Thursday, 2015-10-01

« Wednesday, 2015-09-30 Index Friday, 2015-10-02 »
*** shadower has quit IRC00:23
*** sdake has quit IRC00:23
*** shadower has joined #tripleo00:23
*** shivrao has quit IRC00:26
*** tiswanso has joined #tripleo00:27
openstackgerritMerged openstack/diskimage-builder: Use --nodeps when installing fedora-release  https://review.openstack.org/22712300:30
*** rlandy has quit IRC00:32
*** sthillma has quit IRC00:34
*** tiswanso has quit IRC00:46
*** tiswanso has joined #tripleo01:00
*** julim has quit IRC01:01
*** tiswanso_ has joined #tripleo01:02
*** tiswanso has quit IRC01:05
*** lazy_prince has joined #tripleo01:13
*** julim has joined #tripleo01:20
*** killer_prince has joined #tripleo01:39
*** lazy_prince has quit IRC01:42
*** rhallisey has quit IRC01:53
*** julim has quit IRC01:58
*** jerrygb has joined #tripleo02:09
*** yamahata has quit IRC02:26
*** weshay_xchat has quit IRC02:29
*** lazy_prince has joined #tripleo02:32
*** killer_prince has quit IRC02:32
*** shivrao has joined #tripleo02:38
openstackgerritSteve Baker proposed openstack/tripleo-heat-templates: Ensure present/latest for puppet driven package updates  https://review.openstack.org/19091802:39
openstackgerritSteve Baker proposed openstack/tripleo-heat-templates: Write package names out to flat files  https://review.openstack.org/19434802:39
openstackgerritSteve Baker proposed openstack/tripleo-heat-templates: Make package upgrade pacemaker-aware  https://review.openstack.org/22552402:39
openstackgerritSteve Baker proposed openstack/tripleo-heat-templates: yum_update.sh pacemaker and puppet fixes  https://review.openstack.org/22971002:39
openstackgerritSteve Baker proposed openstack/tripleo-heat-templates: Force stop a single node pacemaker on yum update  https://review.openstack.org/22971102:39
openstackgerritSteve Baker proposed openstack/tripleo-heat-templates: Force stop a single node pacemaker on yum update  https://review.openstack.org/22971102:41
openstackgerritSteve Baker proposed openstack/tripleo-heat-templates: Make package upgrade pacemaker-aware  https://review.openstack.org/22552402:41
*** shivrao has quit IRC02:42
*** tzumainn has quit IRC02:47
*** lazy_prince has quit IRC03:19
*** BharatK has joined #tripleo03:21
*** sthillma has joined #tripleo03:24
*** tiswanso_ has quit IRC03:24
*** Nisha has joined #tripleo03:25
*** sthillma_ has joined #tripleo03:26
*** sthillma has quit IRC03:29
*** sthillma_ is now known as sthillma03:29
*** sthillma has quit IRC03:40
*** lazy_prince has joined #tripleo03:51
*** links has joined #tripleo03:55
*** jerrygb has quit IRC03:57
*** jerrygb has joined #tripleo03:58
*** sdake has joined #tripleo04:00
*** Nisha has quit IRC04:01
*** sdake_ has joined #tripleo04:25
*** aukhan has joined #tripleo04:25
*** Nisha has joined #tripleo04:25
*** sdake has quit IRC04:28
*** jaosorior has joined #tripleo04:31
*** masco_ has joined #tripleo04:34
*** StevenK has quit IRC04:35
*** StevenK has joined #tripleo04:36
*** Marga_ has quit IRC04:47
*** BharatK has quit IRC04:48
*** jaosorior has quit IRC05:02
*** BharatK has joined #tripleo05:19
*** sdake_ is now known as sdake05:21
*** mgrohar has joined #tripleo05:40
*** mgrohar has quit IRC05:51
*** mgrohar has joined #tripleo06:00
*** leanderthal has joined #tripleo06:01
*** lsmola has joined #tripleo06:06
*** rasca has joined #tripleo06:08
*** jaosorior has joined #tripleo06:12
*** jaosorior has quit IRC06:12
*** jaosorior has joined #tripleo06:21
*** shardy has joined #tripleo06:25
*** mgrohar has quit IRC06:27
*** mgrohar has joined #tripleo06:27
*** jprovazn has joined #tripleo06:37
*** gfidente has quit IRC06:39
*** Marga_ has joined #tripleo06:39
*** mgrohar has quit IRC06:40
*** matbu_ has joined #tripleo06:40
*** matbu has quit IRC06:42
*** Marga_ has quit IRC06:47
*** olaph has quit IRC06:55
*** mcornea has joined #tripleo06:56
*** devvesa has joined #tripleo06:58
Nishapino|work, hello07:11
Nishapino|work, Could you review the patches ?07:11
*** aufi has joined #tripleo07:12
*** akrivoka has joined #tripleo07:16
*** penick has joined #tripleo07:17
*** penick has quit IRC07:24
*** ifarkas has joined #tripleo07:25
*** bvandenh has joined #tripleo07:36
jprovaznshardy: morning07:37
*** jtomasek has joined #tripleo07:39
*** jaosorior has quit IRC07:42
*** Marga_ has joined #tripleo07:44
*** jaosorior has joined #tripleo07:45
*** mgrohar has joined #tripleo07:45
*** openstackgerrit has quit IRC07:46
*** openstackgerrit has joined #tripleo07:46
*** Marga_ has quit IRC07:48
*** pinotree has joined #tripleo07:51
*** pinotree has quit IRC07:51
*** pinotree has joined #tripleo07:51
*** pinotree has quit IRC07:51
*** shivrao has joined #tripleo07:52
*** tosky has joined #tripleo07:55
*** rlandy has joined #tripleo07:56
shardyjprovazn: hi!08:05
*** shivrao has quit IRC08:05
*** bvandenh has quit IRC08:05
*** bvandenh has joined #tripleo08:05
pino|workNisha: note i still receive notifications via email... also, these reviews have been tiring, a bit more of testing and reading/understanding of people's review could have helped08:06
*** Nisha has quit IRC08:06
*** olaph has joined #tripleo08:06
jprovaznshardy: hi, do you want me to rebase https://review.openstack.org/#/c/227225/ and update the resource name or are you planning to make this change?08:08
shardyjprovazn: Hey, I actually already did locally, but I haven't quite finished the subsequent patch yet08:08
*** derekh has joined #tripleo08:08
shardyjprovazn: I can push what I have, more coming soon08:09
jprovaznshardy: ha, awesome! thanks08:09
openstackgerritJuan Antonio Osorio Robles proposed openstack/python-tripleoclient: Add debug environment to tox  https://review.openstack.org/22976808:12
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: Wire in NodeExtraConfig interface  https://review.openstack.org/22722508:13
shardyjprovazn: ^^08:13
* jprovazn looks08:13
shardyjprovazn: I was thinking about zaneb's comment re SoftwareComponent in the subsequent review08:13
shardythat could be neater, do you think it's OK to move both registration and unregistration inside the scaled-unit, e.g to the "pre deploy" stage?08:14
shardyI think it's probably OK provided it happens after the NetworkDeployment, and I guess it will help if you e.g want a thin image where RPMs are installed by puppet at deployment time from satellite08:15
shardywhich wouldn't work if registration always happens PostDeploy08:15
jprovaznshardy: IIRC my concern was mainly potential issue with network not being set, but if network is already set at that point, I *think* it's fine08:16
shardyjprovazn: Ok, I'll wire it in like that and we can see how it goes08:16
*** jistr has joined #tripleo08:18
*** tserong has quit IRC08:19
jprovazncool08:20
*** tserong has joined #tripleo08:20
*** tserong has joined #tripleo08:20
openstackgerritPino Toscano proposed openstack/python-tripleoclient: image build: add support for virt-dib  https://review.openstack.org/22254108:27
openstackgerritYanis Guenane proposed openstack/tripleo-puppet-elements: Add puppet-logrotate and puppet-rsyslog to the list of modules  https://review.openstack.org/22786608:36
*** paramite has joined #tripleo08:37
*** killer_prince has joined #tripleo08:37
openstackgerritDerek Higgins proposed openstack/tripleo-common: tripleo.sh Add option to locally build delorean package  https://review.openstack.org/22815708:38
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Use tripleo.sh to build packages  https://review.openstack.org/22942508:38
*** lazy_prince has quit IRC08:41
*** Marga_ has joined #tripleo08:45
*** Marga_ has quit IRC08:50
*** BharatK has quit IRC08:51
*** killer_prince has quit IRC08:56
*** lazy_prince has joined #tripleo08:56
openstackgerritMerged openstack/tripleo-common: Remove support for Tuskar  https://review.openstack.org/22352708:56
*** gfidente has joined #tripleo08:57
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Use tripleo.sh to build packages  https://review.openstack.org/22942508:58
*** jaosorior has quit IRC08:58
openstackgerritPino Toscano proposed openstack/python-tripleoclient: image build: add --builder-extra-args  https://review.openstack.org/22977809:00
*** BharatK has joined #tripleo09:04
openstackgerritMerged openstack/python-tripleoclient: Actually call _validate_args to check for network types mismatch  https://review.openstack.org/22132109:05
openstackgerritIan Wienand proposed openstack/diskimage-builder: Add flake8 to requirements  https://review.openstack.org/22978009:10
derekhshardy: sorry, this is going to be a vague question09:11
shardylol09:12
derekhshardy: there was a issue mentioned a while back when stack deployments were not continuing, as if heat didn't reciever a signal for one of the steps09:12
derekhshardy: that that ring any bells ?09:13
*** shardy_ has joined #tripleo09:13
shardy_derekh: Yeah, there were some internal issues with software deployment signalling (race condition)09:13
shardy_derekh: AFAIK they are all fixed now, let me get the bug, sec09:13
derekhshardy_: cool, ya thats what I'm looking for, thanks09:14
*** shardy has quit IRC09:14
derekhshardy_: I've gotten fixes for the HA job, that gets it back to somtimes passing09:15
derekhshardy_: the times it doesn't pass, heat returns a http 400 to one of the signals during a step and everything stops09:16
derekhshardy_: so I'm just wondering if we have those fixes in the version of heat we are using09:16
*** shardy_ has quit IRC09:18
*** shardy has joined #tripleo09:19
*** Nisha has joined #tripleo09:21
openstackgerritJan Provaznik proposed openstack/tripleo-common: Fix wrong node count when deleting specific nodes  https://review.openstack.org/22668209:21
*** mgrohar has quit IRC09:22
*** andrearosa has joined #tripleo09:23
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Test a update in trunk repository version  https://review.openstack.org/22978909:24
openstackgerritJan Provaznik proposed openstack/tripleo-specs: Spec for running overcloud nodes reboot  https://review.openstack.org/22448109:25
pino|workd0ugal: hi, do you know what's going on in https://review.openstack.org/#/c/229778/ wrt py34?09:28
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: Move RHEL (un)registration to NodeExtraConfig  https://review.openstack.org/22722609:29
shardyjprovazn: ^^ That just moves the registration/unregistration to NodeExtraConfig09:29
shardyI'll do a follow-up patch combining the configs into a SoftwareComponent09:29
d0ugalpino|work: Not sure, I seen that before on another review and it only happened once09:29
jprovaznshardy: cool, thanks, /me tries09:29
d0ugalpino|work: so hopefully a recheck will solve it :)09:29
pino|workd0ugal: your recheck failed the same way :/09:30
shardyjprovazn: I can't test the actual registration, but I stubbed the scripts with an echo/exit and that works OK09:30
d0ugalpino|work: Really< it doesn't look like it has finished to me09:30
d0ugalpino|work: or I can't see the result09:30
d0ugalpino|work: but in that case, I don't know. sorry09:30
*** andrearosa has quit IRC09:30
pino|workd0ugal: https://jenkins03.openstack.org/job/gate-python-tripleoclient-python34/4/09:30
pino|workoh np, just wondering whether was something known09:30
jprovaznshardy: ah, great, good to know09:31
jprovaznmtg, bbl09:31
*** jtomasek has quit IRC09:35
*** bvandenh has quit IRC09:37
*** sdake_ has joined #tripleo09:38
*** pelix has joined #tripleo09:41
shardyDoes anyone want to review https://review.openstack.org/#/q/status:open+project:openstack/tripleo-heat-templates+branch:master+topic:bug/1463092,n,z before it lands?09:41
*** sdake has quit IRC09:42
shardythere are three patches (puppet & update related) that look ready to land and are passing CI09:42
* shardy will approve in a bit if nobody objects09:42
*** Marga_ has joined #tripleo09:46
openstackgerritDerek Higgins proposed openstack/tripleo-common: tripleo.sh Add option to locally build delorean package  https://review.openstack.org/22815709:46
*** bvandenh has joined #tripleo09:49
*** Marga_ has quit IRC09:51
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Nothing to see here  https://review.openstack.org/11101109:51
*** jtomasek has joined #tripleo09:51
*** jaosorior has joined #tripleo09:52
*** mcornea has quit IRC09:53
*** jaosorior has quit IRC09:54
*** jaosorior has joined #tripleo09:54
*** jaosorior has quit IRC09:54
*** jaosorior has joined #tripleo09:55
*** jtomasek has quit IRC09:57
*** mcornea has joined #tripleo10:06
spredzyI am trying to run tripleo.sh, but it fails when doing so os-net-config setup. It is looking for ethX when nics are named enpXsY. Anyone already had that ?10:16
spredzyshardy, ^10:16
shardyspredzy: are you installing the undercloud on baremetal?10:17
spredzythe tripleo.sh is ran on a baremetal, I ran it with --all option10:17
spredzyshardy,10:17
shardyspredzy: I think for baremetal you need to configure the undercloud.conf local_interface10:19
shardyso you may need to run each step (not --all) and do the undercloud install manually10:19
shardyIIRC the undercloud defaults work for the instack-virt-setup VMs, but we don't attempt to detect baremetal nic names10:19
shardyspredzy: also FYI last time I tried the introspection wasn't working, that may be fixed now tho10:20
spredzyshardy, ack, will run the step one by one. Thanks for the info10:21
shardyspredzy: we should probably add an UNDERCLOUD_LOCAL_INTERFACE variable to the script I guess10:21
openstackgerritMerged openstack/tripleo-heat-templates: Write package names out to flat files  https://review.openstack.org/19434810:22
spredzyshardy, IMO one should be able to get the script run it and it should work out of the box with the default values, then if neede tweak it at which10:22
spredzywish*10:22
*** hashar has joined #tripleo10:23
shardyspredzy: well for VMs it does, but it's probably quite hard to automatically detect which nic to use on baremetal10:32
shardye.g if you choose the wrong one you can hose the local network w/dhcp, or take down the primary nic connecting to the local network, etc10:33
shardyspredzy: maybe we should detect that we don't know and prompt the user tho10:33
shardyOr, at least fail when the default nic doesn't even exist ;)10:33
shardyhttps://review.openstack.org/#/c/225096/8/scripts/tripleo.sh10:33
shardyI made similar comments here, FWIW10:34
*** paramite is now known as paramite|afk10:35
*** andrearosa has joined #tripleo10:37
openstackgerritJuan Antonio Osorio Robles proposed openstack/python-tripleoclient: Remove unused cli arguments from overcloud deploy  https://review.openstack.org/22981510:37
*** andrearosa has left #tripleo10:38
*** paramite|afk is now known as paramite10:39
*** Nisha_away has joined #tripleo10:40
spredzyshardy, agree, if one has a box for different purpose one of them being tripleo, it might be a bad idea to auto-guess the NIC10:40
*** Nisha has quit IRC10:40
*** BharatK has quit IRC10:42
openstackgerritMartin Mágr proposed openstack/tripleo-heat-templates: Keystone domain for Heat  https://review.openstack.org/18056610:44
*** Nisha_away has quit IRC10:44
*** Nisha_away has joined #tripleo10:45
*** Marga_ has joined #tripleo10:46
*** Marga_ has quit IRC10:51
openstackgerritGael Chamoulaud proposed openstack/tripleo-heat-templates: Make puppet manifests compliant with Puppet 4.x  https://review.openstack.org/22566410:52
*** jaosorior has quit IRC10:54
*** BharatK has joined #tripleo10:55
*** Nisha_away has quit IRC10:57
*** trown|outttypeww is now known as trown11:00
*** paramite is now known as paramite|afk11:02
*** aukhan has quit IRC11:07
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Test a update in trunk repository version  https://review.openstack.org/22978911:12
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Remove sudo when calling tripleo.sh  https://review.openstack.org/22982311:12
derekhHmm, the repository in delorean-deps seems to have vanished, causing CI falures http://cbs.centos.org/repos/cloud7-openstack-common-testing/x86_64/os/Packages/11:21
trownruh roh... that happened over the weekend too11:24
*** paramite|afk is now known as paramite11:29
slagleoh geez11:32
slaglei'm eating my words when i was arguing for the reliability of the rdo infra11:33
derekhslagle: its not rdo infra, its cenots infra11:34
slagleoh right, thank goodness11:34
derekhslagle: ;-)11:34
derekhslagle: looks like when they generate a new repo (the one specified in delorean-deps), its not  an atomic operation11:35
derekhslagle: so while it happens there is now repo ....11:35
slaglei'd like to find a way to generate the commands in the docs from the tripleo.sh code11:38
slaglewas thinking about using the same awk stuff we have in incubator11:39
derekhslagle: easy     $ cat tripleo.sh11:39
slagleexcept i dont want all the comments and actual documentation to live in tripleo.sh11:39
slagleso i was thinking a way to mark just the line(s) in tripleo.sh that "install the undercloud", then parse that out somehow, and add the result into the tripleo-docs rst11:40
derekhslagle: some kind of tagged markers11:41
slagleright, something like that11:41
derekhslagle: sounds good11:42
slaglei'm going to add this bit to the trello card11:45
slaglethis feels like concensus!11:45
*** dprince has joined #tripleo11:47
*** dtantsur|afk is now known as dtantsur11:47
*** Marga_ has joined #tripleo11:47
*** Nisha has joined #tripleo11:50
*** jtomasek has joined #tripleo11:51
*** Marga_ has quit IRC11:52
openstackgerritMerged openstack/instack-undercloud: Add basic API docs  https://review.openstack.org/22381011:57
derekhOk, so the HA job has 3 problems I know of the fix for the first is here https://review.openstack.org/#/c/229183/ once thats merged I'll track down the 2nd12:00
derekhwhich I believe to be something todo with https://github.com/puppetlabs/puppetlabs-mysql/commit/97b8200a5f93d1a6853a8994229183706e6e989e12:00
derekhWith those two fixed the HA job sometimes passes again (e.g. here https://jenkins05.openstack.org/job/gate-tripleo-ironic-overcloud-f21puppet-ha/174/)12:00
derekhI things the answer to the 3rd problem causing it to sometimes fail is somewhere in heat or our signaling mechanism12:01
trownI had noticed it passed a couple times, and was wondering how12:02
*** rhallisey has joined #tripleo12:03
*** rbrady has joined #tripleo12:04
*** marcusvrn_ has joined #tripleo12:06
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Nothing to see here  https://review.openstack.org/11101112:07
pino|work"nothing to see here" is the best way of saying "everybody, look HERE!" :)12:08
openstackgerritMerged openstack/os-cloud-config: Fix nodes registering with fake_pxe driver  https://review.openstack.org/22304612:08
openstackgerritJames Slagle proposed openstack/python-tripleoclient: Make ssh user configurable  https://review.openstack.org/22841312:10
openstackgerritMerged openstack/tripleo-heat-templates: Ensure present/latest for puppet driven package updates  https://review.openstack.org/19091812:10
*** pblaho has joined #tripleo12:11
openstackgerritMerged openstack/tripleo-heat-templates: Make package upgrade pacemaker-aware  https://review.openstack.org/22552412:11
*** jtomasek has quit IRC12:11
*** morazi has joined #tripleo12:12
*** tzumainn has joined #tripleo12:16
*** masco_ has quit IRC12:17
*** jerrygb has quit IRC12:17
*** jtomasek has joined #tripleo12:17
*** jaosorior has joined #tripleo12:18
*** jayg|g0n3 is now known as jayg12:20
openstackgerritJan Provaznik proposed openstack/python-tripleoclient: Change RHEL subscription template path  https://review.openstack.org/22984712:24
*** weshay_xchat has joined #tripleo12:25
*** weshay_xchat is now known as weshay12:27
*** pradk has joined #tripleo12:28
*** spzala has joined #tripleo12:31
openstackgerritMerged openstack/python-tripleoclient: Fix file checksum function in for python 3.4  https://review.openstack.org/22860812:32
*** mkovacik has joined #tripleo12:32
*** BharatK has quit IRC12:35
shardyzaneb: Hey, I pushed updates with the NodeExtraConfig rename we discussed yesterday12:45
shardyzaneb: I've not forgotten your comments re SoftwareComponent in https://review.openstack.org/#/c/227226/12:45
shardyI just decided it might be cleaner to handle that refactor in a subsequent patch12:45
zanebshardy: np, I think it's much easier to move that later than to rename stuff later12:48
*** Marga_ has joined #tripleo12:48
shardyzaneb: +1, I'll ping the ML about the more general rename later, e.g potentially removing the *Pre suffix12:48
*** BharatK has joined #tripleo12:49
shardyzaneb: we can probably deprecate them and leave some aliases to the old names in place, but it's unfortunate that there's no way to generate a user-visible warning12:49
shardymaybe we could add something to the oscplugin which warns for a couple of cycles, then remove the aliases12:49
zanebshardy: I actually don't think living with the ones we've got is a big problem12:50
zanebI was more worried about minimising any confusion with new ones at a stage where we still had time to fix them12:51
shardyzaneb: ack, I guess I was looking more for consistency, but perhaps we can have both in due course12:51
shardyzaneb: I'll work up a docs patch to at least explain what we have, as it's not all captured atm12:51
zaneb+1 great idea12:51
* zaneb was certainly blundering around cluelessly12:52
*** julim has joined #tripleo12:52
*** mgrohar has joined #tripleo12:53
*** Marga_ has quit IRC12:53
dprinceslagle, shardy: commented here https://review.openstack.org/#/c/222772/512:58
dprincecouple of thoughts on this (interesting) topic. First is would it make sense to put the 'fake server' resource into Heat itself somehow...12:59
dprincebecause a pre-existing server resource in Heat would expose new options for using Heat for configuring only13:00
shardydprince: probably not, because if it's an actual nova resource, we should do something like https://review.openstack.org/#/c/135492/ instead13:00
shardythe dummy server thing is really only needed when the server is something not known to openstack at all13:00
shardydprince: I think the external_id interface proposed by asalkeld pretty much does what you just described13:01
dprinceshardy: I don't know if I care about it being a Nova resource13:01
dprinceshardy: cool, this would be perhaps valuable if we go down this route13:02
shardydprince: In that case, heat doesn't have to care about it, you can just use an empty nested stack, as I did in the examples I provided for slagle13:02
dprinceshardy: yep, got it13:02
shardyhttps://etherpad.openstack.org/p/noop-softwareconfig13:02
dprincethe second idea is what if we architected our templates differently so that we always did this13:02
dprinceshardy: ^^13:03
shardydprince: Yeah, it's certainly worth thinking about, and it makes perfect sense for some things that can be independently created, e.g Ceph clusters13:03
shardyI'm less sure about having a series of separate stacks operating on the same servers, because you lose the dependency information Heat normally has between nested stacks13:04
shardycould be worth investigating though - I'm definitely +1 on anything which gets us incrementally away from the giant stack architecture13:04
dprinceshardy: here is the thing I actually like about that... it gives us options for refactoring in t-h-t13:04
dprinceshardy: our puppet manifests are mostly idempotent, or they should be13:05
dprinceshardy: so it doesn't really matter if I pull the rug on the the Heat config resources and re-run them13:05
dprinceshardy: if you buy that... then I think there could be serious merit to always managing our stacks this way13:06
shardydprince: I dunno, I guess I see it as decoupling something which is in reality still coupled, you'd need something orchestrating each heat stack (wrapper script) which would in the end do exactly what the parent stack already does13:07
*** Goneri has joined #tripleo13:07
shardyI think I'd need to see an example to form a proper opinion, but like I said, probably worth investigating :)13:07
slagledprince: i actually started hacking on a fake server resource in heat initially, until shardy showed me how easy this was without that13:09
dprinceshardy: ack, I may try this dual stack idea out13:09
slaglebut my goal here was to make it work on a server not known to nova. that being said, i think it could be adapted to work both ways. servers known to nova, but not orchestrated via heat, and servers not known to nova at all13:09
dprinceslagle: cool, just asking the question13:09
slaglesure. i'm actually super pleased with how well this is working13:10
dprinceslagle: cool. The idea I mentioned about was to use 2 stacks: one to create the resources, and another to configure them13:11
dprinceslagle: in the second stack (the configuration stack) we would essentially always use your new code13:11
slagleyea, makes sense13:11
*** athomas has quit IRC13:11
dprinceslagle: would be worth prototyping13:11
dprinceslagle: how it effects things like network isolation, etc would be potentially trick. But I actually think doing this would help the Spine Leaf architecture concerns too13:12
dprinceslagle: because the single stack... isn't going to handle that so well I'm afraid13:12
slaglejust b/c we don't always have a full provisioning network to all the nodes in spine/leaf?13:13
dprincespine leaf is like network isolation for each rack13:13
dprinceparameter explosion!!! (or obfuscation) if we do it in a single stack perhaps13:13
dprinceshardy: I also think the dual stack approach could have merit for other reasons as well13:15
*** athomas has joined #tripleo13:17
*** lazy_prince has quit IRC13:18
*** noslzzp has joined #tripleo13:18
jaygjistr: ping, qq - is the NeutronScale resource not included in tht on purpose?  if so, do you know the reason?  it is still in the ref arch is why I ask13:19
*** rpothier has joined #tripleo13:23
slagledprince: what did we end up deciding about requiring adding parameters at the top level? EmilienM has this patch: https://review.openstack.org/229471 and was asking if he needed to update overcloud-without-mergey.yaml as well. i don't recall what the decision here was13:24
dprinceslagle: I'm torn on those being top level13:25
*** Nisha has quit IRC13:25
*** mcornea has quit IRC13:25
shardyslagle: FWIW my understanding was that new parameters to the role templates, e.g puppet/controller.yaml should be surfaced via overcloud-without-mergepy, and the parameter_defaults should be used only for optional additions, e.g *ExtraConfig13:25
shardyslagle: Or, maybe that's just what I was thinking in my head ;)13:25
dprinceshardy: sort of, but if we do that we are going to get a slew of obscure top level parameters13:26
shardyIMHO it will get very confusing if we encourage paramter-defaults to be wired in to the "core" templates, vs for third-party additions13:26
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: Ensure access to RabbitMQ is password protected  https://review.openstack.org/21040213:26
dprinceif the paramter is in any sense implementation specific (related to puppet or docker only) they don't belong at the top level13:27
shardydprince: In the patch slagle referenced, couldn't those just be added to the static hieradata yaml?13:27
dprinceshardy: yes. I think so13:27
dprinceshardy: although that is arguably a bit of an advanced feature13:27
shardydprince: Well, it can be overriden easily enough via the controllerExtraConfig parameter?13:28
dprinceshardy: My only concern is having too many top level parameters13:28
shardydprince: one option would be to start making use of heat ParameterGroups13:28
dprinceshardy: okay, perhaps you can ask that question on the review13:28
shardythen parameters could be grouped and categorized as "core" and "advanced", or whatever13:29
dprinceshardy: I'd be fine using that for these sort of configurations... probably just need to document it somewhere in the "puppet specific" docs13:29
shardydprince: Yeah, I'm planning a docs patch which refreshes the various ExtraConfig (parameter and resource) interfaces13:29
dprinceshardy: yeah, shadower used to have a TODO in the templates to start using parameter groups at one point13:29
slagleshardy: i think i'm with you and i do remember agreeing that the role template parameters should be surfaced. that being said, we can always revisit13:29
shardywhat we have atm is a bit outdated now13:29
*** tiswanso has joined #tripleo13:29
*** jprovazn has quit IRC13:29
*** bvandenh has quit IRC13:29
*** lblanchard has joined #tripleo13:31
jistrjayg: i recall there were some issues with neutron service names and how they reported themselves to neutron API... could that be related?13:32
jistri don't recall exactly what the problem was though13:32
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: Parametrize RabbitMQ FD limit  https://review.openstack.org/20179613:32
dprinceshardy: ++ for Parameter groups at the top level13:33
jistrjayg: so i think it's quite possible that NeutronScale is not there on purpose13:33
jistrjayg: yup here is the removal of NeutronScale https://github.com/openstack/tripleo-heat-templates/commit/1ca7efc5d18c4b5ba5f53753dcfbbe790c42aaad13:35
*** trown is now known as trown|brb13:35
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: Ensure access to RabbitMQ is password protected  https://review.openstack.org/21040213:36
gfidentejayg, I think refarch was meant to support rhos6 where metadata was supposed to take over the hostname from the other hosts13:37
gfidentejayg, with rhos7 we didn't need to that13:37
*** egafford has joined #tripleo13:39
d0ugalslagle: For tripleo.sh, is the usage just clone the repo and run on the undercloud? or?13:40
derekhd0ugal: http://paste.openstack.org/show/475071/13:41
d0ugalakrivoka: ^13:41
d0ugalderekh: Thanks13:42
mgroharhey13:42
akrivokathanks13:42
mgroharcan anyone what is the minumum number of controllers for openstack HA in rhos713:42
mgroharis it 3x?13:43
akrivokaderekh: so in case of virt deployment, things like register nodes and overcloud deploy are also run from the virt host and not from the undercloud vm?13:43
slagled0ugal: yes. i plan on getting this documented at some point13:43
mgroharor do tripleo-heat-templates support 2x controllea HA in active-passive way?13:43
d0ugalslagle: Thanks13:44
slagled0ugal: but that's the idea anyway. it'll be for stuff that doesnt have to run on the undercloud as well, eventually13:44
slaglesuch as the virt setup, and shardy has patches up for the delorean package building too13:44
d0ugalNeat13:44
derekhakrivoka: I've been creating an undercloud vm with instack-virt-setup, then ssh to undercloud and run everything else from there13:46
akrivokaderekh: I see, thanks13:47
jaygjistr gfidente: ok, thanks, do you happen to know if this was discussed with fabio and andrew (and anyone from neutron team, since they added the scale service)?  if it is truly not needed anymore, then maybe we can get the ref arch updated to match reality13:47
gfidentewith amuller and fabbione yes13:47
gfidenteand marios actually :)13:47
*** trown|brb is now known as trown13:47
*** Marga_ has joined #tripleo13:49
jayggfidente: ok, cool, then maybe I'll check in with fabbione (he can notice here if around) to see if he thinks that NeutronScale is needed in the refarch docs anymore based on what you have already discussed13:51
gfidentejayg, yeah I am unsure what will happen with the refarch13:52
gfidentegiven I think there is desire to shift over a different process13:53
gfidentewhere bugs are filed as required when there is something to be changed13:53
*** Marga_ has quit IRC13:53
*** jprovazn has joined #tripleo13:54
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: Deprecate the undercloud templates.  https://review.openstack.org/22989313:54
*** links has quit IRC13:57
jayggfidente: ok, that seems reasonable, I just noticed the diff when comparing tht to refarch for a sprint task I am working on, wanted to see what the reasoning was14:03
dprinceshardy: are you good here now https://review.openstack.org/#/c/226981/14:09
dprinceshardy: if you have a moment you might also review this one https://review.openstack.org/#/c/226968/14:10
*** david-lyle has quit IRC14:11
shardydprince: minor issue with the default in 226981 otherwise lgtm14:12
dprinceshardy: is an empty list not valid then?14:13
*** david-lyle has joined #tripleo14:13
shardydprince: I assumed it wasn't, because the value parsed by that type is a comma delimited string14:13
shardylet me check, sec14:14
dprinceshardy: hmmm, it is valid Json. SO I would assume also a valid list...14:14
shardydprince: comma_delimited_list accepts a string, json accepts either a json string, or a yaml map/list14:14
dprinceshardy: I think this is why I liked json better :)14:17
openstackgerritDerek Higgins proposed openstack/tripleo-common: Fix repo whitelist for heat and ironic  https://review.openstack.org/22990614:18
derekhSo we havn't been pulling heat and ironic from the repo we thougt we were pulling it from ^14:19
shardydprince: Hmm, turns out I'm wrong, it actually works fine, apologies!14:19
dprinceshardy: I was gonna say, I think I tested this...14:19
openstackgerritMerged openstack/os-net-config: ifcfg: Add support for dns_servers  https://review.openstack.org/22696814:20
shardyhttps://github.com/openstack/heat/blob/master/heat/engine/parameters.py#L35914:21
*** mgrohar has quit IRC14:21
shardydprince: yeah, we actually look for a list, then fall back to splitting a string14:21
shardywhich is more permissive than I remembered ;)14:21
spredzyjistr, when running openstack overcloud image upload I have an error stating ironic-python-agent.initramfs does not exist, does that ring a bell to you ?14:23
jistrspredzy: yes14:23
jistrmmt14:23
jistrspredzy: https://github.com/rdo-management/inlunch/blob/62bd434381392814450e201203d77b4baf802ac1/answers.yml.example#L12414:24
jistrspredzy: you could also try the new tripleo.sh script i think. paramite have you had success with it in the end?14:25
jistrspredzy: tripleo.sh would do that automatically14:25
*** leanderthal has quit IRC14:27
spredzyjistr, thanks for the fix, I ran tripleo.sh this morning, but couldn't get what I needed out of it, need to restart it later today on the box14:27
openstackgerritRyan Brady proposed openstack/tripleo-common: [WIP]Adds initial swift integration for plans  https://review.openstack.org/22729714:28
*** paramite is now known as paramite|afk14:34
dprincemarios: are you also okay with this (similar to the instack change): https://review.openstack.org/#/c/226327/14:36
dprinceslagle: you are still cool here right? https://review.openstack.org/#/c/226981/14:36
mariosdprince: looking14:37
openstackgerritMerged openstack/puppet-tripleo: Automatically install packages when upgrading  https://review.openstack.org/22853214:38
openstackgerritMerged openstack/instack-undercloud: swift::storage::all deprecation cleanup  https://review.openstack.org/22631514:38
mariosdprince: not sure if you wana wait for ci on https://review.openstack.org/#/c/226327/1 but last run fail is not relevant afaik14:41
openstackgerritEthan Gafford proposed openstack/tripleo-heat-templates: Sahara Integration  https://review.openstack.org/22086314:41
mariosdprince: so i leave the recheck up to you14:41
dprincemarios: cool. thanks14:42
dprinceslagle: as I'm going through reviews I'm abandoning (and marking tickets as won't fix) for anything older than 6 months14:45
dprincethere aren't that many reviews to just do this manually14:45
openstackgerritMerged openstack/os-cloud-config: register-nodes: add support for pxe_ucs  https://review.openstack.org/21893914:45
openstackgerritMerged openstack/tripleo-heat-templates: swift::storage::all deprecation cleanup  https://review.openstack.org/22632714:46
*** Marga_ has joined #tripleo14:50
*** jtomasek has quit IRC14:50
slagledprince: wfm.14:51
spredzyjistr, bothering you again. Did you encouter this issue while running bulk start ? OperationalError: (OperationalError) no such table: nodes u'DELETE FROM nodes WHERE nodes.uuid = ?' (...74236a566ad3',)14:51
jistrspredzy: yes :)14:51
* spredzy smiled14:52
dprinceslagle: this again too https://review.openstack.org/#/c/226981/14:52
*** Marga_ has quit IRC14:54
openstackgerritMerged openstack/tripleo-specs: Spec for TripleO overcloud deployment library  https://review.openstack.org/21975414:58
openstackgerritMerged openstack/tripleo-heat-templates: Add DnsServers param to network config templates  https://review.openstack.org/22698114:59
*** aufi has quit IRC14:59
*** untriaged-bot has joined #tripleo15:00
untriaged-botUntriaged bugs so far:15:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/149929815:00
openstackLaunchpad bug 1499298 in tripleo "Installers should use pymysql db driver as of Liberty" [Undecided,New]15:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/150137815:00
openstackLaunchpad bug 1501378 in tripleo "pacemaker neutron resource constraints and cleanup" [Undecided,Triaged]15:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/150129615:00
openstackLaunchpad bug 1501296 in tripleo "Stack updates fail due to a new resource type" [Medium,New]15:00
*** untriaged-bot has quit IRC15:00
*** hashar has quit IRC15:00
*** paramite|afk is now known as paramite15:01
paramitejistr, "paramite have you had success with it in the end?" >>> not yet, introspection did not work and I'm trying to make rsync to work now15:03
jistrparamite: for introspection the thing i pasted yesterday should do the trick15:04
jistrspredzy: sorry! :)) i got carried away by another thing and totally forgot to paste it for you15:04
jistrspredzy: https://github.com/rdo-management/inlunch/blob/62bd434381392814450e201203d77b4baf802ac1/answers.yml.example#L106-L11015:04
spredzyjistr, I need to run that and rebuild the images correct ?15:05
paramitejistr, ack, it fixed the issue thx15:05
jistrspredzy: that's not necessary. just re-run the introspection after that15:06
*** sdake has joined #tripleo15:06
egaffordd0ugal: Responded to your -1 on https://review.openstack.org/#/c/221418/. It's possible I'm missing something (and if I am, please let me know,) but I think that the line you're objecting to is in keeping with handling of pwd fields for other services. Thoughts?15:08
*** mkovacik has quit IRC15:09
*** mkovacik has joined #tripleo15:09
dprinceslagle: was there a rush to get the tripleo-common spec in?15:09
d0ugalegafford: oh, so it is. sorry. They should probably all be removed at some point15:09
* dprince would rather see the docker specs land soon too15:10
egaffordd0ugal: NP; your comment totally makes sense.15:10
d0ugalegafford: I'll re-review now.15:10
*** sdake_ has quit IRC15:10
egaffordd0ugal: I fully support a later revision patch to remove the lot of them: they are effectively replacing unset with None.15:10
d0ugal+115:10
dprinceslagle: gotta say, with all the restructuring upstream it still feels a bit early to get the API spec right for tripleo-common. Not saying that work on it can't proceed, just that it is likely to change perhaps...15:11
*** tosky has quit IRC15:13
*** Marga_ has joined #tripleo15:14
*** social has joined #tripleo15:14
*** jerrygb has joined #tripleo15:19
*** jerrygb has quit IRC15:19
*** jerrygb has joined #tripleo15:19
*** shardy_ has joined #tripleo15:22
*** shardy has quit IRC15:24
*** shardy_ has quit IRC15:27
*** shardy has joined #tripleo15:28
* mkovacik is away: I'm busy15:32
*** tosky has joined #tripleo15:35
*** tosky has quit IRC15:35
*** tosky has joined #tripleo15:35
*** persia has quit IRC15:38
*** paramite has quit IRC15:44
*** Marga_ has quit IRC15:44
*** Marga_ has joined #tripleo15:45
*** alop has joined #tripleo15:49
*** tzumainn has quit IRC15:53
*** lazy_prince has joined #tripleo15:53
dprinceshardy: just one additional sentence here and I'm happy I think https://review.openstack.org/#/c/221811/516:01
Slowerdprince: are we back on fedora 22 again?16:03
dprinceSlower: I'm not, why would we?16:03
dprinceSlower: did I miss something?16:03
Slowerwtf16:03
* dprince answers Slower's question w/ 2 questions16:03
Slower52f27d73c61d        dprince/heat-docker-agents                    "/usr/bin/os-collect-"   5 minutes ago       Up 5 minutes                            heat-agents16:03
Slower[heat-admin@overcloud-novacompute-0 ~]$ sudo docker exec -ti heat-agents /bin/bash16:04
Slower[root@overcloud-novacompute-0 /]# cat /etc/redhat-release16:04
SlowerFedora release 22 (Twenty Two)16:04
Slowermaybe I messed something up..16:04
gfidenteI think the idea of naming 22 Twenty Two was the brightest naming ever16:04
gfidentemy 2 cents16:04
dprinceyou should be using dprince/heat-docker-agents-centos16:05
Slowerdprince: doh!16:05
Slowerwell isn't that grand.. how did that happen?16:05
Slowerok16:05
dprinceSlower: sounds like someone needs to work on rebase skills?16:05
Slowerheh, could be!16:06
dprinceSlower: to be fair this patch has evolved a good bit over the past 6 months or so16:06
SlowerI'm still on devtest too so..16:06
SlowerI have some manual changes to make to make it work16:06
*** jaosorior has quit IRC16:07
shardySlower: FYI there's a nice new script if you decide to build a new env:16:08
shardyhttps://github.com/openstack/tripleo-common/blob/master/scripts/tripleo.sh16:08
dprinceSlower: it is time to pay the price man16:08
Slowerugh, I hate prices.. :)16:08
Sloweror paying?16:09
Slowersomething.. :)16:09
dprinceSlower: devtest is banned in tokyo16:09
shardySlower: tl;dr - follow the tripleo-docs virt instructions until instack-virt-setup, ssh onto the undercloud, then run each step of tripleo.sh (the usage is in order)16:09
shardySlower: personally I skipped introspection but that may work now, not sure16:09
Slowershardy: ok cool16:09
slagledprince: i dont think it was a rush to get it in. but it's been around for a month.16:10
slagledprince: that's not necessarily long or short, as far as specs go16:10
dprinceslagle: fair, I'm just concerned it early for that yet16:10
*** killer_prince has joined #tripleo16:10
shardydprince: thanks, I'll update it - it's absolutely my intention that nothing changes except we support some extra branches16:12
shardyso "moving" was probably the wrong word to use there16:12
dprinceshardy: perhaps just me being defensive16:12
dprincerhallisey: what did the kernel panic man have to say today16:12
rhalliseydprince, I'm going to walk over to his desk today16:13
dprincerhallisey: you should take a posse16:13
slagledprince: i can agree that it might be too early to say we know what the implementation will be exactly, i think we got about specific as we felt comfortable with in the spec16:13
rhalliseyjust want to be sure compute node is set16:13
*** lazy_prince has quit IRC16:14
slagledprince: but, it's not too early to start the actual work on solving it, imo anyway16:14
dprinceslagle: agree, part of my wishes we had more than 1 implementation to deal with for tis16:14
dprincethis16:14
*** trown is now known as trown|lunch16:15
dprinceslagle: that would keep us honest I think16:15
slagledprince: indeed16:16
slagledtantsur: so for the virt power key for ironic, the way it used to work is we had a ssh key on the filesystem. in ironic.conf, there was a configuration value set to the path to that key16:20
*** jaosorior has joined #tripleo16:20
slagledtantsur: and that's the key that got used for pxe_ssh nodes. then it changed to be specified on the node itself, and the key saved in the ironic db16:20
slagledtantsur: it just sounded like what you were suggesting is how it used to work, and i assumed was deprecated now16:21
dtantsurslagle, no, I'm just pointing that now you can save a path to the key on a node16:21
dtantsurinstead of saving the key contents16:21
slagleah, i didn't know you could do that16:21
*** yamahata has joined #tripleo16:27
*** jistr has quit IRC16:29
*** dtantsur is now known as dtantsur|afk16:29
*** lazy_prince has joined #tripleo16:31
*** lazy_prince has quit IRC16:32
*** killer_prince has quit IRC16:33
*** lazy_prince has joined #tripleo16:34
*** devvesa has quit IRC16:35
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Pin delorean to a pre mock version  https://review.openstack.org/23001116:37
*** zaneb has quit IRC16:38
derekhThat may be needed soon ^^16:39
*** mgrohar has joined #tripleo16:42
*** mcornea has joined #tripleo16:43
*** mgrohar has quit IRC16:44
*** trown|lunch is now known as trown16:56
*** persia has joined #tripleo16:57
*** derekh has quit IRC17:01
*** sdake_ has joined #tripleo17:02
*** morazi has quit IRC17:03
*** sdake_ has quit IRC17:04
*** jerrygb has quit IRC17:04
*** sdake has quit IRC17:05
*** Marga_ has quit IRC17:07
*** Marga_ has joined #tripleo17:08
*** gfidente has quit IRC17:11
*** Marga_ has quit IRC17:13
openstackgerritJames Slagle proposed openstack/tripleo-docs: Document tripleo.sh  https://review.openstack.org/23002717:17
*** shivrao has joined #tripleo17:19
*** tosky has quit IRC17:21
*** morazi has joined #tripleo17:22
*** athomas has quit IRC17:32
*** akrivoka has quit IRC17:35
*** Nisha_away has joined #tripleo17:35
*** Marga_ has joined #tripleo17:36
*** jaosorior has quit IRC17:38
openstackgerritMerged openstack/tripleo-heat-templates: Deprecate the undercloud templates.  https://review.openstack.org/22989317:41
*** jaosorior has joined #tripleo17:42
*** rhallisey has quit IRC17:50
*** mcornea has quit IRC17:50
*** rhallisey has joined #tripleo17:50
*** rhallisey has quit IRC17:50
*** rhallisey has joined #tripleo17:51
*** pelix has quit IRC17:51
*** dprince has quit IRC17:54
*** rhallisey has quit IRC18:00
*** rhallisey_ has joined #tripleo18:01
*** lazy_prince has quit IRC18:02
*** CheKoLyN has joined #tripleo18:03
*** shardy is now known as shardy_afk18:08
*** BharatK has quit IRC18:09
*** dprince has joined #tripleo18:09
dsneddondprince, EC2MetadataIp and ControlPlaneDefaultRoute have no default values, so you can no longer deploy a stack with just network-isolation.yaml and net-single-nic-with-vlans.yaml, you need a network-environment.yaml file now: https://bugzilla.redhat.com/show_bug.cgi?id=126807018:22
openstackbugzilla.redhat.com bug 1268070 in openstack-tripleo-heat-templates "EC2MetadataIp and ControlPlaneDefaultRoute Have No Default Value" [Unspecified,New] - Assigned to jason.dobies18:22
*** rhallisey_ has quit IRC18:23
dsneddondprince, Rather than set default values, what do you think about creating a sample network-environment.yaml inside of environments/, that would have default values and serve as an example?18:23
dsneddondprince, If you like that idea, I'll submit a patch18:23
dprincedsneddon: values for those are required though18:25
dprincedsneddon: and we don't have a good default18:25
dsneddondprince, 192.0.2.1 is a good default for both, no?18:25
dprincedsneddon: I don't think so18:25
*** penick has joined #tripleo18:25
dsneddondprince, Hmm, then what values would I use there if my undercloud is 192.0.2.1?18:25
dprincedsneddon: I think it is entirely likely that you might cause someone's environment to fail if you set these by default18:25
dsneddondprince, I'm not suggesting we set those as defaults. I'm suggesting that we have an example network-environment.yaml that sets those defaults.18:26
dprincedsneddon: I use my actual router18:26
dprincedsneddon: so these18:26
dprince  ControlPlaneDefaultRoute: 172.19.0.118:26
dprince  EC2MetadataIp: 172.19.0.318:26
dprince  DnsServers: ['8.8.8.8']18:26
dprincedsneddon: 172.19.0.1 is my router18:26
dprince172.19.0.3 is my instack VM (seed)18:26
dsneddondprince, Right, but if we don't even have an example of how to set those values, then it's a mystery that is only answered in docs.18:27
dprincedsneddon: totally fine with the example, perhaps you could add it to network isolation then?18:27
*** rhallisey has joined #tripleo18:27
dprincedsneddon: since they are required?18:27
dprincedsneddon: FWIW I tried making the EC2MetadataIp parameter go away (with a reasoable default).18:28
dsneddondprince, Yeah, I know. A valiant effort.18:28
dprincenot everyone was pleased so I nuked it18:28
jaosoriorexit18:29
*** jaosorior has quit IRC18:29
dprinceI actually think we might be able to "harvest" the Ec2MetadataIp route via os-net-config18:29
dprinceand thus avoid setting it up in template land18:29
dprincedsneddon: ^18:29
dsneddondprince, Interesting. Where would we get that value, from DHCP?18:29
dprincedsneddon: yes, on the first run DHCP has it18:30
dsneddondprince, That's a good idea.18:30
dprincedsneddon: again, perhaps a bit of trickery... so I'm not sure everyone would be happy18:30
dsneddondprince, By the way, on another topic, did you see the reviews for Linux bonds and bridges? It looks like we need additional packages on the CI boxes to make that work. Is there a way to include those dependencies in the os-net-config spec?18:31
*** openstackgerrit has quit IRC18:31
dsneddondprince, ebtables and bridge-utils is what is missing18:31
*** openstackgerrit has joined #tripleo18:31
dprincedsneddon: sure, submit them here https://github.com/openstack-packages/os-net-config18:31
dprincedsneddon: git review... perhaps mark it WIP until we actually land those patches though18:32
dsneddondprince, Will do, thanks.18:32
*** Nisha_away has quit IRC18:46
*** julim has quit IRC18:48
*** rhallisey has quit IRC18:55
*** rhallisey has joined #tripleo18:56
*** rhallisey has quit IRC18:56
dprinceslagle: this patch series looks super useful https://review.openstack.org/#/c/213902/18:57
*** hashar has joined #tripleo19:03
*** rhallisey has joined #tripleo19:14
rhalliseydprince, I think the kernel panic issue has been solved already19:15
rhalliseywill 2x check.  Going to try in rhel atomic 7.119:15
dprincerhallisey: got a BZ? are we just waiting on a kernel update then?19:15
rhalliseycentos has an old kernel in the version we're using19:15
rhalliseyhttps://bugzilla.redhat.com/show_bug.cgi?id=126804719:15
openstackrhallisey: Error: Error getting bugzilla.redhat.com bug #1268047: NotPermitted19:15
dprincerhallisey: so it is an issue with the VMs kernel?19:15
*** sthillma has joined #tripleo19:15
rhalliseydprince, ya19:16
dprincerhallisey: aha, lets bump it then19:16
rhalliseydprince, supposedly this is fixed in rhel and I'm going to check that19:16
dprincerhallisey: cool19:16
rhalliseydprince, he immediately knew it was docker19:17
*** david-lyle has quit IRC19:18
rhalliseynow just need to figure out the neutron-agent issue19:18
*** zaneb has joined #tripleo19:18
dprincerhallisey: docker on the host?19:18
openstackgerritDan Prince proposed openstack-infra/tripleo-ci: Fix pep8 errors in tripleo-jobs.py  https://review.openstack.org/23009119:20
*** matbu_ has quit IRC19:23
*** matbu has joined #tripleo19:25
Slowerrhallisey: dprince: cool!19:25
*** david-lyle has joined #tripleo19:27
*** shardy_afk is now known as shardy19:33
*** julim has joined #tripleo19:33
*** jprovazn has quit IRC19:52
*** ifarkas has quit IRC19:55
*** paramite has joined #tripleo19:58
*** lblanchard has quit IRC19:59
*** jtomasek has joined #tripleo20:10
openstackgerritLokesh Jain proposed openstack/tripleo-heat-templates: Changes for configuring Nuage  https://review.openstack.org/23011620:11
*** tiswanso has quit IRC20:12
Slowerdprince: it looks like the i/o error is another kernel bug20:15
*** paramite is now known as paramite|afk20:16
Slowerdprince: http://fpaste.org/273814/73056414/20:16
openstackgerritRyan Hallisey proposed openstack/tripleo-heat-templates: Docker compute role configured via Puppet  https://review.openstack.org/20950520:26
rhalliseydprince, let's use f22 atomic20:27
rhalliseycentos kernel is giving us grief20:27
Slowerdprince: I think docker is pretty young and the 'new hotness', there's been a lot of bug fixes going on to support it20:31
Slowerdprince: I think it may be more prudent to use the latest and greatest in this situation20:31
*** dsneddon is now known as dsneddon_biab20:34
openstackgerritDerek Higgins proposed openstack-infra/tripleo-ci: Destroy previous test env VM's earlier  https://review.openstack.org/23012920:37
*** derekh has joined #tripleo20:46
*** hashar has quit IRC20:47
openstackgerritimain proposed openstack/tripleo-heat-templates: Docker compute role configured via Puppet  https://review.openstack.org/20950520:48
derekhA commit to switch delorean from using docker to mock could merge at any stage  soon, our CI isn't yet ready to use mock,20:48
derekhI'll work on it from tomorrow but in the meantime we may want to pin delorean20:49
derekhhttps://review.openstack.org/#/c/230011/120:49
derekhdprince: shardy slagle ^20:49
slaglederekh: thanks, sounds good to me20:49
slaglei just +2'd20:49
derekhslagle: thanks20:49
slaglei had that open earlier in a tab, but forgot to review20:49
derekhslagle: cool, I had a -1 on it until just now anyways20:51
slagleoh, maybe that was why :) i don't remember 4 hours ago honestly20:52
derekhslagle: 4 hours is a long time in tripleo20:53
shardyderekh: done20:54
openstackgerritMerged openstack-infra/tripleo-ci: Pin delorean to a pre mock version  https://review.openstack.org/23001120:54
derekhshardy: btw, I finally got CI working today generating packages with tripleo.sh https://review.openstack.org/#/c/229425/20:55
*** rpothier has quit IRC20:55
shardyderekh: nice!  I'll take a look in the morning :)20:55
derekhshardy: cool, I'm trying to decide which is better put the same delorean pin in tripleo.sh now before we merge it20:56
derekhshardy: or hold off and use the mock version instead20:56
derekhshardy: sure take a look in the morning and we can figure out whats best20:56
openstackgerritEmilien Macchi proposed openstack/puppet-tripleo: loadbalancer: fix MySQL timeout HAproxy config  https://review.openstack.org/22322320:57
jpeelerrhallisey: i just got the kernel panic again, guess i'll switch too20:58
rhalliseyjpeeler, ya there's a double kernel panic here20:58
rhalliseycentos atomic has has to old of a kernel20:58
rhalliseydocker blows it up20:58
jpeelerah, well i'm seeing it on centos (not atomic)20:59
jpeelerprobably the same is true there though20:59
rhalliseyjpeeler, I just updated the compute patch. We will use fedora atomic20:59
shardyjpeeler: I had my centos7 undercloud kernel panic earlier20:59
rhalliseyshardy, with containers right?20:59
*** untriaged-bot has joined #tripleo21:00
untriaged-botUntriaged bugs so far:21:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/149929821:00
openstackLaunchpad bug 1499298 in tripleo "Installers should use pymysql db driver as of Liberty" [Undecided,New]21:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/150137821:00
shardyrhallisey: I may have been deploying the docker overcloud, but I wasn't running docker on the undercloud21:00
openstackLaunchpad bug 1501378 in tripleo "pacemaker neutron resource constraints and cleanup" [Undecided,Triaged]21:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/150129621:00
openstackLaunchpad bug 1501296 in tripleo "Stack updates fail due to a new resource type" [Medium,New]21:00
*** untriaged-bot has quit IRC21:00
Slowerwe have 2 different kernel issues21:00
Slowerboth with centos21:00
jpeeleris there somewhere we should be sending this reports?21:00
*** derekh has quit IRC21:00
Slowerinstack seed VM panics and crashes21:00
shardytotally wedged the box tho, I had to reset it21:00
Slowercompute node has XFS i/o errors causing the container to fail21:00
rhalliseyjpeeler, I'm on it. They are aware and have supposedly fixed this in a later kernel21:00
Slowershardy: you had to reset the HW box?21:01
shardySlower: no, it was a VM21:01
jpeelerok, guess i'll just reset and go. this makes working pretty difficult - all i did was one redeploy21:01
Slowershardy: right, I thought it was ironic causing it?21:01
rhalliseythe issue is propagating down21:02
Slowerwhich is also wierd :)21:02
Slowerit shouldn't be able to propogate through a vm?21:02
shardyI don't see how it can be container related, I didn't have containers running on the host or on the undercloud, only in a VM on the host21:02
Slowershardy: agreed21:03
shardyI didn't investigate too much tho tbh21:03
dprinceSlower: what if we make it so that when os-net-config ifdown's eth0 to reconfigure it (the first time) it uses nmcli if networkmanager is enabled?21:04
openstackgerritEmilien Macchi proposed openstack/tripleo-heat-templates: neutron: enable nova-event-callback by default  https://review.openstack.org/22596221:04
rhalliseykernel panic logs -> http://fpaste.org/272780/44354982/21:04
rhalliseyline 6221:04
rhalliseywhen I asked the kernel guy21:04
rhalliseyhe said that is from docker21:04
rhalliseythe vm failure propagates down through instack  and kills it too21:05
rhalliseyeven though docker is only in a vm21:05
Slowerdoes ironic use device mapper?21:05
Slowerdprince: that could work..21:06
shardyrhallisey: That is pretty scary21:06
Slowerdprince: did you see the new centos issue tho?21:06
rhalliseyshardy, it is supposed to be fixed though21:06
dprinceSlower: no, bug/link?21:06
Slowerdprince: http://fpaste.org/273814/73056414/21:07
Slowerdprince: that's the source of our i/o error we're seeing on the neutron container21:07
Slowerdprince: I'm getting the sense that docker has moved pretty fast and exposed a lot of bugs..21:07
rhalliseyhttps://bugzilla.redhat.com/show_bug.cgi?id=126804721:07
openstackrhallisey: Error: Error getting bugzilla.redhat.com bug #1268047: NotPermitted21:07
dprinceSlower: can we pin to a previous version?21:07
rhallisey^ kernel bz21:08
Slowerdprince: and I wonder if we wouldn't be better off using newer upstream21:08
*** jayg is now known as jayg|g0n321:08
Slowerdprince: of docker?  We actually need the latest..21:08
dprinceSlower: or upstream is fine too21:08
Slowerdprince: by that I mean fedora atomic instead of centos (which has older kernel)21:08
dprinceSlower: we can move back to Fedora atomic for now. Fine w/ me21:09
dprinceSlower: just update the README is all that is required there21:09
Slowerdprince: ok good.. cause we already pushed that change ;-)21:09
dprinceSlower: great, press ahead man21:09
dprinceSlower: do you want me to look into the potential around using nmcli in os-net-config?21:09
Slowerhonestly if we are on fedora atomic I think we can just disable NM and it'll work..21:10
dprinceSlower: would that help resolve the DHCP lease issue w/ NetworkManager?21:10
Slowerbut we may need that for rhel, I'm not sure21:10
dprinceSlower: okay, we may need to revisit this.21:10
Slowerdprince: I'll try it out soon (once I get instack going) and let you know21:10
dprinceSlower: might be good to document this "idea", cause I will surely forget it21:11
Slowerdprince: you just want a comment someplace, or?21:11
Slowermaybe in the etherpad21:12
*** dsneddon_biab is now known as dsneddon21:12
openstackgerritEmilien Macchi proposed openstack/tripleo-heat-templates: puppet: run keystone in wsgi  https://review.openstack.org/21317521:12
dprinceSlower: sure, I'll leave a note in the etherpad21:12
*** jpeeler has quit IRC21:12
dprinceSlower: even though we are hitting new errors I'm not sure we should block this heat template patch for those21:13
Slowerdprince: I totally agree21:14
dprinceSlower: getting that landed would probably help more people get involved, so perhaps just organize all the issues as best you can and maybe we can convince shardy it is okay even if it doesn't fully work today21:14
Slowerit's actually a bit confusing as it is, we're all making changes to a patch with no diff tracking..21:14
Slowerit's gonna work by tomorrow darnit!21:15
Slower:)21:15
rhalliseyIt should work now21:15
Slowernever closer!21:15
rhalliseywas working for awhile now, but these kernel bugs came outa no where21:15
rhalliseyahhh21:15
*** paramite|afk has quit IRC21:16
Slowerno they came from centos :)21:18
*** Goneri has quit IRC21:19
*** rhallisey has quit IRC21:23
*** shardy has quit IRC21:26
*** Marga_ has quit IRC21:28
*** trown is now known as trown|outttypeww21:30
*** Marga_ has joined #tripleo21:30
*** david-lyle has quit IRC21:43
*** rlandy has quit IRC21:43
*** weshay has quit IRC21:44
*** Marga_ has quit IRC21:46
*** Marga_ has joined #tripleo21:47
*** spzala has quit IRC21:54
*** spzala has joined #tripleo21:55
*** david-lyle has joined #tripleo21:56
*** Marga_ has quit IRC21:59
*** Marga_ has joined #tripleo21:59
*** CheKoLyN has quit IRC21:59
*** Marga_ has quit IRC22:00
*** Marga_ has joined #tripleo22:00
*** nijaba has quit IRC22:16
*** Goneri has joined #tripleo22:20
*** egafford has left #tripleo22:20
*** nijaba has joined #tripleo22:26
*** nijaba has quit IRC22:26
*** nijaba has joined #tripleo22:26
*** nijaba has quit IRC22:33
*** nijaba has joined #tripleo22:36
*** nijaba has quit IRC22:36
*** nijaba has joined #tripleo22:36
*** sthillma_ has joined #tripleo22:40
*** sthillma has quit IRC22:43
*** sthillma_ is now known as sthillma22:43
*** dprince has quit IRC22:46
*** Goneri has quit IRC22:47
*** hashar has joined #tripleo22:50
*** pradk has quit IRC22:59
*** mudassirlatif has joined #tripleo23:01
mudassirlatifHi, newbie question: How do I find out if tripleo release has been cut? Is it too late for change anything for Liberty? If not, what are the deadlines?23:03
*** chlong has quit IRC23:07
*** hashar has quit IRC23:08
*** sthillma_ has joined #tripleo23:16
*** sthillma has quit IRC23:19
*** sthillma_ is now known as sthillma23:19
*** alop has quit IRC23:25
*** rasca_ has joined #tripleo23:26
*** rasca has quit IRC23:29
*** rhallisey has joined #tripleo23:29
openstackgerritDerek Higgins proposed openstack/diskimage-builder: Ignore ./proc when deleting python object files  https://review.openstack.org/23020223:31
*** Goneri has joined #tripleo23:56
*** bnemec has joined #tripleo23:57
« Wednesday, 2015-09-30 Index Friday, 2015-10-02 »

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