | 12:57 |
| davidkranz | jaypipes: OK, got it. | 12:58 |
| *** tryggvil_ has joined #openstack-dev | 12:58 | |
| jaypipes | davidkranz: BTW, did you see rajalakshmi's email about the test for volume attachments? | 12:58 |
| davidkranz | Yes. I was going to recommend the ssh approach. I don't see what else we can do for cases where the effect of an API call depends on state internal to the instance. | 12:59 |
| *** gakott has joined #openstack-dev | 12:59 | |
| jaypipes | davidkranz: right. it's a tough one. I suppose the only other option would be to put something in the config [compute] section that describes the volume setup for the image_ref and image_ref_alt images? | 13:00 |
| *** flaviamissi has joined #openstack-dev | 13:00 | |
| *** garyk has quit IRC | 13:00 | |
| davidkranz | jaypipes: We could do that but we don't want to have to do lots of tempest configuration for each test target. I really think it is in the spirit of black-box tests to freely ssh into the instance to check stuff. | 13:01 |
| davidkranz | jaypipes: THis stuff is not too hard if we can accept that,. | 13:02 |
| davidkranz | jaypipes: This test should really ssh in after the call as well to check that the device appeared. | 13:02 |
| jaypipes | davidkranz: sure, agreed. just floating the idea :) | 13:02 |
| jaypipes | speaking of which, is daryl's SSH code ready yet? | 13:03 |
| jaypipes | davidkranz: if you get a chance: https://review.openstack.org/#/c/6946/ | 13:03 |
| jaypipes | davidkranz: it's fattarsi's admin identity tests | 13:03 |
| davidkranz | jaypipes: I think the ssh is OK in theory but we are waiting for Daryl's final submission. | 13:04 |
| jaypipes | k | 13:04 |
| *** jsavak has joined #openstack-dev | 13:04 | |
| davidkranz | jaypipes: I'll look at the admin identity submission soon. | 13:05 |
| jaypipes | davidkranz: you want to get back to rajalakshmi or should I? | 13:05 |
| davidkranz | jaypipes: I can do it. | 13:05 |
| jaypipes | ok | 13:06 |
| *** dprince has joined #openstack-dev | 13:09 | |
| *** apevec has joined #openstack-dev | 13:12 | |
| *** alaski has quit IRC | 13:16 | |
| *** alaski has joined #openstack-dev | 13:17 | |
| *** nvez has joined #openstack-dev | 13:18 | |
| davidkranz | jaypipes: Regarding fattarsi's admin identity tests, they look OK but will not cleanup the created roles in some cases. | 13:20 |
| davidkranz | jaypipes: I am now convinced the clients need to handle this transparently unless told not to explicitely. | 13:21 |
| jaypipes | davidkranz: hmm, I think daryl's resource pool is probably better | 13:21 |
| *** blamar has joined #openstack-dev | 13:21 | |
| davidkranz | jaypipes: Do you think I should approve it anyway? | 13:21 |
| *** ayoung_ has quit IRC | 13:22 | |
| jaypipes | davidkranz: the resource pool is the concept I used in this merge prop, too, for the base test class(es): https://review.openstack.org/#/c/7069/2/tempest/test.py | 13:22 |
| davidkranz | jaypipes: Yes, I was talking about Daryl's code but we discussed on Friday the possibility of not making the client have to remember to add each thing. Too error prone. | 13:22 |
| jaypipes | davidkranz: I see what you mean... basically a "create hook" | 13:23 |
| jaypipes | davidkranz: that would auto-register the created resource for deletion. | 13:23 |
| davidkranz | jaypipes: Yes. | 13:23 |
| jaypipes | davidkranz: ++ | 13:23 |
| jaypipes | nice idea. | 13:23 |
| nvez | I've been looking around and not sure.. is it possible to have nova volumes on local storage (on the same hypervisor)? | 13:23 |
| jaypipes | davidkranz: unfortunately, that only works when we're using the Tempest rest client ;) | 13:24 |
| jaypipes | nvez: yup. | 13:24 |
| jaypipes | nvez: that's what devstack does by default. | 13:24 |
| jaypipes | nvez: grep through stack.sh for BACKING_FILE. | 13:24 |
| davidkranz | jaypipes: Right, but most people writing tempest tests will be using it. | 13:24 |
| nvez | good to know, thanks jaypipes -- I see, I will look into how devstack sets it up | 13:24 |
| jaypipes | nvez: you'll see the setup there... | 13:24 |
| jaypipes | davidkranz: agreed. | 13:24 |
| nvez | Sweet, thanks a lot, jaypipes. | 13:24 |
| jaypipes | nvez: no problemo. | 13:25 |
| nvez | Also, I think there's a small bug with DevStack | 13:25 |
| davidkranz | jaypipes: I'll send Darly an email and see where we stand on this and the ssh stuff. I will -1 this review for now. | 13:25 |
| jaypipes | nvez: making it easier for ya: https://github.com/openstack-dev/devstack/blob/master/stack.sh#L1457 | 13:25 |
| *** ayoung has quit IRC | 13:25 | |
| jaypipes | nvez: volume group is initialized there.. | 13:25 |
| nvez | Well, DevStack/keystone, PrettyTable 0.6 doesn't support "printt" and 0.5 does, for some reason, DevStack installed 0.6 and Keystone was not able to respond proprely in the CLI (because printt for tables wasnt working) | 13:26 |
| jaypipes | nvez: and the volume group is what nova-volume uses (the volume group is called nova-volumes by default) | 13:26 |
| *** markmcclain has joined #openstack-dev | 13:26 | |
| nvez | I presume that there should be somewhere to enforce the usage of 0.5 not 0.6, but I don't know python, but I'm going to read that .sh script now | 13:26 |
| jaypipes | nvez: sounds like a bug to me... | 13:26 |
| *** johnpur has quit IRC | 13:27 | |
| nvez | I think it installs PrettyPrint 0.6 when it runs the setup.py develop | 13:27 |
| nvez | but I couldn't find where the requirements are listed so I can try forcing to 0.5 or so | 13:27 |
| nvez | jaypipes: hmm, hosting a local iSCSI volume is the only way, no way to just directly have it assigned to it? | 13:28 |
| nvez | Or is the iscsi_helper only used to expose that volume elsewhere | 13:29 |
| jaypipes | nvez: are you asking if nova-volumes has the ability to set up a volume group on startup itself? | 13:30 |
| nvez | jaypipes: nope, but basically if you can have nova-volume use local storage for specific VMs (ex: create new instance backed by volume, instance is booted on hypervisor X which has local storage on it, the volume is also created on hypervisor X and the server always boots from hypervisor X (because the volume is on it) | 13:32 |
| jaypipes | nvez: sorry, I'm not quite following you... | 13:34 |
| dprince | unnamed glance reviewer: .... send me: https://review.openstack.org/#/c/7167/ | 13:34 |
| jaypipes | dprince: approved. | 13:34 |
| * dprince would like that to re-enable XenServer testing! | 13:34 | |
| dprince | jaypipes: THanks. And good morning! | 13:34 |
| jaypipes | dprince: morning! | 13:35 |
| *** sacharya1 has joined #openstack-dev | 13:38 | |
| *** sacharya has quit IRC | 13:38 | |
| *** mdomsch has quit IRC | 13:38 | |
| *** Lumiere has joined #openstack-dev | 13:42 | |
| *** GheRivero_ has joined #openstack-dev | 13:44 | |
| *** sandywalsh has joined #openstack-dev | 13:44 | |
| *** jgriffith has joined #openstack-dev | 13:45 | |
| *** kbringard has joined #openstack-dev | 13:46 | |
| *** sandywalsh_ has joined #openstack-dev | 13:51 | |
| *** sandywalsh has quit IRC | 13:51 | |
| *** tryggvil_ has quit IRC | 13:55 | |
| nvez | Another question, is multiple fixed IPs currently a possible thing with Nova (or is there somewhere someone working on that?) | 13:56 |
| nvez | And not just necessarily fixed IPs, but basically just a server with multiple IPs (but not NAT'd) | 13:56 |
| *** littleidea has joined #openstack-dev | 13:58 | |
| *** davidha has joined #openstack-dev | 14:00 | |
| *** edygarcia has joined #openstack-dev | 14:11 | |
| *** tong has joined #openstack-dev | 14:15 | |
| jaypipes | nvez: not sure about that one.... might want to ask vishy. | 14:17 |
| tong | hi, folks, any one has a script which restart devstack without destroy the databases? | 14:19 |
| *** dtroyer is now known as dtroyer_zzz | 14:20 | |
| nvez | tong: as far as I've read, devstack was meant to be torn down everytime... | 14:24 |
| tong | @nvez, that is right, | 14:24 |
| nvez | I was looking for that too but really after reading it was mentioned that devstack was really for dev/documentation and was not meant to stay in place | 14:24 |
| *** danwent has joined #openstack-dev | 14:25 | |
| tong | even for development, it will be still nice to have tenant id and image id, etc preserved so that when you create a request (via API), you do not have to search hard each time. | 14:25 |
| tong | I would like to have a script as a client to drive call into Nova Rest APIs. which requires tenant id to make up url and image id in the request. | 14:26 |
| tong | it is a pain to find it each time you run stack.sh. | 14:27 |
| *** pengyong has joined #openstack-dev | 14:27 | |
| nvez | tong: i believe that it does create a demo user for you, but I guess the fact that you just need to get the tenant/image id might be annoying | 14:28 |
| nvez | should be that much of a hassle though :P | 14:28 |
| pengyong | ping Johannes Erdfelt | 14:28 |
| tong | @nvez, yes, getting the tenant id and image id and put in the right place of a request is every annoying and tedious. would rather have it hard coded in the request either url or body if these things do not change from one run to the other. | 14:30 |
| nvez | Just a heads up, the DevStack bug with PrettyTable 0.6 is affecting people so I don't know if someone can update the constraints to force it to use 0.5 | 14:31 |
| zykes- | danwent: ping | 14:32 |
| danwent | zykes-: hey. on phone call, but can chat a bit | 14:33 |
| *** dtroyer_zzz is now known as dtroyer | 14:33 | |
| *** wdunn has joined #openstack-dev | 14:41 | |
| zykes- | danwent: you got a clue on https://bugs.launchpad.net/quantum/+bug/994471 ? | 14:42 |
| uvirtbot | Launchpad bug 994471 in quantum "dnsmasq config is not updated upon instance creation" [Undecided,New] | 14:42 |
| *** cmagina_ is now known as cmagina | 14:46 | |
| jerdfelt | it appears all of the devstack tests are broken | 14:47 |
| *** davidha has quit IRC | 14:47 | |
| jerdfelt | not a single devstack test has passed on any patch today | 14:47 |
| *** hub_cap has joined #openstack-dev | 14:48 | |
| nvez | jerdfelt: i think it has to do with keystone bug i reported | 14:51 |
| nvez | jerdfelt: https://launchpad.net/bugs/995976 -- it cant get the keystone id/etc | 14:51 |
| uvirtbot | Launchpad bug 995976 in keystone "PrettyTable 0.6 breaks Keystone CLI" [Undecided,New] | 14:51 |
| danwent | zykes-: ok will try to take a look today. | 14:52 |
| *** ayoung_ has joined #openstack-dev | 14:52 | |
| *** ayoung has joined #openstack-dev | 14:52 | |
| *** anderstj has joined #openstack-dev | 14:52 | |
| russellb | any devstack reviewers around? this patch has been around for a while ... https://review.openstack.org/#/c/6501/ | 14:55 |
| *** flaviamissi has quit IRC | 14:55 | |
| *** anderstj has quit IRC | 14:55 | |
| *** rnirmal has joined #openstack-dev | 14:55 | |
| zykes- | danwent: ping me if you need more info, i'll be around here during your "daytime" | 14:56 |
| danwent | zykes-: k | 14:56 |
| *** heckj has joined #openstack-dev | 15:00 | |
| jerdfelt | nvez: https://review.openstack.org/7198 | 15:11 |
| nvez | jerdfelt: i did one too :p | 15:14 |
| nvez | https://review.openstack.org/#/q/status:open+project:openstack/python-keystoneclient,n,z looks like we got 3, heh | 15:14 |
| nvez | i should do bug/# to have jerrit and all that shebang automate things? | 15:16 |
| russellb | heh, i think i see 4 patches to fix that bug so far today | 15:27 |
| nvez | tis pretty annoying :p | 15:28 |
| dtroyer | and it's affecting most of the clients | 15:28 |
| nvez | https://review.openstack.org/#/c/7188/ and https://review.openstack.org/#/c/7200/ both fix it | 15:31 |
| *** edygarcia_ has joined #openstack-dev | 15:31 | |
| nvez | dtroyer: might wanna fix yours to be backwards compatible as mentioned in the first one and maybe wed be able to get that merged | 15:31 |
| dtroyer | it's a trivial fix, I don't care which one gets used…using get_string() is backward compatible | 15:32 |
| *** edygarcia has quit IRC | 15:33 | |
| *** edygarcia_ is now known as edygarcia | 15:33 | |
| nvez | ah if get_string is backwards compatible then it should be ok if somemone just ok's it | 15:33 |
| *** Gordonz has joined #openstack-dev | 15:34 | |
| ayoung_ | heckj, I want to answer your question re my Keystone auth prefix...but it doesn't seem to have a comment link or button for me | 15:36 |
| *** thingee has quit IRC | 15:37 | |
| *** thingee has joined #openstack-dev | 15:37 | |
| ayoung_ | Do I have to do a review to comment? | 15:37 |
| ayoung_ | On my own commit? | 15:37 |
| *** jgriffith has quit IRC | 15:38 | |
| heckj | ayoung: yeah, I think you do... | 15:39 |
| heckj | ayoung_: ^^ | 15:39 |
| ayoung_ | heckj, OK. | 15:39 |
| *** anniec72 has joined #openstack-dev | 15:39 | |
| ayoung_ | doing so now | 15:39 |
| heckj | ayoung_: kinda akward, I know… | 15:39 |
| ayoung_ | heckj, so long as it is it the software that is messed up, and not me, I am OK | 15:40 |
| dprince | russellb: I see you've been making use of SmokeStack to test some of your RPC branches.... | 15:40 |
| dprince | russellb: thats great. One thing about the Tempest test suite thought.... | 15:40 |
| dprince | russellb: The volume tests fail because... well, we aren't running the volume service on Fedora. | 15:41 |
| *** alaski has quit IRC | 15:41 | |
| dprince | russellb: I *think* we can turn it on again. There was an issue with tgtm but I think that has been resolved in packaging. | 15:42 |
| *** ayoung has quit IRC | 15:42 | |
| *** ayoung_ is now known as ayoung\ | 15:42 | |
| *** ayoung\ is now known as ayoung | 15:42 | |
| russellb | dprince: gotcha ... i thought it was my fault since fedora 16 libvirt was passing on other branches | 15:42 |
| *** aa__ has joined #openstack-dev | 15:42 | |
| dprince | russellb: Torpedo w/ run fine. As with the Nova Smoke tests. | 15:42 |
| russellb | oh, so it's just my test config | 15:43 |
| *** alaski has joined #openstack-dev | 15:43 | |
| dprince | russellb: A couple weeks ago some of the Volume tests got moved to the 'smoke' category (which is what SmokeStack uses). That happens on any of the branches. So you can probably ignore that failure. | 15:44 |
| dprince | russellb: For Tempest that is. Which you were running on your branch. | 15:44 |
| *** ewindisch has joined #openstack-dev | 15:44 | |
| russellb | right. cool, thanks! it caught a real failure in my code before that. smokestack <3 | 15:45 |
| ayoung | heckj, apevec, does https://review.openstack.org/#/c/7156/ answer your questions? | 15:45 |
| dprince | russellb: great. Glad to see you find it useful. I usually look over any failures so I figured I'd explain the Tempest failure. I'll try to address it soon enough. | 15:46 |
| *** dwalleck has joined #openstack-dev | 15:47 | |
| *** mikal has quit IRC | 15:47 | |
| russellb | dprince: i appreciate it, because i did think it was my fault, so you saved me some time | 15:47 |
| russellb | wasn't sure how it was my fault, but i usually assume it's my fault :) | 15:47 |
| *** mikal has joined #openstack-dev | 15:49 | |
| dtroyer | patches for prettytable 0.6 are up for keystoneclient, glanceclient and novaclient | 15:50 |
| *** tryggvil_ has joined #openstack-dev | 15:50 | |
| dtroyer | heckj: https://review.openstack.org/7200 | 15:50 |
| dtroyer | bcwaldon: https://review.openstack.org/7201 | 15:50 |
| dtroyer | vishy: https://review.openstack.org/7202 | 15:50 |
| dtroyer | jeblair, LinuxJedi or mtaylor are probably going to have to push them through by hand because all three are required to fix the problem | 15:51 |
| *** alaski has quit IRC | 15:51 | |
| *** alaski has joined #openstack-dev | 15:52 | |
| heckj | dtroyer: just +2'd | 15:52 |
| *** markvoelker has quit IRC | 15:53 | |
| bcwaldon | dtroyer: +2'd | 15:53 |
| dtroyer | thx | 15:54 |
| *** anniec72 has quit IRC | 15:55 | |
| Vek_ | mtaylor / jeblair: Seeing a lot of consistent failures on approvals today... | 15:57 |
| bcwaldon | dtroyer: you could lock down prettytable to a previous version, no? | 15:58 |
| bcwaldon | dtroyer: or are you making non-backwards compatabile changes? | 15:58 |
| *** galstrom has joined #openstack-dev | 15:58 | |
| dtroyer | bcwaldon: the changes are backward compatible. print was just a wrapper around geT_string | 15:58 |
| bcwaldon | dtroyer: so you could fix gating by pegging everything at the prev PT version | 15:59 |
| bcwaldon | dtroyer: which blows, but it would work | 15:59 |
| bcwaldon | and I guess this is a good example of why we should ride latest versions of python libraries | 15:59 |
| dtroyer | right, but I believe the pip modules are pre-loaded into the image daily…not sure if I can override that | 16:00 |
| bcwaldon | dtroyer: you are correct | 16:00 |
| *** salgado has quit IRC | 16:01 | |
| *** salgado has joined #openstack-dev | 16:01 | |
| zykes- | danwent: gotten a change? | 16:02 |
| zykes- | s/change/change | 16:02 |
| *** Gordonz has quit IRC | 16:03 | |
| *** Gordonz has joined #openstack-dev | 16:03 | |
| *** asisin has joined #openstack-dev | 16:05 | |
| *** jemartin has joined #openstack-dev | 16:06 | |
| apevec | ayoung, I didn't get it, isn't auth_uri == auth_host:auth_port in auth_token middleware? | 16:07 |
| apevec | and it's always admin endpoint afaict | 16:07 |
| *** dhellmann has joined #openstack-dev | 16:09 | |
| ayoung | apevec, nope | 16:13 |
| ayoung | auth_uri should be | 16:13 |
| ayoung | by defaul;t | 16:13 |
| ayoung | http://hostname:5000 | 16:13 |
| ayoung | that is what is returned in a 403\ | 16:14 |
| ayoung | and different from the uri the service is going to use to authenticate the token | 16:14 |
| ayoung | that URI is by default http://hostname:35757 | 16:14 |
| ayoung | If you look at the area around my changes, you can see that the only change is to inject the prefisx | 16:15 |
| ayoung | which, by default is blank | 16:15 |
| ayoung | so the default behavior doesn't change | 16:15 |
| apevec | but then the code is wrong, if auth_uri is not in config, it will construct default value from host:port ? | 16:15 |
| ayoung | to make things work for httpd | 16:15 |
| ayoung | apevec, no | 16:15 |
| ayoung | you can specify is explicitly | 16:15 |
| ayoung | I'll show what I use, 1 sec | 16:16 |
| ayoung | apevec, heckj http://fpaste.org/KtJ1/ | 16:16 |
| ayoung | that is what I used for glance-api-paste.ini | 16:16 |
| ayoung | the 443 is in the auth-uri is superfluous | 16:17 |
| *** sniperd has joined #openstack-dev | 16:17 | |
| apevec | ayoung, what do you mean no? https://github.com/openstack/keystone/blob/master/keystone/middleware/auth_token.py#L134 | 16:17 |
| ayoung | self.auth_uri = conf.get('auth_uri', default_auth_uri) | 16:17 |
| ayoung | it is configurable if set explicitly | 16:18 |
| apevec | yes | 16:18 |
| ayoung | ah..sorry misread | 16:18 |
| apevec | but if not, default is host:port | 16:18 |
| ayoung | yes..you are correct. | 16:18 |
| ayoung | and I think that is my changes don;t go through, that is a mistake | 16:18 |
| ayoung | If | 16:18 |
| apevec | anyway, I thought it would be better to specify uri and let computers compute host port, no? | 16:19 |
| ayoung | apevec, I was shooting for a "least changes" approach | 16:19 |
| apevec | that's also most flexible, you don't need to introduce yet another parameter for prefix | 16:19 |
| *** shang has joined #openstack-dev | 16:19 | |
| ayoung | the way we get the http connection we'd have to parse the URI | 16:19 |
| apevec | IMHO openstack has too many knobs, adding one more parameter is bad :) | 16:20 |
| ayoung | apevec, yeah, the simplest uri scheme would be to specify auth admin URI and auth main URI | 16:20 |
| ayoung | apevec, I wa trying to only change one thing at a time | 16:21 |
| ayoung | in this case, just allowing access to the admin URI inside HTTPD | 16:21 |
| ayoung | once that works, I am OK for deprecating the old values, parsing, etc | 16:21 |
| ayoung | I don't want to break devstack...yet | 16:22 |
| *** gakott has quit IRC | 16:22 | |
| ayoung | this change goes in, and by default changes nothing. Then I can HTTPDify devstack for keystone....and so on | 16:22 |
| ayoung | apevec, defaulting the auth_uri to port 35757 is a mistake. But I wouldn;t bother fixing that as is, I'd rather get HTTPD as the Keystone default | 16:24 |
| *** rods has joined #openstack-dev | 16:24 | |
| *** salgado is now known as salgado-lunch | 16:27 | |
| *** maplebed has joined #openstack-dev | 16:27 | |
| *** aa__ has quit IRC | 16:28 | |
| *** hitesh has joined #openstack-dev | 16:29 | |
| hitesh | Hello freinds | 16:30 |
| hitesh | I got this error | 16:30 |
| hitesh | + echo 'g-api did not start' | 16:30 |
| hitesh | g-api did not start | 16:30 |
| hitesh | + exit 1 | 16:30 |
| hitesh | on devstack | 16:30 |
| apevec | ayoung, ok, +1 for get this in as-is and refactor later, now you just need keystone-core buy-in :) | 16:30 |
| hitesh | please suggest some steps? | 16:30 |
| apevec | hitesh, see subject | 16:31 |
| hitesh | yes | 16:31 |
| hitesh | I knew, still recommend me some about it? | 16:32 |
| hitesh | I restart apache | 16:32 |
| *** timello has quit IRC | 16:32 | |
| *** hitesh_ has joined #openstack-dev | 16:32 | |
| *** hitesh has quit IRC | 16:33 | |
| *** zaitcev has joined #openstack-dev | 16:33 | |
| *** sniperd has quit IRC | 16:33 | |
| *** sniperd has joined #openstack-dev | 16:33 | |
| hitesh_ | echo 'g-api did not start' | 16:33 |
| hitesh_ | can anyone tell me what should I do? | 16:34 |
| jerdfelt | https://review.openstack.org/#/c/7200/ | 16:34 |
| jerdfelt | would be nice to get that keystoneclient fix merged so we can get the devstack failures fixed | 16:35 |
| *** littleidea has quit IRC | 16:35 | |
| *** vincentricci has joined #openstack-dev | 16:35 | |
| *** littleidea has joined #openstack-dev | 16:36 | |
| *** hitesh_ has quit IRC | 16:37 | |
| *** andrewsben_zz is now known as andrewsben | 16:37 | |
| dprince | jerdfelt: Its for this very reason I've always liked integration/functional tests running in "advisory mode"... | 16:38 |
| dprince | I mean its important not to break things across the components. But when the house of cards falls down its nice to be able to put it back together in a sane manner. | 16:40 |
| jerdfelt | dprince: yeah. i also understand why pip freeze defaults to == | 16:42 |
| *** sniperd has quit IRC | 16:43 | |
| *** sniperd has joined #openstack-dev | 16:44 | |
| dprince | jerdfelt: Well. I differ there too. I don't like using pip for integration/function tests. For unit tests runs a simple PIP_DOWNLOAD_CACHE seems to do the trick though. | 16:44 |
| *** markvoelker has joined #openstack-dev | 16:45 | |
| *** rafaduran has joined #openstack-dev | 16:47 | |
| *** apevec has quit IRC | 16:50 | |
| *** thingee has joined #openstack-dev | 16:50 | |
| *** jgriffith has joined #openstack-dev | 16:53 | |
| *** spiffxp has joined #openstack-dev | 16:57 | |
| *** CatKiller is now known as Guest25068 | 16:58 | |
| *** shang has quit IRC | 16:59 | |
| *** shang has joined #openstack-dev | 17:02 | |
| *** utlemming has joined #openstack-dev | 17:04 | |
| *** rbasak has joined #openstack-dev | 17:05 | |
| *** rbasak has joined #openstack-dev | 17:05 | |
| *** davidha_who_took has joined #openstack-dev | 17:05 | |
| *** milner has joined #openstack-dev | 17:08 | |
| *** davidha_who_took is now known as davidha | 17:09 | |
| *** jdurgin has joined #openstack-dev | 17:10 | |
| *** Guest25068 has quit IRC | 17:13 | |
| *** Mandell has joined #openstack-dev | 17:13 | |
| *** utlemming has quit IRC | 17:15 | |
| *** harlowja has joined #openstack-dev | 17:17 | |
| *** salgado-lunch is now known as salgado | 17:17 | |
| *** anderstj has joined #openstack-dev | 17:19 | |
| harlowja | sup dawgs | 17:20 |
| asisin | harlowja: wasssaa | 17:20 |
| harlowja | waasssuppp | 17:20 |
| harlowja | ha | 17:20 |
| fattarsi | jaypipes: ping | 17:21 |
| mtaylor | Vek_: looking | 17:25 |
| jeblair | mtaylor: what are you looking at? | 17:25 |
| mtaylor | jeblair: nothing yet - still reading scrollback | 17:27 |
| jeblair | dtroyer: i think we can push those changes through if the integration test is currently broken by external factors; but some of them don't seem to have adequate votes on code review | 17:27 |
| jeblair | mtaylor: ^ | 17:27 |
| jeblair | mtaylor: those changes are 7189-7201 | 17:27 |
| *** dachary has quit IRC | 17:29 | |
| jeblair | dtroyer: also, all of the caching that the devstack gate does is strictly for performance reasons -- you can always change versions of things in pip/test-requires and it will use that | 17:30 |
| *** alrs has joined #openstack-dev | 17:31 | |
| jeblair | dtroyer: so yes, changing the reqs in devstack to require the old version, then making the code changes to the new api, then upgrading the version would be a good way of solving this problem without asking us to override the tests | 17:31 |
| *** gakott has joined #openstack-dev | 17:32 | |
| *** timello has joined #openstack-dev | 17:32 | |
| *** reed has quit IRC | 17:33 | |
| *** GheRivero_ has quit IRC | 17:33 | |
| *** dwalleck has quit IRC | 17:33 | |
| *** aa has joined #openstack-dev | 17:37 | |
| harlowja | ack, what required dep is changing | 17:42 |
| davidha | notmyname: hey, saw your cleanup email. Did you guys have time to discuss the webserver frontend as well? | 17:42 |
| *** torgomatic has joined #openstack-dev | 17:44 | |
| *** alrs has quit IRC | 17:46 | |
| *** jog0 has joined #openstack-dev | 17:50 | |
| *** johnpostlethwait has joined #openstack-dev | 17:54 | |
| *** mdrnstm has joined #openstack-dev | 17:55 | |
| *** pvo is now known as pvo-away | 17:56 | |
| *** rbasak has quit IRC | 17:57 | |
| *** mdomsch has joined #openstack-dev | 17:58 | |
| *** dwalleck has joined #openstack-dev | 18:00 | |
| notmyname | davidha: there isn't anything preventing you from doing that now. you would have to write your own startup scripts (mod_wsgi wrappers), but it could be done. in the words of one of the other devs "there's no reason why you couldn't, but there's been no compelling reason why you should. if someone has a compelling reason, they are free to do it" | 18:02 |
| *** Ryan_Lane has joined #openstack-dev | 18:02 | |
| *** sniperd has quit IRC | 18:03 | |
| *** sniperd has joined #openstack-dev | 18:03 | |
| *** hub-cap has joined #openstack-dev | 18:03 | |
| *** armax has joined #openstack-dev | 18:03 | |
| *** armax has left #openstack-dev | 18:03 | |
| *** hub-cap has quit IRC | 18:04 | |
| *** hub-cap has joined #openstack-dev | 18:04 | |
| *** Stackops-Jorge has quit IRC | 18:05 | |
| bcwaldon | Vek_ or jk0: can you +2 this little guy https://review.openstack.org/#/c/7201/ | 18:05 |
| jk0 | sure | 18:05 |
| *** hub_cap has quit IRC | 18:06 | |
| *** hub-cap is now known as hub_cap | 18:06 | |
| *** jgriffith has quit IRC | 18:07 | |
| *** jgriffith has joined #openstack-dev | 18:09 | |
| *** maplebed has quit IRC | 18:10 | |
| *** utlemming has joined #openstack-dev | 18:11 | |
| *** littleidea has quit IRC | 18:12 | |
| *** jakedahn_zz is now known as jakedahn | 18:12 | |
| bcwaldon | jk0: thanks! | 18:16 |
| jk0 | you betcha~ | 18:17 |
| bcwaldon | jk0: this is an easy one, too https://review.openstack.org/#/c/6920/ | 18:17 |
| bcwaldon | jk0: I promise this is all I'll ask for today! | 18:17 |
| jk0 | done | 18:18 |
| *** maplebed has joined #openstack-dev | 18:18 | |
| *** sniperd has quit IRC | 18:19 | |
| *** sniperd has joined #openstack-dev | 18:19 | |
| *** jgriffith has quit IRC | 18:19 | |
| *** jgriffith has joined #openstack-dev | 18:21 | |
| *** jgriffith is now known as jgriffith_away | 18:21 | |
| *** crobinso has quit IRC | 18:23 | |
| *** reed has joined #openstack-dev | 18:23 | |
| *** camm has quit IRC | 18:25 | |
| dtroyer | jeblair: I just pushed up https://review.openstack.org/7205 for devstack that forces prettytable 0.5 | 18:28 |
| dtroyer | it worked in my test oneiric VM | 18:28 |
| *** timello has quit IRC | 18:30 | |
| harlowja | so going back to dependency management, this prettyable thing seems to make me think that we need a common list with versions... | 18:31 |
| *** vincentricci has quit IRC | 18:34 | |
| *** vincentricci has joined #openstack-dev | 18:34 | |
| *** rbasak has joined #openstack-dev | 18:36 | |
| *** rbasak has joined #openstack-dev | 18:36 | |
| zykes- | danwent: ping | 18:36 |
| *** novas0x2a|laptop has joined #openstack-dev | 18:36 | |
| Vek_ | jeblair / mtaylor: Have ya'll figured out what's wrong with the gate? | 18:39 |
| davidkranz | jaypipes, mtaylor: Looks like the tempest gate on stable/essex just failed with some infrastructure problem: https://jenkins.openstack.org/job/gate-tempest-devstack-vm/451/ | 18:42 |
| jaypipes | davidha: been failling for a while (as has the integration-tests-devstack-vm one). Infra guys are aware of it. | 18:43 |
| jaypipes | davidkranz: sorry... ^^ | 18:43 |
| jaypipes | davidkranz: seems to be a Keystone problem: http://paste.openstack.org/show/16000/ | 18:44 |
| *** timello has joined #openstack-dev | 18:44 | |
| *** openstackjenkins has quit IRC | 18:45 | |
| davidkranz | jaypipes: OK. I didn't realize we were doing this already! | 18:46 |
| *** openstackjenkins has joined #openstack-dev | 18:46 | |
| jaypipes | davidkranz: we weren't supposed to be :( | 18:46 |
| jaypipes | davidkranz: I never said to enable the gate for anything other than Tempest. and it's not enabled for Tempest but is apparently enabled for other core projects.. | 18:47 |
| davidkranz | jaypipes: Oops | 18:47 |
| *** mnewby has joined #openstack-dev | 18:47 | |
| jaypipes | davidkranz: yeah :) | 18:47 |
| *** mnewby has quit IRC | 18:47 | |
| *** mnewby has joined #openstack-dev | 18:48 | |
| *** reed has quit IRC | 18:49 | |
| *** mdrnstm has quit IRC | 18:50 | |
| *** rbasak has quit IRC | 18:50 | |
| Vek_ | ah, ok. And at least they're aware of it... | 18:50 |
| *** gabrielhurley has joined #openstack-dev | 18:50 | |
| *** mdrnstm has joined #openstack-dev | 18:52 | |
| *** dolphm has joined #openstack-dev | 19:01 | |
| *** milner has quit IRC | 19:08 | |
| *** rbasak has joined #openstack-dev | 19:12 | |
| *** littleidea has joined #openstack-dev | 19:13 | |
| *** reed has joined #openstack-dev | 19:13 | |
| *** rods has quit IRC | 19:19 | |
| *** rods has joined #openstack-dev | 19:19 | |
| *** dprince has quit IRC | 19:20 | |
| *** pengyong has quit IRC | 19:21 | |
| *** eglynn has quit IRC | 19:22 | |
| *** eglynn has joined #openstack-dev | 19:22 | |
| *** jgriffith_away has quit IRC | 19:24 | |
| *** torgomatic_ has joined #openstack-dev | 19:28 | |
| *** rgoodwin is now known as rgoodwin_away | 19:28 | |
| *** anderstj has quit IRC | 19:30 | |
| *** camm has joined #openstack-dev | 19:30 | |
| *** torgomatic has quit IRC | 19:30 | |
| *** torgomatic_ is now known as torgomatic | 19:30 | |
| *** shang has quit IRC | 19:31 | |
| *** rkukura has quit IRC | 19:31 | |
| *** rgoodwin_away is now known as rgoodwin | 19:33 | |
| ayoung | heckj, termie I'd like to change the default back end for tokens from KVS to SQL | 19:34 |
| ayoung | KVS doesn't survive a reboot | 19:34 |
| ayoung | and SQL is shared between multiple processes, which means it works for HTTPD, but KVS does not | 19:34 |
| jeblair | dtroyer: 19:23:07 error: Installed distribution prettytable 0.5 conflicts with requirement PrettyTable>=0.6 | 19:37 |
| jeblair | dtroyer: https://jenkins.openstack.org/job/gate-integration-tests-devstack-vm/3901/consoleFull | 19:37 |
| *** anderstj has joined #openstack-dev | 19:37 | |
| jeblair | dtroyer: looks like something has a pinned prettytable requirement? | 19:38 |
| dtroyer | wtf? who did that? | 19:38 |
| mtaylor | I guess we should get single-global-list implemented | 19:38 |
| mtaylor | single-global-dep-list | 19:38 |
| jeblair | mtaylor: +1 | 19:39 |
| jeblair | dtroyer: i haven't found it yet.. i wonder if it's a dep of a dep. | 19:39 |
| * mtaylor cries | 19:39 | |
| dtroyer | found it... | 19:40 |
| mtaylor | cmd2 | 19:40 |
| dtroyer | no, cliff | 19:41 |
| mtaylor | oh, right. yeah | 19:41 |
| dtroyer | dhellmann rev'ed it, I'll need to pin it back now too... | 19:41 |
| mtaylor | dtroyer: dependencies are great aren't they? | 19:41 |
| dhellmann | is the simple fix for me to change the cliff requirements? | 19:42 |
| dtroyer | for a while, yes | 19:42 |
| dhellmann | I set a version because I was afraid of breakage due to the API change | 19:43 |
| dhellmann | and I wanted the Python 3 support in 0.6 | 19:43 |
| dhellmann | ok, I'll do a point release of cliff with that taken out | 19:43 |
| dtroyer | it will not be an issue once all of the clients get the api fixes in | 19:43 |
| dhellmann | makes sense | 19:43 |
| dhellmann | let me go do that right now... | 19:43 |
| *** sniperd has quit IRC | 19:46 | |
| dhellmann | ok, try with 0.5.1 of cliff | 19:47 |
| harlowja | singe global list somewhere ftw! | 19:47 |
| mtaylor | harlowja: ++ | 19:48 |
| mtaylor | harlowja: https://bugs.launchpad.net/openstack-ci/+bug/995607 | 19:48 |
| uvirtbot | Launchpad bug 995607 in openstack-ci "implement a single global dependency list" [Medium,Triaged] | 19:49 |
| harlowja | maybe something like, https://github.com/yahoo/Openstack-Anvil/tree/master/conf/distros or something more generic, idk, but something, pwease, lol | 19:49 |
| harlowja | thx! | 19:50 |
| mtaylor | harlowja: interesting ... first pass though is just to get the mechanics of a single pip-requires file that's used by all of the projects | 19:50 |
| harlowja | yes, pips are a start, its not just pips though :-p | 19:50 |
| mtaylor | indeed - one step at a time though | 19:51 |
| harlowja | sure | 19:51 |
| *** jgriff has joined #openstack-dev | 19:51 | |
| harlowja | in anvil, i am doing stuff like the following, "WARNING: @anvil.packager : A request has come in for a 'potentially' newer version of python-pastedeploy v(1.5.0*), when v(1.5*) was previously installed! ", thats pretty useful for seeing version mismatches, possibly the same in devstack... | 19:54 |
| zykes- | danwent: was the info I added enough ? | 19:54 |
| *** anderstj has quit IRC | 19:55 | |
| *** sniperd has joined #openstack-dev | 19:55 | |
| *** rbasak has quit IRC | 19:57 | |
| *** dolphm has quit IRC | 20:02 | |
| bcwaldon | dtroyer: you saw this, right? https://jenkins.openstack.org/job/gate-integration-tests-devstack-vm/3901/console | 20:04 |
| bcwaldon | dtroyer: the conflict in prettytable versions | 20:04 |
| *** utlemming has quit IRC | 20:04 | |
| *** reed has quit IRC | 20:04 | |
| mtaylor | bcwaldon: yeah - we're working on it | 20:04 |
| bcwaldon | kk, good luck | 20:05 |
| *** milner has joined #openstack-dev | 20:05 | |
| novas0x2a|laptop | hm, the devstack gate seems to be asking for an oneiric image that doesn't exist? (https://jenkins.openstack.org/job/gate-tempest-devstack-vm/450/console) is that signal or noise? | 20:05 |
| mtaylor | novas0x2a|laptop: I believe that is "we're out of nodes" ... which is probably related to the fact that three of our four cloud providers have ceased to be usable | 20:07 |
| *** Ruetobas has quit IRC | 20:07 | |
| mtaylor | three of the four cloud environments/accounts | 20:07 |
| novas0x2a|laptop | ah, okay, wasn't sure if it was something i'd done :) | 20:07 |
| *** hub-cap has joined #openstack-dev | 20:08 | |
| *** hub_cap has quit IRC | 20:11 | |
| *** hub-cap is now known as hub_cap | 20:11 | |
| *** dwalleck has quit IRC | 20:13 | |
| *** Ruetobas has joined #openstack-dev | 20:15 | |
| *** anderstj has joined #openstack-dev | 20:22 | |
| *** gyee has joined #openstack-dev | 20:23 | |
| *** heckj has quit IRC | 20:25 | |
| *** jaypipes is now known as jaypipes-afk | 20:30 | |
| *** heckj has joined #openstack-dev | 20:35 | |
| *** heckj has quit IRC | 20:35 | |
| *** hub_cap has quit IRC | 20:39 | |
| creiht | mtaylor: looking at a completed review for an openstack project, is there a way that I can determine what actual commit that code made it in to master? | 20:40 |
| vishy | creiht: git log | grep <changeid> -B10 ? | 20:41 |
| creiht | vishy: yeah that finds the original commit, but not the jenkins merge of that | 20:43 |
| creiht | The commit id shows up on the review page | 20:43 |
| vishy | creiht: really? ah because jenkins is doing merge commits now? | 20:43 |
| creiht | for that commit | 20:43 |
| creiht | I guess so? | 20:43 |
| *** dwalleck has joined #openstack-dev | 20:43 | |
| vishy | creiht: oh so then you have it? | 20:43 |
| creiht | well for example | 20:43 |
| creiht | https://review.openstack.org/#/c/6511/ | 20:43 |
| creiht | What I'm trying to figure out is what actual commit that was merged into master | 20:44 |
| vishy | creiht: so it wasn't 407e16b863bac1dfbf4e954837009abf9c17f018 | 20:44 |
| mtaylor | creiht: yeah, not a great way - we have an open ticket to include that info on the change-merged event ... I could include that we should include that sha in the merged message | 20:45 |
| * creiht cries | 20:45 | |
| creiht | :) | 20:45 |
| mtaylor | creiht: bug 929723 | 20:45 |
| uvirtbot | Launchpad bug 929723 in openstack-ci "gerrit should include merge commit sha1 in change-merged events" [High,In progress] https://launchpad.net/bugs/929723 | 20:45 |
| vishy | creiht: that is bad | 20:45 |
| vishy | git log | grep "Convert Volume" -B6 | 20:45 |
| creiht | vishy: yeah that's his initial merge, but I want to know when jenkins merged into master | 20:45 |
| vishy | yeah it should put the change id in the merge commit too... | 20:46 |
| mtaylor | creiht: I believe jeblair has the code 95% done, actually | 20:46 |
| vishy | or not do merge commits | 20:46 |
| vishy | mtaylor: can you set jenkist to always rebase instead of doing merge commits? | 20:46 |
| creiht | vishy: ok that worked and thanks | 20:46 |
| mtaylor | vishy: well, we could have it not do merge commits and instead cherry-pick/rebase onto the top of master | 20:46 |
| vishy | mtaylor: i mean it is always one changeset anyway, so why do a merge commit at all? | 20:46 |
| creiht | but down the road, I may need to do things like this programatically | 20:46 |
| creiht | mtaylor: so I look forward to those improvements :) | 20:47 |
| vishy | creiht: I think the real answer is not doing merge commits | 20:47 |
| vishy | mtaylor: didn't it used to do that? Or does it only do it in some special situations? | 20:47 |
| mtaylor | vishy: the biggest problem there is if people have stuff that depends on the thing being merged... | 20:47 |
| creiht | I'll leve that discussion to the git masters :) | 20:47 |
| mtaylor | vishy: it only merges if it has to merge - it tries to fast-forward apply the patch if it can | 20:47 |
| vishy | mtaylor: if everything is rebased then depedent branches should be fine as well? | 20:48 |
| mtaylor | vishy: but we figured given the rate of changes of things, that having gerrit rebase/cherry-pick apply each patch might confuse people tracking the dependent patches (it's totally just a config switch) | 20:48 |
| mtaylor | vishy: potentially ... it's more things that people might be working on locally | 20:48 |
| vishy | mtaylor: or will they have to be updated sometimes? | 20:48 |
| vishy | mtaylor, considering we rebase everything during a git-review anyway | 20:49 |
| vishy | I'm not too worried about that | 20:49 |
| mtaylor | vishy: ok. let me get a POC up for you, just so that we can make sure it does what we want | 20:49 |
| creiht | vishy: thanks again for your help :) | 20:50 |
| vishy | mtaylor: cool | 20:50 |
| vishy | creiht: no worries | 20:50 |
| bcwaldon | harlowja: do you think you'll have a merge prop up for this bp soon? https://blueprints.launchpad.net/glance/+spec/import-dynamic-stores | 20:50 |
| mtaylor | creiht: and I'll include your usecase on the change-merged but as well | 20:50 |
| harlowja | ya | 20:50 |
| bcwaldon | harlowja: cool cool | 20:50 |
| creiht | mtaylor: awesome | 20:50 |
| harlowja | i'm working on getting it for essex first, the darn tests are dying right now cause of it :-p | 20:51 |
| creiht | mtaylor: I look forward to that :) | 20:51 |
| bcwaldon | harlowja: ok | 20:51 |
| bcwaldon | harlowja: why focus on essex first? | 20:51 |
| bcwaldon | harlowja: you need to get it into master first :) | 20:51 |
| harlowja | i know :-p | 20:51 |
| bcwaldon | harlowja: and I don't think thats something we should backport, as its a feature | 20:51 |
| harlowja | agreed, yahoo is using essex for getting something out, and i'd like to have this there, at least in our packages, so that i don't have to fork essex/glance | 20:52 |
| harlowja | but the patch should be similar for master, i just think tests have moved around | 20:52 |
| bcwaldon | harlowja: ok, understood | 20:52 |
| bcwaldon | harlowja: thanks for taking the lead on this, btw | 20:52 |
| harlowja | ha, np, its needed cause i don't want to maintain syncing a github glance fork :-p | 20:53 |
| bcwaldon | harlowja: yep, definitely understand that | 20:53 |
| jeblair | vishy, mtaylor: the fact that jenkins does as many merge commits as it does indicates that a switch to ff-only would be very disruptive to developers | 20:53 |
| *** sacharya1 has quit IRC | 20:53 | |
| zykes- | ff-only, what does that mean ? | 20:54 |
| vishy | jeblair: i think the discussion was to switch to rebase/cherry-pick. Is that the same thing? | 20:54 |
| jeblair | vishy: i'm just catching up. :) | 20:54 |
| mtaylor | jeblair: yeah, I was talking about cherry-pick rather than merge - ff-only would be totally unworkable | 20:54 |
| bcwaldon | guys, something just merged! | 20:55 |
| bcwaldon | woo! | 20:55 |
| mtaylor | bcwaldon: w00t! | 20:55 |
| jeblair | cherry-pick is undesirable too -- it means that every commit is _not_ the one submitted | 20:55 |
| jeblair | it really makes the git log look a _lot_ different than it does now. :( | 20:55 |
| jeblair | zykes-: 'fast-forward only' | 20:56 |
| vishy | bcwaldon: holy sh** | 20:56 |
| mtaylor | jeblair: does it make it different in a way that is actually undesirable? as in, is the jenkins merge info actually interesting to people? | 20:56 |
| bcwaldon | vishy: my thoughts exactly | 20:56 |
| *** GheRivero_ has joined #openstack-dev | 20:59 | |
| *** maplebed has quit IRC | 21:00 | |
| *** milner has quit IRC | 21:01 | |
| *** ywu has joined #openstack-dev | 21:02 | |
| *** utlemming has joined #openstack-dev | 21:02 | |
| mtaylor | vishy: https://review.openstack.org/#/c/7164/ | 21:02 |
| mtaylor | vishy: turns out we're not testing libvirt on our unittest nodes :( | 21:02 |
| mtaylor | vishy: that patch SHOULD fix that | 21:03 |
| vishy | mtaylor: but it depends on the pep8 branch which disables the hacking tests :( | 21:04 |
| mtaylor | vishy: it does? well that's silly | 21:05 |
| mtaylor | vishy: lemme fix that - the two are not actually dependent | 21:06 |
| mtaylor | vishy: (and let me re-address the pep8 branch too) | 21:06 |
| mtaylor | vishy: ok. https://review.openstack.org/#/c/7164/ fixed. | 21:07 |
| *** jgriff has quit IRC | 21:17 | |
| *** GheRivero_ has quit IRC | 21:19 | |
| *** johnpur has joined #openstack-dev | 21:21 | |
| *** ChanServ sets mode: +v johnpur | 21:21 | |
| *** kbringard has quit IRC | 21:23 | |
| zykes- | danwent: hints on my bug report ? | 21:27 |
| danwent | just replied asking if you were using melange | 21:28 |
| danwent | otherwise, not really sure. Its working fine on my setup, so I'd probably start trying to trace through your setup and see if the linux_net.py methods that write dnsmasq config are getting called, and if they aren't, figure out why. | 21:28 |
| *** markvoelker has quit IRC | 21:29 | |
| jeblair | mtaylor: the bug about change-merged was to include the sha1 of the merge commit in the change-merged event. that's a little different than what i believe creiht was asking about. | 21:29 |
| zykes- | what method specifically is that danwent ? | 21:29 |
| harlowja | bcwaldon: do u know if i can upload a essex patch to gerrit, just for commenting purposes, for this thingy before i move it over :-p | 21:29 |
| *** tong has quit IRC | 21:29 | |
| harlowja | bcwaldon: should be straightfoward i guess, just haven't done it | 21:29 |
| danwent | zykes-: get_dhcp_hosts() | 21:30 |
| *** kbringard has joined #openstack-dev | 21:30 | |
| bcwaldon | harlowja: yep, use 'git review stable/essex' | 21:30 |
| harlowja | k | 21:30 |
| harlowja | thx | 21:30 |
| danwent | zykes-: or get_dhcp_hosts_text() | 21:30 |
| harlowja | bcwaldon: couple tests left and will do that | 21:30 |
| bcwaldon | harlowja: awesome | 21:31 |
| danwent | zykes-: check out nova/network/quantum/manager.py, and start with enable_dhcp() | 21:32 |
| danwent | and update_dhcp() | 21:32 |
| *** azbarcea has quit IRC | 21:33 | |
| *** utlemming has quit IRC | 21:37 | |
| *** dolphm has joined #openstack-dev | 21:38 | |
| *** lts has quit IRC | 21:39 | |
| *** milner has joined #openstack-dev | 21:39 | |
| *** dwalleck has quit IRC | 21:40 | |
| *** utlemming has joined #openstack-dev | 21:41 | |
| jeblair | vishy: https://github.com/openstack/openstack-ci-puppet/commit/0d716c1fe1830b47a44541cd0db35248cb52cd87 | 21:42 |
| jeblair | vishy, mtaylor: there's an example of a cherry-picked commit from gerrit | 21:42 |
| *** Gordonz has quit IRC | 21:43 | |
| jeblair | vishy, mtaylor: note that when cherry-picking, gerrit ignores patch dependencies. | 21:43 |
| *** alaski has quit IRC | 21:44 | |
| *** ohnoimde1d has quit IRC | 21:45 | |
| *** reed has joined #openstack-dev | 21:46 | |
| *** bsza has quit IRC | 21:51 | |
| zykes- | hmmm, danwent I'm trying to add some print / log statements to the enable_dhcp method but it's not logged ? | 21:52 |
| *** sniperd has quit IRC | 21:52 | |
| danwent | zykes-: in your logs, I saw output from the following log statement: LOG.info("Using DHCP for network: %s" % network_ref['label']) within enable_dhcp in QuantumManager | 21:53 |
| danwent | if you use the same format, and you're modified the right code location, you should see it. | 21:54 |
| *** reed has quit IRC | 21:56 | |
| *** utlemming has quit IRC | 21:56 | |
| *** milner has quit IRC | 22:00 | |
| *** kbringard has quit IRC | 22:01 | |
| zykes- | Hmmm danwent I'm hitting this if subnet is None or subnet['cidr'] is None: | 22:01 |
| danwent | zykes-: k, in meeting right now. will look when i get out. | 22:02 |
| zykes- | danwent: long time left ? | 22:03 |
| *** gyee has quit IRC | 22:03 | |
| danwent | zykes-: just started :( | 22:03 |
| zykes- | doh ;p | 22:04 |
| zykes- | bed time here soon | 22:04 |
| *** milner has joined #openstack-dev | 22:04 | |
| *** milner has quit IRC | 22:04 | |
| *** milner has joined #openstack-dev | 22:04 | |
| *** sacharya has joined #openstack-dev | 22:05 | |
| *** rnirmal has quit IRC | 22:06 | |
| mnewby | mtaylor: git-review installable from pypi could use an update - it has a blocker bug. | 22:06 |
| mnewby | mtaylor: version from github is fine | 22:06 |
| *** jemartin_ has joined #openstack-dev | 22:06 | |
| *** maplebed has joined #openstack-dev | 22:08 | |
| *** maplebed has joined #openstack-dev | 22:08 | |
| *** utlemming has joined #openstack-dev | 22:08 | |
| danwent | zykes-: ok… do you know which of those two cases are true in your setup? | 22:09 |
| *** cmagina has quit IRC | 22:09 | |
| *** jemartin has quit IRC | 22:11 | |
| *** jemartin_ is now known as jemartin | 22:11 | |
| *** cmagina has joined #openstack-dev | 22:12 | |
| *** rbasak has joined #openstack-dev | 22:12 | |
| *** jsavak has quit IRC | 22:13 | |
| salgado | danwent, thanks for the review on https://review.openstack.org/#/c/7019/! I'm just wondering if I need to do anything else now or is it just a matter of waiting for it to be approved/committed? | 22:13 |
| *** cmagina_ has joined #openstack-dev | 22:17 | |
| zykes- | danwent http://pastebin.com/wNwXzt6m | 22:18 |
| *** cmagina has quit IRC | 22:18 | |
| *** dolphm has quit IRC | 22:18 | |
| danwent | salgado: we need one more core dev to review. that core dev will approve. | 22:19 |
| *** s0mik has joined #openstack-dev | 22:19 | |
| harlowja | bcwaldon: https://review.openstack.org/#/c/7216/ | 22:19 |
| harlowja | feedback welcome from all who are willing, ha | 22:19 |
| salgado | danwent, ok, will see if I find other things to work on in the meantime then :) | 22:20 |
| bcwaldon | harlowja: ok, will look it over | 22:20 |
| bcwaldon | harlowja: next time, try to uplaod it as a DRAFT | 22:20 |
| harlowja | ah, is that in the topic? | 22:20 |
| zykes- | So I'm not hit by the continue at least danwent for the ipv4 net | 22:20 |
| bcwaldon | harlowja: I dont know how to do it, but Gerrit will now let you do it | 22:21 |
| *** shang has joined #openstack-dev | 22:21 | |
| *** anderstj has quit IRC | 22:21 | |
| harlowja | kk, gave it a -1 for now anyway | 22:21 |
| danwent | zykes-: yeah, so perhaps that's not really the issue, as the dnsmasq config doesn't even have v4 addresses | 22:22 |
| mnewby | danwent, salgado: I'm going to -1, it needs a bit more work. | 22:22 |
| zykes- | going further down then danwent | 22:22 |
| harlowja | ah, "--draft" option, good to know | 22:23 |
| *** galstrom has quit IRC | 22:23 | |
| zykes- | danwent: what does get_dhcp_hosts_text do ? | 22:24 |
| danwent | zykes-: it generates the text that will go in the config file that seems to be empty for you | 22:25 |
| zykes- | uhuh, it is totally empty hosts = ... returns nada | 22:26 |
| *** dachary has joined #openstack-dev | 22:28 | |
| *** dachary has quit IRC | 22:28 | |
| *** dachary has joined #openstack-dev | 22:28 | |
| bcwaldon | harlowja: you've got comments! | 22:29 |
| bcwaldon | harlowja: looking good, btw | 22:29 |
| harlowja | bcwaldon: yippe, comments | 22:29 |
| *** thingee has quit IRC | 22:29 | |
| harlowja | i thought about the "get_store_location_class" method, but then i was like thats more breaking changes, haha | 22:30 |
| bcwaldon | harlowja: not necessarily, you can look for StoreLocation in the same module by default | 22:30 |
| bcwaldon | harlowja: define that default in the superclass | 22:30 |
| harlowja | bcwaldon: k | 22:30 |
| bcwaldon | harlowja: then new store classes can implement it however they wayt | 22:30 |
| bcwaldon | want* | 22:30 |
| harlowja | wfm | 22:31 |
| *** rbasak has quit IRC | 22:31 | |
| *** utlemming has quit IRC | 22:32 | |
| harlowja | bcwaldon: "Really, this is all a bit confusing...having a Location class and then a store-specific Location class", 150% agree, stinky code smell | 22:32 |
| bcwaldon | harlowja: yeah, it was waaay too hard for me to understand wtf was going on | 22:32 |
| bcwaldon | harlowja: some good refactoring is in order here | 22:33 |
| harlowja | bcwaldon: agreed | 22:33 |
| bcwaldon | harlowja: if you think you have a good grasp on it, feel free to file a bug | 22:33 |
| harlowja | bcwaldon: sure, can do that, not 100% sure on the implications of removing it although i sort of understand what its doing, tiny little proxy class really | 22:34 |
| *** sniperd has joined #openstack-dev | 22:34 | |
| bcwaldon | I don't really know either. | 22:35 |
| zykes- | hmmmms, danwent >> "ips = self.ipam.get_allocated_ips(context, subnet_id, project_id)" that yields no ips | 22:35 |
| harlowja | bcwaldon: for the "glance/tests/utils.py" i think there isn't a config file when tests are running, there afaik is a mock that simulates the config | 22:35 |
| bcwaldon | harlowja: can you not do it with the mock? | 22:36 |
| bcwaldon | harlowja: or did I miss something | 22:36 |
| harlowja | bcwaldon: those configs eventually end up in the mock, so i think its the same thing :) | 22:36 |
| danwent | zykes-: key, sounds like you've almost tracked it down. look in nova_ipam_lib.get_allocated_ips() | 22:36 |
| bcwaldon | harlowja: hmm, I guess you're right | 22:37 |
| zykes- | :p | 22:37 |
| danwent | zykes-: things don't go much deeper than that, so you're probably almost there | 22:37 |
| harlowja | either or | 22:37 |
| *** sniperd_ has joined #openstack-dev | 22:37 | |
| *** sniperd has quit IRC | 22:38 | |
| *** dtroyer is now known as dtroyer_zzz | 22:42 | |
| *** utlemming has joined #openstack-dev | 22:43 | |
| bcwaldon | harlowja: headed out for a few, feel free to email or comment on the review and I'll get back to you | 22:43 |
| harlowja | bcwaldon: will do | 22:43 |
| *** sniperd_ has quit IRC | 22:51 | |
| zykes- | danwent: what in there to look at ? | 22:53 |
| *** rbasak has joined #openstack-dev | 22:54 | |
| danwent | zykes-: presumably there actually are IPs allocated in the subnet, right? So I'd try to figure out why those IPs aren't being returned. Do you see allocated IPs when you run: mysql -uroot -p<your password> nova -e "select address, instance_id, allocated from fixed_ips limit 10"? | 22:54 |
| zykes- | if ip['network_id'] == network['id']: | 22:56 |
| zykes- | It ain't hitting anything there | 22:56 |
| *** rafaduran has quit IRC | 23:00 | |
| zykes- | I don't see why it fails danwent | 23:03 |
| *** jemartin has quit IRC | 23:05 | |
| *** jemartin has joined #openstack-dev | 23:05 | |
| zykes- | for some reason danwent none of the ip's are marked as allocated | 23:09 |
| *** jgriff has joined #openstack-dev | 23:09 | |
| zykes- | in that network that's having problems atm | 23:09 |
| *** rbasak has quit IRC | 23:12 | |
| *** cmagina_ has quit IRC | 23:12 | |
| *** dtroyer_zzz is now known as dtroyer | 23:14 | |
| *** edygarcia has quit IRC | 23:14 | |
| danwent | zykes-: k, then that's the problem. seems like empty dhcp hosts file is a symptom, not a cause. | 23:15 |
| *** pmezard has quit IRC | 23:15 | |
| zykes- | danwent: yeah | 23:15 |
| zykes- | any clue on why no instance ip is assigned ? | 23:15 |
| *** rbasak has joined #openstack-dev | 23:16 | |
| *** rbasak has joined #openstack-dev | 23:16 | |
| *** dtroyer is now known as dtroyer_zzz | 23:16 | |
| danwent | zykes-: checkout allocate_for_instance in QuantumManager. Is it possible that you're spinning up these VMs on a project that has no networks? And that there aren't any global (i.e., non-project specific) networks? | 23:17 |
| zykes- | danwent: I know there's a network assigned | 23:18 |
| danwent | zykes-: ok, so you're seeing entries in the virtual_interfaces tables? | 23:19 |
| danwent | tables -> table? | 23:19 |
| *** reed has joined #openstack-dev | 23:19 | |
| zykes- | quantum list_nets f8109734926d4b1c91d6e7b5d1cbd23a > gives me a net id and if i do select * from networks where project_id = "f8109734926d4b1c91d6e7b5d1cbd23a"; I get the network in nova | 23:21 |
| zykes- | ports you mean ? | 23:23 |
| danwent | zykes-: yes, but can you check the virtual_interfaces table? I just want to eliminate that. this is a nova table. | 23:23 |
| zykes- | does it store info in both? | 23:23 |
| zykes- | http://pastebin.com/iqURQTPU | 23:24 |
| zykes- | net nr 9 there are my 2 machines | 23:26 |
| danwent | so is this code using the patch from somik? I think that is busted, and it could lead to the oddness you are seeing. I am in the middle of working on another patch for that issue. | 23:28 |
| *** rgoodwin is now known as rgoodwin_away | 23:28 | |
| zykes- | the manager.py is yes | 23:29 |
| danwent | you should just be able to remove the code the patch added in 'deallocate_for_instance' (see: https://review.openstack.org/#/c/6949/4/nova/network/quantum/manager.py) | 23:29 |
| danwent | simply deleting the release_fixed_ip method in QuantumManager is sufficient for me. | 23:29 |
| *** rgoodwin_away is now known as rgoodwin | 23:30 | |
| zykes- | wait a minute | 23:30 |
| danwent | when using dhcp. I still need to do some testing for when dhcp is not in use though. | 23:30 |
| danwent | i need to run to an interview. | 23:30 |
| zykes- | darn :( | 23:31 |
| zykes- | and no btw, I was using the code from trunk now lately actually, I replaced the code from somik to test this stuff in case I had broken it | 23:32 |
| zykes- | anyways danwent I'm using the manager.py from trunk | 23:37 |
| zykes- | it doesn't seem to be setting allocated = 1 now either | 23:38 |
| zykes- | and it's not freeing addresses after I removed the release_fixed_ip method | 23:39 |
| zykes- | so back to square one :) | 23:39 |
| *** galstrom has joined #openstack-dev | 23:45 | |
| mtaylor | jeblair: yes, the bug is a little different ... BUT - I was hoping if we include the merge id in the event stream, then the jenkins message could display it | 23:45 |
| mtaylor | except - I just realized, that's a chicken-and-egg, so never mind | 23:45 |
| mtaylor | jeblair: it's a different bug, just potentially touches similar chunks of code? | 23:45 |
| *** galstrom has quit IRC | 23:45 | |
| *** galstrom has joined #openstack-dev | 23:45 | |
| *** vincentricci has quit IRC | 23:48 | |
| *** dtroyer_zzz is now known as dtroyer | 23:49 | |
| *** shang has quit IRC | 23:52 | |
| *** reed has quit IRC | 23:52 | |
| *** bhuvan_ has joined #openstack-dev | 23:53 | |
| *** galstrom has quit IRC | 23:54 | |
| *** utlemming has quit IRC | 23:58 | |
| *** dtroyer is now known as dtroyer_zzz | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!