Thursday, 2015-04-23

*** Marga_ has quit IRC00:11
*** Marga_ has joined #tripleo00:13
*** adrianopetrich has quit IRC00:17
*** MasterPiece has quit IRC00:18
openstackgerritgreghaynes proposed openstack/diskimage-builder: Split the debootstrap functions into an element  https://review.openstack.org/16665000:20
greghaynesSpamapS: wondering if you have any great ideas about https://review.openstack.org/#/c/166650/14/elements/debian/package-installs.yaml00:31
*** Marga_ has quit IRC00:31
greghaynesSpamapS: I was considering adding to package-installs a pass of python templating where it passes in an arch var00:31
greghaynesalthough well have the whole x86_64 vs amd64 issue00:32
openstackgerritMartin AndrĂ© proposed openstack/tripleo-image-elements: Add openstack-keystone to pkg-map  https://review.openstack.org/15968800:33
clarkbgreghaynes: what about IA64?00:33
clarkbgreghaynes: and arm64?00:33
clarkb>_>00:33
greghaynesnow, lets be serious00:33
openstackgerritMonty Taylor proposed openstack/diskimage-builder: Port centos-minimal to yum-minimal  https://review.openstack.org/16664100:34
greghayneswe do kind of support arm though00:37
greghaynesI believe it has been used00:37
*** adrianopetrich has joined #tripleo00:39
*** sdake has joined #tripleo00:44
openstackgerritgreghaynes proposed openstack/diskimage-builder: Support arch-specific package-installs  https://review.openstack.org/17656000:49
greghaynesmordred: ^00:50
*** sdake has quit IRC00:50
greghaynesyou dont have to use that, but its an option00:50
openstackgerritgreghaynes proposed openstack/diskimage-builder: Support arch-specific package-installs  https://review.openstack.org/17656000:51
mordredgreghaynes: I'm assuming ARCH will be amd64 on debian and x86_64 on redhat?00:57
greghaynesmordred: should always be amd64 I think01:00
greghaynes(dont I sound confident)01:01
mordredsweet01:01
SpamapSIt's one of those ??? bonkers things about Debian01:01
SpamapSthat they'd have arch-specific linux packages01:01
greghaynesmordred: yea, itll be amd64 everywhere01:01
greghaynesif not then its a bug01:01
SpamapSI brought it up once w/ some Debian/Ubuntu people and they were like "don't even ask man."01:02
greghayneshaha01:02
*** dsneddon has quit IRC01:04
mordredSpamapS: it has something to do with a floppy disk I'm sure01:04
*** dsneddon has joined #tripleo01:04
SpamapSmordred: and hurd01:04
SpamapSdefinitely hurd01:04
SpamapSActually word is hurd may finally be dropped.01:05
mordredSpamapS: hurd is important01:05
SpamapSIt missed squeeze and jessie.01:05
mordredSpamapS: what if infra started running all its servers on debian gnu/hurd01:05
SpamapStwo releases missed == nobody cares enough to fix RC == take a hike hosers01:05
SpamapSmordred: the entire infra team would instantaneously burst into flames01:06
greghaynesonly one way to prove that theory01:07
SpamapSyes01:07
SpamapSit's my birthday01:07
SpamapSdo it01:07
SpamapSTHIsssssssssssssss I command01:08
greghaynesnew priority effort!01:08
SpamapShttp://statici.behindthevoiceactors.com/behindthevoiceactors/_img/chars/char_1127.jpg01:08
greghaynesalso, happy birthday :)01:08
greghaynesfor your birthday, I have broken dib's ability to merge anything01:09
greghaynesenjoy01:09
SpamapSgreghaynes: excellent smithers01:10
*** eghobo has quit IRC01:11
openstackgerritMartin AndrĂ© proposed openstack/diskimage-builder: Fix init system detection on Docker images  https://review.openstack.org/15278501:17
*** wuhg has joined #tripleo01:21
*** sdake has joined #tripleo01:31
*** Marga_ has joined #tripleo01:49
*** david-lyle has quit IRC01:59
*** eghobo has joined #tripleo01:59
*** rhallisey has quit IRC02:18
*** tzumainn has quit IRC03:00
openstackgerritMerged openstack/diskimage-builder: Create the dracut directory if not existing already  https://review.openstack.org/17011403:46
*** Marga_ has quit IRC03:50
*** Marga_ has joined #tripleo03:50
*** eghobo has quit IRC03:51
openstackgerritMerged openstack/diskimage-builder: Use shorter temporary file names for kpartx  https://review.openstack.org/17576703:51
openstackgerritMerged openstack/diskimage-builder: Allow setting DIB_PIP_RETRIES  https://review.openstack.org/15758403:51
openstackgerritMerged openstack/diskimage-builder: Dont try to unmount if were not using tmpfs  https://review.openstack.org/16735803:55
openstackgerritMerged openstack/diskimage-builder: Generate the default en_US locale  https://review.openstack.org/17084603:55
openstackgerritEmilien Macchi proposed stackforge/puppet-tripleo: Bump rspec-puppet to 2.1.0  https://review.openstack.org/17661003:58
openstackgerritMerged openstack/diskimage-builder: Fix dhclient in Fedora ramdisks  https://review.openstack.org/15358103:58
openstackgerritMerged openstack/diskimage-builder: Add generic devuser element  https://review.openstack.org/15343903:58
*** yuanying has quit IRC04:08
*** spzala has quit IRC04:23
*** sdake has quit IRC04:41
*** sdake has joined #tripleo04:46
*** sdake_ has joined #tripleo04:49
*** sdake has quit IRC04:50
*** david-lyle has joined #tripleo04:59
*** sdake has joined #tripleo05:05
*** yuanying has joined #tripleo05:09
*** sdake_ has quit IRC05:10
*** jcoufal has joined #tripleo05:12
*** yuanying has quit IRC05:13
*** jcoufal has quit IRC05:15
*** rwsu has quit IRC05:21
*** eghobo has joined #tripleo05:21
*** jcoufal has joined #tripleo05:25
*** masco has joined #tripleo05:50
*** Goneri has joined #tripleo06:00
*** yuanying has joined #tripleo06:10
*** Goneri has quit IRC06:13
*** bvandenh has joined #tripleo06:13
*** yog__ has joined #tripleo06:14
*** yuanying has quit IRC06:14
*** bvandenh has quit IRC06:18
*** jtomasek has joined #tripleo06:24
spredzy|afkmorning all06:28
*** spredzy|afk is now known as spredzy06:28
*** Marga_ has quit IRC06:31
*** eghobo has quit IRC06:42
*** rdopiera has joined #tripleo06:43
*** jtomasek has quit IRC06:44
*** jprovazn has joined #tripleo06:45
dasmspredzy: morning07:02
*** ifarkas has joined #tripleo07:07
*** jcoufal_ has joined #tripleo07:07
*** akrivoka has joined #tripleo07:09
*** funzo has joined #tripleo07:11
*** jcoufal has quit IRC07:11
*** jcoufal_ has quit IRC07:15
*** jcoufal has joined #tripleo07:15
*** weshay has joined #tripleo07:17
*** aufi has joined #tripleo07:17
openstackgerritMerged openstack/tuskar-ui: Removed Less remains, updated MANIFEST.in  https://review.openstack.org/17596407:20
*** akrivoka has quit IRC07:25
*** Goneri has joined #tripleo07:26
*** dsneddon has quit IRC07:36
*** jistr has joined #tripleo07:38
openstackgerritJan Provaznik proposed openstack/tripleo-common: Initial Cookiecutter Commit.  https://review.openstack.org/17664907:39
*** yuanying has joined #tripleo07:39
*** jcoufal has quit IRC07:40
*** jcoufal has joined #tripleo07:41
*** ukalifon has joined #tripleo07:43
*** Goneri has quit IRC07:44
*** jcoufal has quit IRC07:45
*** morazi has joined #tripleo07:45
*** Goneri has joined #tripleo07:47
spredzyjistr, based on what we talked yesteray with dprince (About moving more logic to puppet-tripleo) and EmilienM (for having a clean wrapper), what would you think of an approach that looked like that -> https://etherpad.openstack.org/p/tripleo_pacemaker_pattern07:47
spredzy(hello by the way ;) )07:47
jistrmorning :)07:47
*** jcoufal has joined #tripleo07:54
*** akrivoka has joined #tripleo07:56
*** aufi_ has joined #tripleo07:56
*** ukalifon has quit IRC07:58
*** aufi has quit IRC08:00
jistrspredzy: the wrapping approach looks very good to me, but i'm not sure about the contents of the tripleo::pacemaker_service definition... not sure if we should be writing our own ocf files...08:02
jistrspredzy: we should follow the documented architecture, e.g. this one is for glance https://github.com/beekhof/osp-ha-deploy/blob/master/pcmk/glance.scenario08:03
*** gfidente has joined #tripleo08:03
*** gfidente has joined #tripleo08:03
spredzyjistr, this just suppose to provide a way for us to specify one if the packaging doesn't provide one. If we need to rely on systemd/already provided ocf we will prob. never use it08:04
spredzygfidente, o/08:04
* spredzy looking08:04
gfidentehi guys08:04
jistrspredzy: yeah i think we'll rely on provided resource agents all the time08:05
gfidenteis it a review you're talking about or general stuff?08:05
jistrso, i guess you won't like it, but we should probably take a lot of inspiration from here https://github.com/redhat-openstack/astapor/tree/master/puppet/modules/quickstack/manifests/pacemaker08:05
spredzygfidente,  https://etherpad.openstack.org/p/tripleo_pacemaker_pattern08:05
*** jtomasek has joined #tripleo08:06
spredzyjistr, why wouldnt i like it ;) I am open minded :D and not knowledgeable enough in pacemaker to argue about which direction to take08:06
jistrspredzy: e.g. because of this :) https://github.com/redhat-openstack/astapor/blob/master/puppet/modules/quickstack/manifests/pacemaker/resource/generic.pp#L55-L6808:07
spredzyjistr, ahaha. indeed I don't like that one :p (let me read a bit of context)08:07
*** ukalifon has joined #tripleo08:07
spredzyI like the wai + in case you didn't get it, *really* wait08:08
gfidenteI am missing some context I think08:08
gfidentewhat is the goal?08:08
jistrgfidente: spredzy suggested a way to extract bits into puppet-tripleo08:08
jistrgfidente: so that we can deal with both having pacemaker enabled and disabled, without getting crazy in our one overcloud_controller.pp file08:09
spredzygfidente, yesterday dprince and jistr were mentioning that we should start moving logic into puppet-tripleo else manifest will become too complex. which is true. So hence the puppet-tripleo bits. At our daily EmilienM suggested that we use a wrapper for the pacemaker bits so if we want to enable possible ocf file customization it make sense, if it is to have a simple wraper around pcmk_resource, I am not sure anymore08:10
* spredzy reads more into how quickstack does it to have a better overview08:10
jistrwell, quickstack has their own wrappers too https://github.com/redhat-openstack/astapor/tree/master/puppet/modules/quickstack/manifests/pacemaker/resource08:10
gfidenteyou have ownership of the puppet bits08:10
gfidentebut I am against it08:10
jistrbut i'm not sure if we need to replicate the approach totally08:11
gfidenteI think this is precisely what we did not want08:11
*** mcornea has joined #tripleo08:11
gfidentebut I might not understand entirely the pros and cons08:11
jistrgfidente: ok, why are you against though?08:11
gfidenteso I will adapt to whatever is the recommendation08:12
*** athomas has joined #tripleo08:12
gfidentejistr, because I thought initial goal was to not have another composition layer08:12
gfidenteand this looks like moving in that direction08:12
gfidenteBUT I am serious, I do not understand fully the pros and cons08:14
jistryeah it is sorta moving in that direction, i think that if we want to support multiple optional configs (e.g. with pacemaker vs. without pacemaker), then we'd better start extracting things into smaller files08:14
gfidenteexactly this is something I can't have a serious opinion about08:14
gfidenteso I trust the decision08:14
gfidentethis is the reason why yesterday I didn't get into conversation either08:14
gfidenteit's just something I don't feel like I can add much08:14
jistrgfidente: do you think having a composition layer is a bad thing for some reason? It could be that we're just overlooking something.08:15
gfidentejistr, I thought we wanted to move complexity into the modules, if generally applicable to other circumstances, or the orchestration layer08:16
gfidentewhere the .pp shipping in t-h-t were just 'include' or minor hacks08:17
gfidentes/complexity/logic/08:17
jistrgfidente: btw to add more context -- the discussion yesterday was about making puppet modules more hackable in the tripleo context. We'd like to have the *option* to upload puppet modules via rsync when the stack is being created. This would allow us to hack on puppet modules without the need to rebuild images every time we make a change.08:17
*** lucasagomes has joined #tripleo08:17
jistrgfidente: yeah i agree, move as much stuff into the modules where the stuff belongs08:17
gfidentejistr, but I don't want you to waste time explaining me things really08:18
*** pblaho__ is now known as pblaho08:18
gfidenteand I am not kidding, you know puppet better08:18
gfidenteso I am perfectly fine adpating to whatever choice you make08:18
*** ukalifon has quit IRC08:18
jistrgfidente: it's not a waste of time ;) you make good points and i think it's important that we all have similar views in terms of where things are going08:19
spredzyInitially I had in mind puppet-tripleo as a kind of library for t-h-t with helpers that complexity would make overcloud_*.pp unreadable08:19
gfidentespredzy, yeah this was my understanding as well08:19
jistrgfidente: yeah i agree with spredzy ^08:19
openstackgerritJan Provaznik proposed openstack/tripleo-common: Initial Cookiecutter Commit.  https://review.openstack.org/17664908:20
jistrgfidente: well that's what we're doing then...08:20
gfidenteokay so maybe talking is making things look different than what they really are08:20
spredzyand then some overcloud_rabbit.pp that would get concat with the main overcloud_controller (I know you dont like that jistr :)08:20
jistrthere's a thin line between "library for helpers" and "a composition layer" imo08:20
gfidenteyeah that is it probably08:20
gfidentehow about we try with an actual code change?08:20
spredzybut my main concern, is the future, if puppet-tripleo might end up like a full blown project like astapor / puppet-openstack-cloud / etc...08:20
spredzya project on its own08:20
gfidentesomething which can easily be added, like keystone08:21
gfidentesomeone wants to try it out and should how it could look with a code change?08:21
gfidente*show08:22
spredzyagain I'll go with wathever we decide. But this is my main concern. I understand that if not having a composition layer add enormous complexity in the logic it would be beter to go with one08:22
jistrthe practical issue i see is that we'll probably need some code (repeatable bits, which we can include from various other spots) to deal with the "we want to support deployments both with and without pacemkaker"08:22
jistrthat's not a thing we can push to the puppet-nova etc. modules i'd say08:23
spredzyjistr, correct and I agree that those bits if generic enough should be in puppet-tripleo - no question ask on that08:23
spredzywhere I am more 'incline' is to move rabbitmq whole management into puppet-tripleo08:24
gfidentejistr, I was faced as well with the terrible monster that is fencing08:24
gfidentejistr, more on this next week, I think08:24
jistrgfidente: re practical code example, i think what spredzy suggested here looks good https://etherpad.openstack.org/p/tripleo_pacemaker_pattern08:24
jistr(but as i wrote above, i think we'll not write our own ocf files probably)08:25
gfidentejistr, oh I have to study that to understand it :)08:25
gfidenteI was thinking to use ocf only for galera/rabbit08:25
jistrgfidente: and yeah fencig will be fun :D08:25
spredzyjistr, if so a wrapper is not needed08:26
spredzypcmk_resource will be enough08:26
spredzythe wrapper in openstack_extras was made to handle custom ocf files08:26
spredzywhat do you think ? jistr08:26
gfidentespredzy, openstack_extra uses pcs?08:26
jistrspredzy, gfidente: re galera https://github.com/beekhof/osp-ha-deploy/blob/master/pcmk/galera.scenario#L97-L16708:27
spredzygfidente, no https://github.com/stackforge/puppet-openstack_extras/blob/master/manifests/pacemaker/service.pp#L147-L17308:27
spredzybut this could be replaced with a pcmk_resource08:27
spredzyclone is part of the type itself08:27
spredzyjistr, not sure I understand the link between the wrapper thing and the scenario chunk. Can you detail ?08:29
gfidentemaking it depend optionally on puppet-pacemaker so that it would use pcs08:29
gfidentespredzy, ?08:29
spredzygfidente, no no, my idea was to create a wrapper in puppet-tripleo that was inspired from the one in openstack_extras08:30
spredzybut rather of relying on cs_primitive and cs_clone, we would directly rely on pcmk_resource08:30
gfidenteah without importing openstack_extra then08:30
gfidenteI see, thanks08:30
spredzyyes totally, just as an inspiration08:30
jistrspredzy: no i can't because i've no idea what the scenario chunk does tbh :D just wanted to point you guys to it so that we have a better picture of what needs to actually happen on the system to get the recommended arch running for galera08:30
openstackgerritdave-mcnally proposed openstack/tripleo-image-elements: Add ability to disable stunnel  https://review.openstack.org/13605208:31
jistri'm just a newbie with pacemaker :))08:31
gfidentejistr, looks like a cleanup script?08:32
gfidentewhat does this mysqltest user does?08:33
gfidenteis it needed by the resource agent maybe?08:33
gfidentewe'd need to check that spredzy08:33
gfidente:(08:33
jistrhmm i don't see any mention of mysqltest in astapor...08:34
gfidentenot this is the clustercheck user08:34
gfidentemysql -e "CREATE USER 'clustercheck'@'localhost' IDENTIFIED BY '${PHD_VAR_env_password}';08:34
spredzygfidente, user used for clustrchk command08:35
gfidenteoh mysqltest is the password .... root is the user :P08:35
gfidentespredzy, yeah which needs to be added .... better we start experimenting with that as well08:36
spredzygfidente, yep I am reading this document and will update my review accordingly08:37
gfidentespredzy, review?08:38
spredzybased on that document, I need to redo globally the galera review I have pending08:38
gfidentespredzy, ah yeah08:38
gfidentethat is terrible08:38
spredzyto matches this scenario as it is the reference scenario for Galera installation08:38
gfidenteI know08:38
gfidenteI know08:38
spredzySame will be needed for MongoDB and Redis08:38
*** akrivoka has quit IRC08:40
gfidenteso it's not entirely clear to me the corelation in between the https://github.com/redhat-openstack/astapor/blob/master/puppet/modules/quickstack/manifests/pacemaker/galera.pp and the scenarios08:40
gfidentefor rabbitmq I opted again to include just the minimal bits to make it working with the idea of 'adding specific knowledge' in later steps08:41
openstackgerritJan Provaznik proposed openstack/tripleo-common: Initial Cookiecutter Commit.  https://review.openstack.org/17664908:41
gfidentejprovazn, hey there08:42
jprovazngfidente: Hi08:42
mariosohnoe http://docs.openstack.org/developer/tripleo-incubator/devtest.html down? (gives me 500)08:44
*** aufi has joined #tripleo08:44
gfidentemarios rst was created to be readable in a text editor! ENJOY! :)08:45
jistrbtw i just reported this https://bugs.launchpad.net/tripleo/+bug/144749708:46
openstackLaunchpad bug 1447497 in tripleo "Corosync.conf should be set up with hostnames, not IPs" [Medium,Triaged]08:46
mariosgfidente: heh thanks \o/08:46
jistrunfortunately it's not as easy to fix as i thought it would be08:46
*** aufi_ has quit IRC08:46
gfidentemarios, sorry :) https://github.com/openstack/tripleo-incubator/blob/master/scripts/devtest.sh08:47
spredzyjistr, are you on it or you want us to look ?08:47
mariosgfidente: what my looking at08:47
jistrspredzy: i'm on it08:47
spredzyjistr, ack08:47
gfidentejistr, why not easy as expected?08:47
gfidentejistr, I was trying to avoid names though and generally preferring IPs everywhere08:48
gfidentepartially because of the networking changes coming08:48
gfidentepartially because the hostnames we get have uppercase letters forcing us to downcase08:48
gfidentepartially because it would make things more resistent to name resolution failures08:49
gfidenteso, is there a particular reason why we should use hostnames there instead?08:49
*** adrianopetrich has quit IRC08:49
mariosgfidente: \o/ http://docs-draft.openstack.org/45/173645/6/check/gate-tripleo-incubator-docs/c1a6b72//doc/build/html/08:49
gfidente(about networking, that means enforcing communication over the network we want, instead of relying on the name mapping to the appropriate subnet)08:50
gfidentemarios, uauhuh also good, better actually :)08:50
jistrgfidente: that's a good question, i don't know for sure, but i though Pacemaker printing this isn't a good thing: "WARNING: corosync and pacemaker node names do not match (IPs used in setup?)"08:50
jistrgfidente: also astapor and spinalstack both use names there08:50
gfidenteI saw the warning as well, but for all those reasons08:50
gfidenteuntil I figure why we should use hostnames08:50
gfidenteat least why corosync prefers hostnames08:51
gfidenteI tend not to use them08:51
mariosgfidente: so the ip could change right?08:51
mariosgfidente: i mean if it is rebooted, could get a different allocation08:51
mariosgfidente: unless you enforce static dhcp08:51
*** dtantsur is now known as dtantsur|brb08:52
mariosgfidente: btw, i don't have any context here, am just reading the last few lines08:52
marios:)08:52
gfidentemarios, in theory yes but 1) we'd get the change in heat so services would be reconfigured (not nice, but that would actually happen) 2) the changes for the network refarch coming are meant to assign static IPs and multiple IPs for different networks08:52
gfidenteso if I want to make sure ceph is reaching it's osd nodes on the storage network, I can't rely on the hostname mapping to the storage network ip08:53
mariosgfidente: i see. so yeah if you have static address allocation then that will work ok08:53
gfidentebecause hostname will probably point to controlplane08:53
gfidenteso I am forced to give ceph the IPs which the OSDs have on the dedicated storage network08:53
gfidenteand this same problem applies in other circumstances08:54
*** pelix has joined #tripleo08:54
gfidentejistr, for now I will just add my comments on the bug08:55
*** Matt2 has quit IRC08:56
*** untriaged-bot has joined #tripleo09:00
untriaged-botNo untriaged bugs so far! \o/09:00
*** untriaged-bot has quit IRC09:00
gfidentejistr, actually dsneddon can probably help us figure the best approach09:00
jistryeah i wonder who's the best person to ask about these things in our timezone?09:01
gfidenteand anyway, corosync having a warning means we might want to fix that for corosync anyway09:01
gfidentejistr, for the switch to work, I think this is just because you need to downcase((09:02
gfidentedowncase()09:02
gfidenteI faced that before with rabbitmq09:03
jistrgfidente: nope, i did that :( http://fpaste.org/214640/77980214/09:03
jistrgfidente: all nodes think that they are the DC and they never form a cluster with the rest09:03
jistrgfidente: i wonder if the *new warning* ( :D ) has something to do with it09:04
gfidentestonith is disabled though09:04
gfidenteor at least it should be09:04
jistryeah that's puzzling, i saw it in the manifest, but it doesn't get disabled for me, or at least not in the right moment maybe09:05
jistrit might be an ordering issue09:05
* jistr looks further09:05
gfidentettyl :)09:05
*** rdopiera has quit IRC09:09
*** jcoufal_ has joined #tripleo09:10
*** jcoufal has quit IRC09:13
*** rdopiera has joined #tripleo09:14
*** shardy has joined #tripleo09:18
*** d0ugal has quit IRC09:29
*** adrianopetrich has joined #tripleo09:29
*** d0ugal has joined #tripleo09:29
*** d0ugal is now known as Guest8147209:29
*** Guest81472 is now known as d0ugal209:40
*** derekh has joined #tripleo09:41
*** d0ugal2 is now known as d0ugal09:47
*** d0ugal has quit IRC09:47
*** d0ugal has joined #tripleo09:47
*** jcoufal_ is now known as jcoufal09:55
jprovaznshardy: morning, have a heat stack tricky question for you09:58
shardyjprovazn: morning, sure!09:59
* shardy wonders if he needs more coffee ;)09:59
*** pcrews has quit IRC09:59
*** cody-somerville has joined #tripleo09:59
*** cody-somerville has joined #tripleo09:59
jprovaznshardy: well, it's not so tricky, but... when checking if stack is waiting on a hook, would it be safe to check only resources which have updated_time||created_time > stack's updated_time||created_time?10:01
*** dtantsur|brb is now known as dtantsur10:01
jprovaznIOW is it safe to assume, that update/create timestamp for stack is always older than on any resource inside that stack?10:01
*** pdb has quit IRC10:02
shardyjprovazn: Probably not, because the hooks don't change the state of the resource when you hit them10:03
jprovaznhmpf10:03
shardyjprovazn: however, we should be able to check if the *event* timestamp is greater than the stack updated time10:03
shardythat's exactly what I'm trying to do now with the hook-poll patch10:03
shardybut our timestamp filtering in the event API is broken10:04
jprovaznshardy: yep, for events I think this is even necessary (I added few comments to your heatclient patch)10:04
shardyit would be possible to filter locally in heatclient as a temporary workaround10:04
jprovaznoh, I had to miss the bit where you already do it10:04
shardyjprovazn: ah, thanks not seen those yet, and yes it's definitely needed10:04
shardyjprovazn: I don't do it yet, I've been trying to fix the event API10:04
jprovaznshardy: aah, ok10:05
shardyI noticed the problem when I had an update time out, then there was a bogus pending pre-update hook on the subsequent update from the failed state10:05
jprovaznshardy: yes, this is exactly the situation I hit it10:05
jprovaznshardy: do you have a link for the event bug you mentioned?10:06
shardyjprovazn: I'm hoping to push something (either API fix or heatclient workaround) later today10:06
*** pdb has joined #tripleo10:06
jprovaznshardy: cool10:06
shardyhttps://bugs.launchpad.net/heat/+bug/144410010:06
openstackLaunchpad bug 1444100 in heat "Event filtering is non-obvious and broken" [Medium,Triaged] - Assigned to Satyanarayana Patibandla (satya-patibandla)10:06
*** rdopiera has quit IRC10:13
*** rdopiera has joined #tripleo10:14
*** Goneri has quit IRC10:34
gfidentespredzy, can you review this please https://review.openstack.org/11909310:46
gfidenteI took over existing change because couldn't get in touch with the guy for days10:47
gfidenteand it was 'abandoned' from weeks10:47
spredzygfidente, sure. Will review after lunch10:47
*** adrianopetrich has quit IRC11:03
*** rdopiera has quit IRC11:04
*** rdopiera has joined #tripleo11:05
*** yuanying has quit IRC11:10
*** rdopiera has quit IRC11:16
*** rdopiera has joined #tripleo11:16
jistroh my :/11:18
jistri've found why setting up cluster using hostnames doesn't work11:18
jistrhttp://fpaste.org/214695/29787673/11:18
jistr^ that's how /etc/hosts looks like on the deployed nodes :-/11:19
jistrafter removing this:11:20
jistr 127.0.0.1       ov-rl5i5saoc6h-1-hj5tzsbrdv4c-controller-dy6nuyarqy5z.novalocal ov-rl5i5saoc6h-1-hj5tzsbrdv4c-controller-dy6nuyarqy5z11:20
jistrit works11:20
*** subscope has quit IRC11:21
jistri wonder where/how does that line even get generated, and under what cicumstances is it ever a good idea to bind FQDN to loopback address :D11:21
*** rhallisey has joined #tripleo11:25
*** thrash|g0ne is now known as thrash11:28
gfidentederekh, does my comment https://review.openstack.org/#/c/173940 look valid?11:28
*** david-lyle has quit IRC11:29
gfidentejistr, looks like something from cloud-init ?11:29
derekhgfidente: dunno havn't used --abort, I'll take a look, thanks11:29
jistrgfidente: looking for it, thx for the suggestion, will try to find it in cloud-init11:30
gfidentederekh, actually adding it in place of || true would make it execute even if the fetch failed11:32
gfidentewhich we don't want11:32
gfidentelooks like it needs a more controlled case11:32
gfidente&& (git cherry-pick FETCH_HEAD || git cherry-pick --abort) || true11:34
gfidentederekh, ^^11:34
gfidentemerge > https://review.openstack.org/#/c/171411/11:40
*** jdob has joined #tripleo11:40
*** weshay has quit IRC11:41
*** funzo has quit IRC11:41
*** funzo has joined #tripleo11:41
*** subscope has joined #tripleo11:42
*** weshay has joined #tripleo11:46
*** Goneri has joined #tripleo11:49
*** yog__ has quit IRC11:51
jprovaznI wonder what is the minimum supported tuskarclient version which should live in openstack/requirements? https://review.openstack.org/#/c/174233/11:56
jprovaznjdob: ^ any opinion?11:56
jdobnot off the top of my head, though for a first pass I'd just aim for the latest one11:57
jdob0.1.1711:57
jprovaznok, thx11:58
*** mmagr has joined #tripleo11:58
*** david-lyle has joined #tripleo12:05
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: puppet: install Horizon on overcloud-controller  https://review.openstack.org/16439212:06
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: puppet: install Horizon on overcloud-controller  https://review.openstack.org/16439212:07
*** masco has quit IRC12:10
*** prad has quit IRC12:17
*** dasm is now known as dasm|afk12:20
*** tzumainn has joined #tripleo12:21
*** dprince has joined #tripleo12:23
*** mmagr is now known as mmagr|afk12:27
*** MasterPiece has joined #tripleo12:28
jprovaznshardy: I have a proposal, I bet you will love it ;). It would be really great (I'm not afraid to use "necessary") to add an optional param to the "do_hook_poll"method which would allow checking for hooks on particualar resources. This would be a regexp (or at least some wildcard support), so then I could specify that I want to check for hooks only on "*NodesPostDeployment" resources for stepped creation story, or on "update_deployment"12:30
jprovaznfor package update story. This would significantly improve performance because in the do_hook_poll we would just list all resources (with nested_depth=x) and then checked events only on resources with the matched name. In compare to current situation where you get all events recursively12:30
*** david-lyle has quit IRC12:31
*** trown|outttypeww is now known as trown12:33
openstackgerritDerek Higgins proposed openstack/tripleo-incubator: Drop usage of keystoneclient  https://review.openstack.org/17671512:34
openstackgerritDerek Higgins proposed openstack/tripleo-incubator: Don't install the keystone client cli  https://review.openstack.org/17671612:34
openstackgerritDerek Higgins proposed openstack/tripleo-image-elements: Drop usage of keystoneclient  https://review.openstack.org/17671712:35
openstackgerritDerek Higgins proposed openstack/tripleo-image-elements: Don't install the keystone client cli  https://review.openstack.org/17671812:35
*** mmagr|afk is now known as mmagr12:35
shardyjprovazn: +1, I was thinking similar, but I wanted to get the global-poll working first12:36
shardyjprovazn: if you'd like to have a go at implementing that, feel free :)12:36
shardyjprovazn: Or I can take a look (probably tomorrow, I'm looking at the heat API part today)12:36
shardyjprovazn: based on my testing, I'm not certain doing multiple resource-filtered calls for events will actually be faster though12:37
jprovaznshardy: I can give it a shot, I have PTO tomorrow afternoon, I will sync with you tomorrow morning with what I have (for now I was trying to merge/re-use your code in my hook polling code)12:39
shardyjprovazn: Ok, sounds good, thanks!12:42
jprovazncool, np12:42
*** rlandy has joined #tripleo12:47
*** jayg|g0n3 is now known as jayg12:49
*** adrianopetrich has joined #tripleo12:50
openstackgerritEmilien Macchi proposed openstack/tripleo-puppet-elements: elements: drop puppet-openstack module  https://review.openstack.org/17672812:54
*** jprovazn has quit IRC13:00
*** sdake_ has joined #tripleo13:10
*** sdake has quit IRC13:14
*** Marga_ has joined #tripleo13:14
*** mmagr is now known as mmagr|afk13:26
*** lblanchard has joined #tripleo13:32
*** mmagr|afk is now known as mmagr13:35
*** jprovazn has joined #tripleo13:38
*** wuhg has quit IRC13:39
*** Marga_ has quit IRC13:52
*** shardy_ has joined #tripleo13:57
*** sdake_ has quit IRC13:58
*** shardy has quit IRC13:59
*** MasterPiece has quit IRC14:01
*** shardy_ has quit IRC14:02
*** shardy has joined #tripleo14:03
*** ChuckC has quit IRC14:13
*** rhallisey has quit IRC14:16
shardyslagle, dprince: Hey I was wondering about https://review.openstack.org/#/c/170137/3 (and the subsequent patch)14:16
shardyAre you planning to land those, or did you have some other plan for allowing deployer extra-config hooks?14:17
*** rwsu has joined #tripleo14:24
openstackgerritDerek Higgins proposed openstack/tripleo-incubator: Drop usage of keystoneclient  https://review.openstack.org/17671514:31
openstackgerritDerek Higgins proposed openstack/tripleo-incubator: Don't install the keystone client cli  https://review.openstack.org/17671614:31
slagleshardy: they are looking good to me14:31
slagleshardy: i still need to update the rhel reg one based on some of your feedback14:32
*** aufi has quit IRC14:33
*** Marga_ has joined #tripleo14:34
*** jistr is now known as jistr|mtg14:35
*** BadCub_Hiding is now known as BadCub14:36
*** Marga_ has quit IRC14:36
*** rdopiera has quit IRC14:37
*** Marga_ has joined #tripleo14:37
*** mmagr is now known as mmagr|afk14:38
*** mmagr|afk is now known as mmagr14:46
*** ChuckC has joined #tripleo14:48
gfidentejistr|mtg, can you help me figure14:52
gfidentewhy the tests https://review.openstack.org/#/c/119093/ are failing?14:52
shardyslagle: Ok, sounds good, just wanted to check they weren't waiting on any rework from me14:53
*** trown is now known as trown|brb14:57
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: puppet: install Horizon on overcloud-controller  https://review.openstack.org/16439214:59
*** untriaged-bot has joined #tripleo15:00
untriaged-botNo untriaged bugs so far! \o/15:00
*** untriaged-bot has quit IRC15:00
*** Goneri has quit IRC15:01
*** trown|brb is now known as trown15:01
*** pcrews has joined #tripleo15:02
EmilienMslagle: I have a stupid question for you.15:03
EmilienMslagle: I would like to install the script that informs bugs15:04
EmilienMI found the script, I ran it, it work. My question is: where do you run it?15:04
EmilienMin os-infra?15:04
EmilienM(context: having it in puppet-openstack)15:04
slagleEmilienM: i'm actually pretty clueless about that15:04
slaglei think rpodolyaka set it up15:05
*** Marga_ has quit IRC15:13
*** Marga_ has joined #tripleo15:13
*** rhallisey has joined #tripleo15:17
*** Marga_ has quit IRC15:18
openstackgerritJames Slagle proposed openstack/tripleo-image-elements: Make delorean.repo overrideable  https://review.openstack.org/17682415:28
openstackgerritOpenStack Proposal Bot proposed openstack/diskimage-builder: Updated from global requirements  https://review.openstack.org/17682615:28
*** david-lyle has joined #tripleo15:29
openstackgerritOpenStack Proposal Bot proposed openstack/os-cloud-config: Updated from global requirements  https://review.openstack.org/15777615:36
openstackgerritOpenStack Proposal Bot proposed openstack/os-collect-config: Updated from global requirements  https://review.openstack.org/17684715:36
*** derekh has quit IRC15:36
openstackgerritOpenStack Proposal Bot proposed openstack/tuskar: Updated from global requirements  https://review.openstack.org/17685415:37
*** athomas has quit IRC15:43
*** morazi has quit IRC15:47
*** funzo has quit IRC15:48
*** david-lyle has quit IRC15:56
*** mmagr is now known as mmagr|afk15:56
*** mmagr|afk is now known as mmagr15:58
*** rhallisey has quit IRC16:00
*** trown is now known as trown|lunch16:02
*** mmagr is now known as mmagr|afk16:04
*** masco has joined #tripleo16:05
shardydprince: I had another (simpler) idea about the NodeIndex IP mappings16:06
shardyhttp://fpaste.org/214819/42980518/16:06
shardyThis currently doesn't work, but I'm looking into what needs to change in Heat to make that possible16:07
shardyWe should be able to reparse the get_param after the %index% substitution16:07
shardythen the mapping is just a lookup from a data file16:07
shardyobviously ip_mappings could get big, but it's going to be trivial to generate16:08
*** mmagr|afk has quit IRC16:14
*** mcornea has quit IRC16:15
*** jistr|mtg is now known as jistr16:18
dprinceshardy: could work, the IpMappings could get quite large though. It might be cleaner if we just passed around cidr, index pairs for each network instead16:19
*** weshay has quit IRC16:23
*** rhallisey has joined #tripleo16:27
openstackgerritBen Nemec proposed openstack/os-cloud-config: Enable ssl registration for Heat  https://review.openstack.org/17648316:29
openstackgerritBen Nemec proposed openstack/os-cloud-config: Don't register https endpoints for non-ssl services  https://review.openstack.org/17690016:29
*** dtantsur is now known as dtantsur|afk16:30
shardydprince: Yeah, but then we still have the problem that the actual IP of the node isn't obvious to the operator16:38
shardydprince: I was aiming to solve that by using the simple static lookup16:38
shardyas it means you can expose the node IPs very easily, just by looking at the attributes of the resource (e.g using outputs in the template representing the node inside the group)16:39
shardyeither way should work I guess16:39
dprinceshardy: so I gather from #heat that the function plugability is really about pinning them to API versions?16:40
*** jcoufal has quit IRC16:41
shardydprince: sort of, they're pluggable internally so we can easily support different functions per HOT version16:41
shardyso, I guess that pattern breaks if you import non-heat functions16:41
dprinceshardy: right, but not externally, yep16:42
shardyWe definitely could do it with a custom resource, but to me that feels like the wrong approach atm16:42
shardyI'll try to catch stevebaker later to discuss the idea of writable outputs from SoftwareDeployments16:43
*** jistr has quit IRC16:43
shardyprobably not the short-term solution tho16:43
dprinceshardy: cool. If it is too late for you I can try him too?16:43
shardydprince: sure, feel free16:44
shardyI'm going afk for a bit but will most likely be around later16:44
*** bvandenh has joined #tripleo16:44
dprinceshardy: sure, thanks16:44
*** eghobo has joined #tripleo16:44
shardydprince: one way to do it would be a WaitCondition which gets a signal per-group-member16:45
shardybecause that already exposes the data sent via the signal16:45
shardybut it's kinda messy mixing the two signalling mechanisms16:45
shardyso I'd prefer not to unless we really have to16:45
*** cody-somerville has quit IRC16:45
dprinceshardy: agree16:45
*** penick has joined #tripleo16:47
shardydprince: Ok, cool, I'll think about it some more, and I'll probably try to get the %index% json lookup working anyway, as I think that would be useful regardless of if we use it for this16:47
*** alop has joined #tripleo16:47
*** shardy is now known as shardy_afk16:48
*** masco has quit IRC16:58
*** masco has joined #tripleo17:01
*** bvandenh has quit IRC17:05
*** dsneddon has joined #tripleo17:05
*** lucasagomes has quit IRC17:05
*** rhallisey has quit IRC17:14
*** trown|lunch is now known as trown17:14
*** pelix has quit IRC17:27
*** sdake has joined #tripleo17:28
*** rhallisey has joined #tripleo17:29
*** masco has quit IRC17:36
*** dprince has quit IRC17:38
*** jprovazn has quit IRC17:48
*** jprovazn has joined #tripleo17:51
*** dprince has joined #tripleo17:53
*** morazi has joined #tripleo18:05
*** jprovazn has quit IRC18:06
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: puppet: install Horizon on overcloud-controller  https://review.openstack.org/16439218:07
*** jprovazn has joined #tripleo18:07
*** david-lyle has joined #tripleo18:07
*** bvandenh has joined #tripleo18:08
*** jehb has quit IRC18:11
*** bvandenh has quit IRC18:16
*** derekh has joined #tripleo18:17
openstackgerritMerged stackforge/puppet-tripleo: Bump rspec-puppet to 2.1.0  https://review.openstack.org/17661018:23
*** morganfainberg is now known as grebniafnagrom18:24
*** grebniafnagrom is now known as morganfainberg18:24
*** ifarkas has quit IRC18:25
*** Goneri has joined #tripleo18:28
*** bvandenh has joined #tripleo18:34
*** jdob has quit IRC18:40
*** bvandenh has quit IRC18:40
*** jdob has joined #tripleo18:40
*** gfidente has quit IRC18:46
*** derekh has quit IRC18:49
*** jdob has quit IRC18:49
*** jdob has joined #tripleo18:49
*** lblanchard has quit IRC18:59
*** derekh has joined #tripleo19:00
*** lblanchard has joined #tripleo19:02
*** lblanchard has quit IRC19:03
*** spzala has joined #tripleo19:03
*** sdake_ has joined #tripleo19:09
*** sdake has quit IRC19:12
*** jtomasek has quit IRC19:20
*** trown is now known as trown|outttypeww19:23
*** penick has quit IRC19:32
*** penick has joined #tripleo19:37
*** bvandenh has joined #tripleo19:43
*** jtomasek has joined #tripleo19:47
openstackgerritJames Slagle proposed openstack/os-cloud-config: Create flavors from registered nodes  https://review.openstack.org/17696620:18
*** alop has quit IRC20:19
*** jprovazn has quit IRC20:21
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add functional smoke test for disk-image-create  https://review.openstack.org/16916720:27
EmilienMdprince: are you using packaging or source for Puppet modules *today* ?20:31
EmilienMin OOO CI jobs ?20:31
EmilienMslagle: ^20:31
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add functional smoke test for disk-image-create  https://review.openstack.org/16916720:32
*** alop has joined #tripleo20:34
dprinceEmilienM: source20:35
EmilienMdprince: good then.20:35
dprinceEmilienM: all of our CI uses source20:35
dprinceEmilienM: and I think most devs are using it for now too20:35
dprinceEmilienM: once the packages is rev'd it should be fine too though20:35
EmilienMbecause if we switch to package (OPM), the job for puppet- * modules would be useless20:36
* greghaynes loves when fixing up tests finds bugs20:36
EmilienMor maybe I'm missing something20:36
*** trown|outttypeww is now known as trown20:37
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add functional smoke test for disk-image-create  https://review.openstack.org/16916720:38
openstackgerritJames Slagle proposed openstack/os-cloud-config: Remove Nova Baremetal support  https://review.openstack.org/17697320:40
slagleEmilienM: i also noticed there were some modules in the puppet-modules element that I dont think are used20:41
EmilienMslagle: I realized that this morning with puppet-openstack20:42
EmilienMI'll have a look20:42
EmilienMand patch20:42
slagleEmilienM: yea, that was the one i was wondering about20:42
EmilienMslagle: I did a patch already20:42
slagleok20:43
*** rhallisey_ has joined #tripleo20:45
*** rhallisey has quit IRC20:48
*** derekh has quit IRC20:51
openstackgerritMerged openstack/tripleo-puppet-elements: elements: drop puppet-openstack module  https://review.openstack.org/17672820:54
*** alop has quit IRC20:56
*** rhallisey_ has quit IRC20:56
*** untriaged-bot has joined #tripleo21:00
untriaged-botNo untriaged bugs so far! \o/21:00
*** untriaged-bot has quit IRC21:00
*** shardy_afk is now known as shardy21:05
*** alop has joined #tripleo21:07
*** alop has quit IRC21:19
*** alop has joined #tripleo21:23
*** spzala has quit IRC21:31
*** alop has quit IRC21:33
*** MasterPiece has joined #tripleo21:35
*** Goneri has quit IRC21:48
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add functional smoke test for disk-image-create  https://review.openstack.org/16916721:49
*** Goneri has joined #tripleo21:50
*** derekh has joined #tripleo21:51
*** alop has joined #tripleo21:53
*** dprince has quit IRC21:53
*** jtomasek has quit IRC21:53
openstackgerritgreghaynes proposed openstack/diskimage-builder: Fail functests if no qemu-img is found  https://review.openstack.org/17700221:57
greghaynesok, looks like those two ^ will pass, can I please get some reviews on them to unwedge DIB?21:58
greghaynesI tried to explain what is going on in the second commit msg21:58
*** derekh has quit IRC22:04
*** openstackstatus has quit IRC22:09
*** Goneri has quit IRC22:09
*** greghaynes has quit IRC22:18
*** greghaynes has joined #tripleo22:19
*** trown is now known as trown|outttypeww22:31
*** MasterPiece has quit IRC22:42
*** jayg is now known as jayg|g0n322:50
*** noslzzp has quit IRC22:51
*** jdob has quit IRC22:53
*** jdob has joined #tripleo22:54
*** yuanying has joined #tripleo23:00
*** penick has quit IRC23:01
*** eghobo has quit IRC23:02
*** derekh has joined #tripleo23:02
*** derekh has quit IRC23:04
*** ukalifon has joined #tripleo23:04
*** alop has quit IRC23:04
*** rlandy has quit IRC23:09
openstackgerritSteve Baker proposed openstack/tripleo-incubator: QuintupleO setup environment  https://review.openstack.org/17703223:14
*** sdake has joined #tripleo23:16
*** sdake_ has quit IRC23:20
*** wuhg has joined #tripleo23:28
*** julim has quit IRC23:28
*** panda has quit IRC23:30
*** panda has joined #tripleo23:30
*** alop has joined #tripleo23:32
*** Marga_ has joined #tripleo23:38
*** eghobo has joined #tripleo23:58
*** sdake_ has joined #tripleo23:59

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