Monday, 2016-01-18

« Sunday, 2016-01-17 Index Tuesday, 2016-01-19 »
openstackgerritSteve Baker proposed openstack/tripleo-common: Convenience methods in base model class  https://review.openstack.org/26882100:07
openstackgerritSteve Baker proposed openstack/tripleo-common: Implement ViewManager list_software_deployments  https://review.openstack.org/26882200:07
*** mcornea has quit IRC00:17
*** mkovacik has quit IRC00:30
openstackgerritDan Sneddon proposed openstack/os-net-config: Add ability to set IPv6 routes in os-net-config templates  https://review.openstack.org/26880600:42
*** Marga_ has joined #tripleo00:43
*** Marga_ has quit IRC00:51
*** Marga_ has joined #tripleo00:52
*** davidlenwell has quit IRC01:05
*** davidlenwell has joined #tripleo01:07
*** ChanServ sets mode: +v davidlenwell01:07
openstackgerritDan Sneddon proposed openstack/tripleo-heat-templates: Add all isolated networks to all nodes.  https://review.openstack.org/26883301:12
openstackgerritDan Sneddon proposed openstack/os-net-config: Add ability to set IPv6 routes in os-net-config templates  https://review.openstack.org/26880601:41
openstackgerritOpenStack Proposal Bot proposed openstack/diskimage-builder: Updated from global requirements  https://review.openstack.org/26843801:41
*** yuanying has quit IRC02:04
*** yuanying has joined #tripleo02:05
openstackgerritDan Sneddon proposed openstack/os-net-config: Add ability to set IPv6 routes in os-net-config templates  https://review.openstack.org/26880602:13
*** yamahata has joined #tripleo02:15
*** yamahata has quit IRC02:22
*** Goneri has joined #tripleo02:25
*** morazi has quit IRC02:43
*** yuanying has quit IRC03:00
*** yuanying has joined #tripleo03:01
*** yuanying has quit IRC03:20
*** Marga_ has quit IRC03:23
*** links has joined #tripleo03:26
*** Goneri has quit IRC03:45
*** trozet_ has joined #tripleo04:00
*** trozet has quit IRC04:04
*** trozet_ is now known as trozet04:04
*** yuanying has joined #tripleo04:06
*** panda_ has quit IRC04:06
*** panda_ has joined #tripleo04:07
*** jbonjean has joined #tripleo04:12
*** olap_ has quit IRC04:18
*** olap__ has quit IRC04:18
*** coolsvap|away is now known as coolsvap05:11
*** jaosorior has joined #tripleo05:41
*** trozet has quit IRC05:58
*** anande has joined #tripleo06:24
*** anande has quit IRC06:24
*** chlong has quit IRC07:00
openstackgerritIsaac Beckman proposed openstack/diskimage-builder: Move speedup section to image building guide  https://review.openstack.org/26876907:02
*** coolsvap is now known as coolsvap|away07:18
*** ifarkas has joined #tripleo07:22
*** rasca has joined #tripleo07:36
*** rcernin has joined #tripleo07:41
*** hewbrocca has quit IRC07:41
*** jtomasek has joined #tripleo07:46
*** jprovazn has joined #tripleo07:53
*** ukalifon1 has joined #tripleo07:54
*** paramite has joined #tripleo07:59
*** liverpooler has joined #tripleo08:02
*** fgimenez has joined #tripleo08:05
*** fgimenez has joined #tripleo08:05
*** aufi has joined #tripleo08:17
*** jcoufal has joined #tripleo08:24
*** dshulyak_ has joined #tripleo08:24
*** mcornea has joined #tripleo08:25
*** hewbrocca has joined #tripleo08:26
*** jaosorior has quit IRC08:29
*** jaosorior has joined #tripleo08:30
*** shardy has joined #tripleo08:40
*** devvesa has joined #tripleo08:41
*** akrivoka has joined #tripleo08:46
*** mkovacik has joined #tripleo08:46
*** nico_auv has joined #tripleo08:56
*** mbound has joined #tripleo09:02
*** mbound has quit IRC09:02
*** mbound has joined #tripleo09:02
*** bvandenh has joined #tripleo09:02
devvesaping marios09:05
*** olap_ has joined #tripleo09:10
*** derekh has joined #tripleo09:11
derekhshardy: gfidente in CI undercloud cpu == 2 and  MEM == 5G09:11
gfidentederekh++ :)09:12
derekhI rebuild them last night, I've just now noticed that no tests have passed since the rebuild09:12
derekh2016-01-18 05:59:50.524 | No image with a name or ID of 'bm-deploy-kernel' exists.09:12
*** bvandenh has quit IRC09:12
derekhit doesn't look related but may be09:12
derekhwill be back in a few minutes to see if I can find out09:13
shardyderekh: nice, thanks!09:13
shardywill be interesting to see how the performance of the ha job changes09:13
derekhyup09:14
shardydtantsur may have some ideas re the deploy ramdisk thing09:14
dtantsurmorning folks! who can make a DIB release please?09:14
*** jistr has joined #tripleo09:14
dtantsurderekh, hey, what's the context here? is it 'image upload' command misbehaving?09:15
*** mkovacik has quit IRC09:16
*** olap_ has quit IRC09:17
*** olap_ has joined #tripleo09:17
*** jaosorior has quit IRC09:17
*** jaosorior has joined #tripleo09:18
dtantsurshardy, hey, are you able to make DIB releases?09:19
*** olap_ has quit IRC09:19
shardydtantsur: in theory, but I've never done one so not familiar with the process09:19
shardymarios may be able to do one quicker as IIRC he was cutting some releases recently09:20
dtantsurmarios, morning, around?09:20
*** olap_ has joined #tripleo09:20
*** bvandenh has joined #tripleo09:24
*** athomas has quit IRC09:28
*** athomas has joined #tripleo09:34
gfidentederekh, is 5G enough with 2 cores btw? i remember a lot of subsystems scaling the number of processes based on core count so memory consumption going up as well09:35
derekhgfidente: good point, I guess we'll find out09:38
shardyderekh: we may want to land https://review.openstack.org/#/c/268253/ which reverts testing deletes for now09:38
derekhdtantsur: looks like its related to the instance changes I made09:38
shardyI'll re-propose when I've figured out how to retain the overcloud logs, and quantified the actual time taken in CI09:39
shardythat will also give us a bit more time towards testing updates09:39
shardyperhaps when I un-revert, we can just test deletes on the nonha job or something09:40
gfidenteshardy, so you want to land the rever first anyway?09:40
derekhshardy: ya, makes sense09:40
derekhalthough the resn for todays failuers seems to be09:40
derekherror: Couldn't fork %postun(nss-3.19.1-18.el7.x86_64): Cannot allocate memory09:40
shardygfidente: Yeah I think we should land the revert, because we need to retain overcloud logs09:41
derekhand various errors like this one09:41
gfidenteshardy, ack09:41
gfidentederekh, yeah I think it's running oom09:41
shardydoh09:41
derekhmoving from 4G to 5G instance must have push it into using the tmpfs09:41
shardyneed moar memory09:41
derekhfor image build09:41
openstackgerritMerged openstack-infra/tripleo-ci: Revert "Add CI coverage of deleting the deployed overcloud"  https://review.openstack.org/26825309:41
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Set DIB_NO_TMPFS=1  https://review.openstack.org/26895509:45
derekhshardy: gfidente lets see if that helps09:45
derekhI didn't put it into tripleo-common as it could be just a sweet spot the CI undercloud falls into09:46
*** mkovacik has joined #tripleo09:51
*** mgould has joined #tripleo09:59
*** rasca has quit IRC10:04
*** teju has joined #tripleo10:07
*** pblaho has joined #tripleo10:07
*** marcusvrn_ has joined #tripleo10:13
*** cd-undercloud has joined #tripleo10:17
cd-undercloud************** overcloud complete status=1 ************10:17
*** cd-undercloud has quit IRC10:17
derekhits been a while since I saw that ^10:19
*** akrivoka has quit IRC10:19
openstackgerritVishal kumar mahajan proposed openstack/python-tuskarclient: Replace assertEqual(None, *) with assertIsNone in tests  https://review.openstack.org/26896710:19
*** rasca has joined #tripleo10:21
*** cd-undercloud has joined #tripleo10:23
cd-undercloud************** overcloud complete status=1 ************10:23
*** cd-undercloud has quit IRC10:23
dtantsurby the way, what does it ^^ mean?10:23
dtantsurderekh, ^^10:23
derekhsomebody is running an old script, one sec and I'll find it10:23
derekhdtantsur: http://git.openstack.org/cgit/openstack/tripleo-image-elements/tree/elements/tripleo-cd/deploy-ci-overcloud#n11010:24
dtantsuroh nice. shouldn't we kill it with fire already?10:25
derekhWho ever is running it should no that it hasn't been used in a loooong time, probably doesn't work and we should just remove it10:25
derekhdtantsur: we'll most of whats in that repository could be10:26
derekh*could have the same said about it10:27
*** cd-undercloud has joined #tripleo10:27
cd-undercloud************** overcloud complete status=1 ************10:27
*** cd-undercloud has quit IRC10:27
*** akrivoka has joined #tripleo10:32
* afazekas funny this cd-undercloud , my machine sent this, sorry I just running some code I have never seen :)10:32
afazekasWhat is the correct command and/or procedure to set up tripleo-ci worker , which listens on gearman ?10:33
derekhafazekas: sorry for not getting back to you last week, I got bogged down tring to see if I could get the overcloud actually working before you started trying tepest at it10:39
derekhathomas: this is the simple ping test we've been adding https://review.openstack.org/#/c/262028/10:39
derekhafazekas: ^ athomas sorry10:39
derekhafazekas: you shouldn't need to set things up exactly like CI10:39
derekhafazekas: whats important is that you can run this script http://git.openstack.org/cgit/openstack-infra/tripleo-ci/tree/toci_instack.sh10:40
afazekasit requires a TE_DATAFILE10:40
derekhafazekas: with $TE_DATAFILE set point to a json file that holds the details of what nodes to use10:40
afazekasit looks like an instack env json in the other jobs10:40
derekhafazekas: exactly so you should be able to just use one set up by instack10:41
derekhafazekas: actually If you would like, I can sping you up a VM on our ci cloud so you can get a CI env10:41
*** regebro has joined #tripleo10:41
afazekasWhat is the recommended instack setting, and which version should be used ?10:42
derekhafazekas: we can do a quick shared screen session and I'll should you how to get an env10:42
afazekasok10:43
derekhafazekas: ok, I'll spin up a VM can you send me your public key10:43
afazekasderekh, sent10:47
derekhgot it, will send you details in a few minutes10:48
*** lucas-dinner is now known as lucasagomes11:09
*** mburned is now known as mburned_out11:11
*** shardy_ has joined #tripleo11:30
*** shardy has quit IRC11:33
*** xinwu has quit IRC11:42
*** spredzy has quit IRC11:44
*** spredzy has joined #tripleo11:44
*** mburned_out is now known as mburned11:58
*** panda_ has quit IRC12:05
*** panda_ has joined #tripleo12:06
*** bvandenh has quit IRC12:13
derekhshardy_: gfidente this unblocked ci https://review.openstack.org/#/c/268955/112:14
gfidentederekh, it also doesn't seem to have slown down the deployment12:17
derekhgfidente: yup, it seems to be fast then usual, the question is will it still be fast when we're running lots of jobs and the CPU on the host is more over commited12:20
derekh*faster12:21
gfidenteah right cause it's wasting more I/O you mean?12:23
*** bvandenh has joined #tripleo12:28
*** mgould has quit IRC12:34
*** chlong has joined #tripleo12:35
*** electrofelix has joined #tripleo12:37
*** thrash|g0ne is now known as thrash12:46
openstackgerritLennart Regebro proposed openstack/python-tripleoclient: More tests for overcloud deploy  https://review.openstack.org/25476712:50
openstackgerritMerged openstack-infra/tripleo-ci: Set DIB_NO_TMPFS=1  https://review.openstack.org/26895512:50
*** jdob has joined #tripleo12:53
shardy_derekh: nice!  That does seem to be a nice improvement in terms of speed12:55
shardy_derekh: can we tune things to queue jobs vs making individual jobs slow down and potentially time out?12:55
shardy_e.g by looking at the host load average or something?12:56
derekhshardy_: don't think so the number of jobs run at any one time is in a infra repository , it can't be dynamicly tweaked (if thats what you mean)12:57
*** dprince has joined #tripleo13:04
*** tzumainn has joined #tripleo13:06
shardy_derekh: Ah, yeah I was hoping it was something we could control dynamically & start queueing more stuff when the host is overloaded13:12
shardy_I assume it's something we can set manually though based on knowledge of the number of concurrent jobs before the host gets too heavily loaded?13:13
*** morazi has joined #tripleo13:14
*** lblanchard has joined #tripleo13:15
*** rcernin has quit IRC13:19
*** shardy_ has quit IRC13:23
*** shardy has joined #tripleo13:24
openstackgerritLennart Regebro proposed openstack/python-tripleoclient: Add support for an answers file  https://review.openstack.org/24922213:26
*** trown|outttypeww is now known as trown13:29
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: Start RabbitMQ with -proto_dist inet6_tcp for compat with IPv6  https://review.openstack.org/26904113:29
*** weshay_xchat has joined #tripleo13:31
*** weshay_xchat is now known as weshay13:31
*** links has quit IRC13:33
jistrdtantsur: heya, the `<PROFILE>_profile:1,<PROFILE2>_profile:1` format isn't used for scheduling directly, but needs to be transformed into `profile=<PROFILE>` form by using `openstack overcloud profiles match` command, did i get it right?13:38
dtantsurjistr, yes, which is what my patch does13:39
jistrdtantsur: yeah, just wanted to make sure i understand it correctly :) thx13:41
*** jprovazn has quit IRC13:43
openstackgerritMartin Mágr proposed openstack/tripleo-heat-templates: Add Rabbit IPv6 only support  https://review.openstack.org/26905813:58
*** Goneri has joined #tripleo14:06
*** spredzy has quit IRC14:06
*** spredzy has joined #tripleo14:08
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Cinder settings  https://review.openstack.org/26906814:11
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Ceph settings  https://review.openstack.org/26906914:11
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Glance settings  https://review.openstack.org/26907014:11
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Heat settings  https://review.openstack.org/26907114:11
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Keystone settings  https://review.openstack.org/26907214:11
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Neutron settings  https://review.openstack.org/26907314:11
*** rlandy has joined #tripleo14:14
*** rbrady has joined #tripleo14:15
*** dtantsur is now known as dtantsur|brb14:17
*** jeh has joined #tripleo14:22
*** akuznetsov has joined #tripleo14:26
*** trozet has joined #tripleo14:26
*** julim has joined #tripleo14:26
*** lucasagomes is now known as lucas-hungry14:28
*** liverpooler has quit IRC14:37
*** akuznetsov has quit IRC14:39
*** pradk has joined #tripleo14:49
*** akuznetsov has joined #tripleo14:50
*** teju has quit IRC14:51
*** egafford has joined #tripleo14:52
jaosoriordprince: The cinder patch https://review.openstack.org/#/c/269068/1 in the overcloud volume already had the cinder::config resource14:57
dprincejaosorior: okay, then we should remove it there I think14:58
*** mgould has joined #tripleo14:58
jaosoriordprince: I'm referring to this https://review.openstack.org/#/c/269068/1/puppet/manifests/overcloud_volume.pp14:58
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Keystone settings  https://review.openstack.org/26907214:59
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Neutron settings  https://review.openstack.org/26907314:59
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Heat settings  https://review.openstack.org/26907114:59
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Glance settings  https://review.openstack.org/26907014:59
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Ceph settings  https://review.openstack.org/26906914:59
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Cinder settings  https://review.openstack.org/26906814:59
dprincejaosorior: thanks, fixed14:59
jaosoriordprince: Alright, the rest look good. Only detail is gfidente's comment on the keystone::config related patch15:01
*** akuznetsov has quit IRC15:03
dprincejaosorior: oh dear. So if a user sets eystone::config::keystone_config wouldn't it actually override what you are setting in https://review.openstack.org/#/c/267399/2/puppet/hieradata/controller.yaml15:03
dprincegfidente: ^^15:03
dprinceI think we might want to consider a case for potentially merging these...15:04
dprincelooks like we already use nova::config in a similar fashion as well15:04
*** ukalifon1 has quit IRC15:04
jaosoriordprince: The case is the same for cinder15:04
dprincejaosorior: perhaps on volume, but it wasn't on the controllers I think so you could use controller only settings there15:05
dprincejaosorior: I still think it might be reasoable to expose the ability to set ad-hoc settings15:05
jaosoriorit's not on controllers, you're right15:05
dprincejaosorior: just that we might want to document the expected behavior with regards to it potentially overritting stuff that was getting automatically set15:06
jaosoriordprince: The other option would be to remove those settings in hieradata, and pass them to the puppet manifests (overcloud_*.pp), not as <service>::config resources15:07
jaosoriorbut as <service>_config resources, which is a provider object15:07
jaosoriordprince: Such as this one https://github.com/openstack/puppet-keystone/blob/master/lib/puppet/provider/keystone_config/ini_setting.rb15:08
gfidentedprince, for keystone I think we can use the submission from jaosorior15:08
gfidentewe were talking about it the other day cause we were doing it already for cinder on the cinderstorage nodes but not on the controller nodes15:08
gfidentebetter later than never!15:09
dprincegfidente: understood, but if we do that and I set a custom value for keystone::config::keystone_config would it overwrite what is already getting set in https://review.openstack.org/#/c/267399/2/puppet/hieradata/controller.yaml15:09
dprincegfidente: like if I set it with ExtraData...15:09
jaosoriorI think that'll be the case15:09
gfidenteI think that depends on the hierarchy no?15:10
dprincegfidente: extradata comes first15:10
jaosoriordoesn't ExtraData come later? So it would take that one?15:10
gfidentemaybe we have to tweak that too>15:10
jaosorioraaah15:10
jaosoriordamn15:10
dprincegfidente: I don't think this is a hiera ordering issue15:10
dprincegfidente: it is a hiera merging issue15:10
gfidenteI remember some hiera_array which returns all values of a given key15:11
jaosoriordprince, gfidente: Well, an easy way to solve it would be to replace those values in hiera, with code in puppet that would use the <service>_config providers, instead of the <service>::config resources. The downside being that the puppet code gets more cluttered :/15:12
dprincegfidente: I think hiera has some merging configs which would allow us to merge these hashes15:12
dprincegfidente: just saying we need to be careful here because we are using the ::config classes internally as well. Which is fine... I like the stylistic benefits of it15:12
dprincejaosorior: yeah, that essentially undoes what your keystone patch is doing15:13
dprincejaosorior: I'm fine w/ what you are doing I think15:13
jaosoriordprince: Pretty much. I would need to refactor that patch, and we would need to refactor other classes15:13
jaosoriordprince: Alright15:13
dprincejaosorior: and I don't think exposing the new ::config modules is cause for concern, but we might need to document the behavoir and perhaps make sure hiera merging is behaving like we want it to15:14
gfidentehiera_hash https://docs.puppetlabs.com/references/latest/function.html#hierahash15:15
gfidentesays :merge_behaviour can be controlled as pointed by dprince15:15
gfidente:merge_behavior: deeper seems to do it?15:16
dprincegfidente: yep, I think so. No 'u' in that behavior BTW :)15:17
gfidenteyes, sorry :015:17
gfidente;)15:17
*** gchamoul is now known as gchamoul|afk15:18
dprinceEmilienM: hi, we are having a discussion about using the various ::config classes (cinder::config for example). Would setting our :merge_behavior to 'deeper' make sense to allow end users to inject custom settings without overriding ones that we possibly set inside of our modules as well via Hiera.15:19
dprinceEmilienM: or is the default 'native' setting what we actually want to stick with?15:20
EmilienMit would make sense, but I personally would try to use ::config thing the less possible15:20
EmilienMusing ::config should be workarounds (temporary) until we have the actual parameter in the classes15:20
dprinceEmilienM: agree. Internally it is cleaner to use class parameters I think15:20
dprinceEmilienM: and then ::config can be for end users15:21
dprinceEmilienM: to set ad-hoc things only15:21
EmilienMI'm afraid if we offer ::config thing too much, people will just use it without patching our actual interface to parametrize it15:21
EmilienMdprince: we're on the same page15:21
EmilienMdprince: +1 for ad-hoc15:21
dprinceEmilienM: so question is, if we do this (temporarily until we add these options to puppet-keystone) https://review.openstack.org/#/c/267399/2/puppet/hieradata/controller.yaml15:23
dprinceEmilienM: what happens if a TripleO users specifies a custom keystone::config::keystone_config setting via ExtraConfig then?15:23
EmilienMdprince: dup resource15:24
EmilienMbecause the resource will be twice in the same catalog15:24
EmilienMand boom15:24
dprinceEmilienM: oh, but the user is setting different keys15:24
dprinceEmilienM: will they merge?15:24
EmilienMdprince: no15:24
EmilienMdprince: dup resource15:24
EmilienMwait15:25
EmilienMdifferent key or value?15:25
EmilienMkey == parameter to me so if the user adds keystone::config::keystone_config in ExtraConfig for another param, it will work15:25
dprinceEmilienM: different keys and values15:25
EmilienMexcept if they try to set a param that is already set in Hieradata15:25
dprinceEmilienM: I get the duplicate resource thing15:25
dprinceEmilienM: If that is the behaviour I think it is actually correct15:26
EmilienMdprince: it should work15:26
dprincegfidente: Yeah, lets just go with this. I think we are all set. I will delete my puppet-keystone patch in favor of jaosorior's15:27
jaosoriordprince: Alrighty15:27
*** lucas-hungry is now known as lucasagomes15:28
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: puppet: allow config of ad-hoc Neutron settings  https://review.openstack.org/26907315:28
gfidentedprince, thanks15:28
dprinceEmilienM: thanks too15:30
EmilienMdprince: cool15:30
gfidentehey guys can I ask something completely ot15:30
gfidentecause I am checking breaking bad15:31
*** chlong has quit IRC15:31
gfidentethe guy who always says yo , jesse pinkman15:31
gfidentewhich part of us is that?15:31
gfidenteUS15:31
dprincegfidente: sorry man, don't watch that show15:32
jaosoriorgfidente: no clue dude :/15:38
openstackgerritBen Nemec proposed openstack/instack-undercloud: Remove answers file support  https://review.openstack.org/26914715:40
bnemecAn even 40 open patches!15:41
* bnemec sees rebase hell in his future15:41
*** ukalifon1 has joined #tripleo15:44
*** gchamoul|afk is now known as gchamoul15:45
*** akuznetsov has joined #tripleo15:50
*** jprovazn has joined #tripleo15:55
*** rhallisey has joined #tripleo15:59
*** links has joined #tripleo16:00
*** lazy_prince has joined #tripleo16:01
*** dtantsur|brb is now known as dtantsur16:02
*** akuznetsov has quit IRC16:04
dtantsurdprince, or someone: could you please release DIB with the inspector gate fix?16:04
*** panda_ has quit IRC16:06
jaosoriorbnemec: ouch :/16:06
*** panda_ has joined #tripleo16:06
*** mkovacik has quit IRC16:09
*** aufi has quit IRC16:09
*** Slower has joined #tripleo16:10
openstackgerritMartin Mágr proposed openstack/tripleo-heat-templates: Add Rabbit IPv6 only support  https://review.openstack.org/26905816:13
*** lazy_prince has quit IRC16:22
*** lazy_prince has joined #tripleo16:22
openstackgerritMerged openstack/tripleo-docs: Replace doc for overcloud validate with manually running Tempest  https://review.openstack.org/22166416:22
*** links has quit IRC16:23
*** lazy_prince has quit IRC16:25
*** bnemec has quit IRC16:29
*** stendulker has joined #tripleo16:29
*** jaosorior has quit IRC16:34
*** jaosorior has joined #tripleo16:35
*** eggmaster has quit IRC16:35
*** bnemec has joined #tripleo16:37
bnemecderekh: So it looks like we're still having memory issues on the undercloud.  I just saw a bunch of messages like this during an image build:16:42
bnemecInstalling : gcc-4.8.5-4.el7.x86_64                                     57/6516:42
bnemecerror: Couldn't fork %post(gcc-4.8.5-4.el7.x86_64): Cannot allocate memory16:42
derekhbnemec: was that this morning? if so this fixed them https://review.openstack.org/#/c/268955/16:43
*** akuznetsov has joined #tripleo16:44
bnemecderekh: Ah, yes.  I guess I should have set --min-tmpfs for the ramdisk build too. :-/16:44
bnemecI should have refreshed the status page.  Now I see a bunch of successful runs. :-)16:45
*** jaosorior has quit IRC16:46
*** jaosorior has joined #tripleo16:47
-openstackstatus- NOTICE: Gerrit is restarting quickly as a workaround for performance degradation16:50
*** julim has quit IRC16:54
*** anande has joined #tripleo16:57
*** julim has joined #tripleo16:57
*** anande has quit IRC16:58
*** bandini has quit IRC17:01
derekhbnemec: ahh, that explains why it didn't work, I hadn't realized you didn't set it for both image builds17:06
derekhundercloud in CI is now set to 2xCPU and 5G RAM17:06
derekhAre the min reqs in a doc somewhere ? we should up date them too17:07
*** bandini has joined #tripleo17:08
bnemecderekh: Yeah, I'll push a patch to apply that to the agent ramdisk too.  It's a nasty problem to run into if you're not familiar with how DIB works.17:09
bnemecderekh: The baremetal requirements for the undercloud are probably the thing that needs to be updated: http://docs.openstack.org/developer/tripleo-docs/environments/environments.html#baremetal-environment17:10
bnemecAlthough the virt ones will need to be bumped as well if we increase the default undercloud size.17:10
bnemecWe should probably make that requirement an even 16 GB anyway: http://docs.openstack.org/developer/tripleo-docs/environments/environments.html#virtual-environment17:10
*** akuznetsov has quit IRC17:13
slaglebut then i won't be able to run tripleo anymore17:15
openstackgerritBen Nemec proposed openstack/python-tripleoclient: Apply --min-tmpfs 5 to agent ramdisk build too  https://review.openstack.org/26919217:15
bnemecslagle: Did your mini-Dell lose a stick of ram or something? :-)17:15
ryansbtime to buy bigger iron17:18
slaglebnemec: no, i'll just need to plug it back in :)17:19
*** bandini has quit IRC17:19
trownslagle: bnemec, ya I have been running a 16GB undercloud on my dell mini... it works much better17:21
*** athomas has quit IRC17:21
trowntempest smoke actually runs in a reasonable amount of time ~10min17:22
bnemecslagle: Fair enough. :-)17:22
trownalthough that is probably more about 4 cpus17:22
bnemectrown: You have another box too then?17:22
bnemecOr does your mini have more memory than mine?17:23
trownbnemec: mine has 32GB17:23
*** devvesa has quit IRC17:23
*** bandini has joined #tripleo17:23
trownand I just overcommit by about 1.517:23
bnemectrown: Putting in my purchase order now. ;-)17:23
*** fgimenez has quit IRC17:24
*** bandini has quit IRC17:24
bnemecMine is in an OVB environment with a 32 GB box I built out of mostly old parts.17:24
trownbnemec: ya my dell mini runs the exact same job as centosci in half the time17:24
bnemecSo I have 48 GB total.  64 would be better though.17:25
trownya 64 would be ideal17:25
bnemecI've been lusting after an X99 setup for my standalone OpenStack too.  I keep running out of memory on that box, even though it has 32 GB.17:25
derekhbnemec: Have you come across problems with OVB with multiple compute nodes?17:26
bnemecI think X99 maxes out at 128, which would be fantastic for parallel testing in multiple environments at once.17:26
bnemecderekh: Not really.  The only major issue I had was MTU settings, but that's not OVB-specific.17:27
bnemecIn fact, I just saw there's a thread about it on -dev.17:27
derekhbnemec: thanks will take a look, my OVB cloud has 2 compute nodes, thanis work if the bm nodes and the undercloud are on the same compute node only17:30
derekhs/thanis/things17:30
*** bandini has joined #tripleo17:31
derekhsome traffic is getting across compute nodes on the provisioning network so not exactly sure whats being blocked/lost17:31
*** athomas has joined #tripleo17:32
*** mbound has quit IRC17:32
*** trown is now known as trown|lunch17:33
bnemecderekh: Do both nodes have the noop firewall and prevent_arp_spoofing settings?17:39
bnemecMTU is also a possible culprit.  Check that you can ssh between nodes on the different hosts.  If it hangs you likely have an MTU problem.17:39
derekhbnemec: both noth should have both of those settings, I'll check and poke at the MTU later17:40
openstackgerritBen Nemec proposed openstack/instack-undercloud: Allow setting the hostname during undercloud install  https://review.openstack.org/24945717:47
*** hewbrocca has quit IRC17:54
*** stendulker has quit IRC17:55
*** ukalifon1 has quit IRC17:58
*** jaosorior has quit IRC17:59
*** derekh has quit IRC18:03
openstackgerritBen Nemec proposed openstack/instack-undercloud: Add undercloud config wizard UI  https://review.openstack.org/23121218:04
openstackgerritBen Nemec proposed openstack/instack-undercloud: Add validation of network configuration opts  https://review.openstack.org/23121118:04
*** paramite has quit IRC18:08
*** eil397 has joined #tripleo18:09
*** electrofelix has quit IRC18:10
*** yamahata has joined #tripleo18:14
bnemecDB errors seem to be killing the ha job now. :-(18:16
*** jistr has quit IRC18:17
*** nico_auv has quit IRC18:21
*** trown|lunch is now known as trown18:22
*** dtantsur is now known as dtantsur|afk18:22
*** davidlenwell has quit IRC18:22
egaffordHi all. Trying to reproduce http://logs.openstack.org/40/233240/7/check-tripleo/gate-tripleo-ci-f22-ha/b9fa6c0/console.html.gz (failure on the HA pipeline, which looks very much like a legitimate failure caused by my changes on trove-manage db_sync.) I'm unable to reproduce it using tripleo.sh on my instack-virt-setup env. Thoughts on how I might reproduce this / what kind of env or config I'd need to do so?18:23
*** david-lyle has quit IRC18:23
*** davidlenwell has joined #tripleo18:24
*** ChanServ sets mode: +v davidlenwell18:24
*** rwsu has joined #tripleo18:27
*** akuznetsov has joined #tripleo18:30
openstackgerritMiles Gould proposed openstack/python-tripleoclient: Use Ironic API v1.11 to support ENROLL state  https://review.openstack.org/26383118:33
egaffordshardy: ^(01:23:09) any tips on how to repro HA pipeline failures in tripleo-heat-templates?18:35
openstackgerritMiles Gould proposed openstack/python-tripleoclient: Use Ironic API v1.11 to support ENROLL state  https://review.openstack.org/26383118:35
gfidenteegafford, deploy passing -e tht/environments/puppet-pacemaker.yaml18:38
gfidenteegafford, https://github.com/openstack-infra/tripleo-ci/blob/master/toci_gate_test.sh#L4118:39
egaffordgfidente: Ack; many thanks.18:39
gfidenteyou don't need 3 controllers18:39
gfidentepuppet-pacemaker.yaml will use same code with 1 controller as well18:39
*** david-lyle has joined #tripleo18:39
egaffordgfidente: Useful; thanks.18:40
*** david-lyle has quit IRC18:46
*** mgould is now known as mgould|afk18:46
*** rebrego has joined #tripleo18:47
*** regebro has quit IRC18:47
*** rebrego is now known as regebro18:47
*** dmsimard is now known as dmsimard|afk18:47
*** akuznetsov has quit IRC18:48
*** Marga_ has joined #tripleo18:56
*** Marga_ has quit IRC18:57
*** Marga_ has joined #tripleo18:57
*** jcoufal has quit IRC19:03
bnemecegafford: I wouldn't be so quick to blame your patch.  Quite a few of the spurious ha failures in CI right now are db-related.19:04
bnemecI do see some ha jobs have passed recently though, so it might be worth a recheck if you haven't already.19:04
egaffordbnemec: I have enqueued a recheck; thanks. We Shall See. :)19:04
openstackgerritMerged openstack/instack-undercloud: Bump up undercloud VM specs  https://review.openstack.org/24314819:05
*** david-lyle has joined #tripleo19:10
*** dshulyak_ has quit IRC19:12
*** lblanchard has quit IRC19:16
*** david-lyle_ has joined #tripleo19:28
*** lblanchard has joined #tripleo19:28
*** david-lyle has quit IRC19:31
*** Marga_ has quit IRC19:31
*** david-lyle_ has quit IRC19:37
*** david-lyle has joined #tripleo19:39
*** dmsimard|afk is now known as dmsimard19:40
*** nico_auv has joined #tripleo19:42
*** gfidente is now known as gfidente|afk19:43
*** mkovacik has joined #tripleo19:46
*** david-lyle has quit IRC19:46
*** ifarkas has quit IRC19:50
openstackgerritMerged openstack/tripleo-heat-templates: Enable keystone handling of X-Forwarded-Proto header  https://review.openstack.org/26739919:53
*** egafford has quit IRC20:00
openstackgerritJames Slagle proposed openstack/os-cloud-config: Add []'s around ipv6 addresses when creating url's  https://review.openstack.org/26925420:01
*** panda_ has quit IRC20:06
*** panda_ has joined #tripleo20:06
*** ukalifon1 has joined #tripleo20:09
*** ukalifon1 has left #tripleo20:09
*** eil397 has quit IRC20:13
openstackgerritJames Slagle proposed openstack/os-cloud-config: Add []'s around ipv6 addresses when creating url's  https://review.openstack.org/26925420:15
*** xinwu has joined #tripleo20:17
*** dprince has quit IRC20:17
*** nico_auv has quit IRC20:19
openstackgerritJames Slagle proposed openstack/os-cloud-config: Add []'s around ipv6 addresses when creating url's  https://review.openstack.org/26925420:20
openstackgerritClark Boylan proposed openstack/diskimage-builder: Actually prune stale branches in src-repos  https://review.openstack.org/26824120:21
*** liverpooler has joined #tripleo20:28
*** lucasagomes is now known as lucas-dinner20:39
*** weshay has quit IRC20:40
*** Goneri has quit IRC20:52
*** weshay has joined #tripleo20:55
*** eil397 has joined #tripleo21:02
*** eil397 has quit IRC21:04
*** eil397 has joined #tripleo21:05
*** Marga_ has joined #tripleo21:20
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: Update yaml-validate.py to accept files or directories  https://review.openstack.org/26928121:24
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: Add simple parameter test to yaml-validate.py  https://review.openstack.org/26928221:24
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: Remove unused parameters from controller template  https://review.openstack.org/26928321:24
*** Marga_ has quit IRC21:25
*** mbound has joined #tripleo21:26
*** olap_ has quit IRC21:26
*** bvandenh has quit IRC21:28
*** shivrao has joined #tripleo21:29
*** Marga_ has joined #tripleo21:29
*** mcornea has quit IRC21:35
*** jprovazn has quit IRC21:37
*** shivrao has quit IRC21:45
*** Marga_ has quit IRC21:47
*** weshay has quit IRC21:49
*** jtomasek has quit IRC21:50
*** shardy has quit IRC21:51
*** rwsu has quit IRC21:55
*** akrivoka has quit IRC21:58
*** weshay has joined #tripleo22:03
*** chlong has joined #tripleo22:11
*** eil397 has quit IRC22:17
*** julim has quit IRC22:19
*** rwsu has joined #tripleo22:24
*** eil397 has joined #tripleo22:26
*** dsneddon is now known as dsneddon_biab22:38
*** thrash is now known as thrash|g0ne22:43
*** lblanchard has quit IRC22:45
*** weshay has quit IRC22:50
*** mbound has quit IRC22:57
*** eil397 has quit IRC23:04
openstackgerritMerged openstack/tripleo-heat-templates: Set the name property for all deployment resources  https://review.openstack.org/25564423:16
*** davidlenwell has quit IRC23:38
*** dsneddon_biab is now known as dsneddon23:38
openstackgerritBen Nemec proposed openstack/tripleo-heat-templates: Fix neutron-nova notifications  https://review.openstack.org/26937023:53
openstackgerritBen Nemec proposed openstack/tripleo-heat-templates: Fix tunnel_types hieradata on compute nodes  https://review.openstack.org/26937123:53
*** yuanying has quit IRC23:54
bnemectrown: You'll probably be interested in ^23:56
*** yuanying has joined #tripleo23:56
bnemecIt should allow us to turn on nova-neutron notifications and re-merge those reverted neutron patches and still have a working overcloud.23:57
*** davidlenwell has joined #tripleo23:58
*** ChanServ sets mode: +v davidlenwell23:58
« Sunday, 2016-01-17 Index Tuesday, 2016-01-19 »

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