| *** Marga_ has quit IRC | 00:01 | |
| *** trozet has quit IRC | 00:02 | |
| *** tiswanso has joined #tripleo | 00:28 | |
| *** rcernin has quit IRC | 00:30 | |
| *** tiswanso has quit IRC | 00:37 | |
| *** panda_ has quit IRC | 00:42 | |
| *** panda_ has joined #tripleo | 00:42 | |
| *** mbound has quit IRC | 00:46 | |
| *** bswartz has joined #tripleo | 00:51 | |
| *** bswartz has left #tripleo | 00:51 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder: [WIP] Add systemd support to growroot https://review.openstack.org/271862 | 01:00 |
|---|---|---|
| *** dmacpher has joined #tripleo | 01:18 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder: [WIP] Add systemd support to growroot https://review.openstack.org/271862 | 01:28 |
| *** tiswanso has joined #tripleo | 01:34 | |
| *** tiswanso has quit IRC | 01:43 | |
| *** mbound has joined #tripleo | 01:46 | |
| *** mbound has quit IRC | 01:51 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder: [WIP] Add systemd support to growroot https://review.openstack.org/271862 | 01:57 |
| *** shivrao has joined #tripleo | 02:02 | |
| *** trozet has joined #tripleo | 02:11 | |
| openstackgerrit | Merged openstack/puppet-tripleo: Drop webmock dependency https://review.openstack.org/271553 | 02:27 |
| *** tiswanso has joined #tripleo | 02:40 | |
| *** tiswanso has quit IRC | 02:52 | |
| *** trozet has quit IRC | 03:11 | |
| *** yuanying_ has joined #tripleo | 03:18 | |
| *** yuanying has quit IRC | 03:19 | |
| *** yuanying has joined #tripleo | 03:21 | |
| *** yuanying_ has quit IRC | 03:23 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder: [WIP] Add systemd support to growroot https://review.openstack.org/271862 | 03:38 |
| *** dmacpher has quit IRC | 03:45 | |
| *** tiswanso has joined #tripleo | 03:49 | |
| *** tiswanso has quit IRC | 04:03 | |
| *** dmacpher has joined #tripleo | 04:05 | |
| *** yuanying_ has joined #tripleo | 04:07 | |
| *** yuanying has quit IRC | 04:08 | |
| *** julim has quit IRC | 04:14 | |
| *** Marga_ has joined #tripleo | 04:19 | |
| *** Marga_ has quit IRC | 04:23 | |
| *** cwolferh has joined #tripleo | 04:37 | |
| *** dmacpher has quit IRC | 04:45 | |
| *** chlong has joined #tripleo | 04:52 | |
| *** cwolferh has quit IRC | 04:55 | |
| *** tiswanso has joined #tripleo | 05:00 | |
| *** dmacpher has joined #tripleo | 05:02 | |
| *** tiswanso has quit IRC | 05:13 | |
| *** coolsvap|away is now known as coolsvap | 05:19 | |
| *** oshvartz has quit IRC | 05:32 | |
| *** Goneri has quit IRC | 05:39 | |
| *** masco has joined #tripleo | 05:46 | |
| *** jaosorior has joined #tripleo | 05:52 | |
| *** rcernin has joined #tripleo | 05:54 | |
| *** Marga_ has joined #tripleo | 06:05 | |
| *** Marga_ has quit IRC | 06:09 | |
| *** tiswanso has joined #tripleo | 06:10 | |
| *** tiswanso has quit IRC | 06:20 | |
| *** jaosorior has quit IRC | 06:22 | |
| *** mkovacik has quit IRC | 06:24 | |
| *** chlong has quit IRC | 06:28 | |
| *** jaosorior has joined #tripleo | 06:30 | |
| *** marios_ has joined #tripleo | 06:33 | |
| *** marios_ has quit IRC | 06:33 | |
| *** shivrao has quit IRC | 06:35 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder: Add systemd/fedora support to growroot https://review.openstack.org/271862 | 06:42 |
| *** aufi has joined #tripleo | 06:57 | |
| *** davidlenwell has quit IRC | 07:00 | |
| *** dshulyak_ has joined #tripleo | 07:01 | |
| *** davidlenwell has joined #tripleo | 07:02 | |
| *** ChanServ sets mode: +v davidlenwell | 07:02 | |
| *** oshvartz has joined #tripleo | 07:05 | |
| *** dmacpher has quit IRC | 07:06 | |
| *** Marga_ has joined #tripleo | 07:06 | |
| *** ukalifon has joined #tripleo | 07:13 | |
| *** tiswanso has joined #tripleo | 07:17 | |
| *** teju has joined #tripleo | 07:21 | |
| *** jhenner has joined #tripleo | 07:22 | |
| *** davidlenwell has quit IRC | 07:24 | |
| *** coolsvap is now known as coolsvap|away | 07:25 | |
| *** davidlenwell has joined #tripleo | 07:26 | |
| *** ChanServ sets mode: +v davidlenwell | 07:26 | |
| *** akuznetsov has joined #tripleo | 07:27 | |
| *** leanderthal has joined #tripleo | 07:27 | |
| *** tiswanso has quit IRC | 07:29 | |
| *** davidlenwell has quit IRC | 07:29 | |
| *** jtomasek has joined #tripleo | 07:29 | |
| *** davidlenwell has joined #tripleo | 07:33 | |
| *** ChanServ sets mode: +v davidlenwell | 07:33 | |
| *** liverpooler has joined #tripleo | 07:34 | |
| *** davidlenwell has quit IRC | 07:41 | |
| *** davidlenwell has joined #tripleo | 07:45 | |
| *** ChanServ sets mode: +v davidlenwell | 07:45 | |
| openstackgerrit | Matthew Thode proposed openstack/diskimage-builder: add support for package-installs to gentoo https://review.openstack.org/270597 | 07:49 |
| *** mkovacik has joined #tripleo | 07:51 | |
| *** Marga_ has quit IRC | 07:54 | |
| openstackgerrit | xin wu proposed openstack/tripleo-heat-templates: Add extra config yaml files for big switch agents. https://review.openstack.org/271922 | 07:54 |
| *** jcoufal has joined #tripleo | 07:54 | |
| openstackgerrit | xin wu proposed openstack/tripleo-heat-templates: Add extra config yaml files for big switch agents. https://review.openstack.org/271922 | 07:59 |
| *** pblaho has joined #tripleo | 08:00 | |
| *** akuznetsov has quit IRC | 08:01 | |
| *** mkovacik has quit IRC | 08:01 | |
| *** ifarkas has joined #tripleo | 08:01 | |
| *** fgimenez has joined #tripleo | 08:02 | |
| *** fgimenez has quit IRC | 08:02 | |
| *** fgimenez has joined #tripleo | 08:02 | |
| *** jhenner has quit IRC | 08:04 | |
| *** jprovazn has joined #tripleo | 08:08 | |
| *** akuznetsov has joined #tripleo | 08:09 | |
| *** bvandenh has joined #tripleo | 08:10 | |
| *** regebro has joined #tripleo | 08:11 | |
| *** jhenner has joined #tripleo | 08:12 | |
| *** akuznetsov has quit IRC | 08:13 | |
| *** coolsvap|away is now known as coolsvap | 08:16 | |
| *** akuznetsov has joined #tripleo | 08:24 | |
| *** tiswanso has joined #tripleo | 08:26 | |
| *** akuznetsov has quit IRC | 08:30 | |
| *** shardy has joined #tripleo | 08:30 | |
| *** rwsu has joined #tripleo | 08:34 | |
| *** gfidente has joined #tripleo | 08:37 | |
| *** jcoufal has quit IRC | 08:37 | |
| openstackgerrit | xin wu proposed openstack/tripleo-heat-templates: Include big switch puppet modules for deploying overcloud https://review.openstack.org/271940 | 08:38 |
| *** jcoufal has joined #tripleo | 08:38 | |
| *** tiswanso has quit IRC | 08:39 | |
| openstackgerrit | xin wu proposed openstack/tripleo-heat-templates: Include big switch puppet modules for deploying overcloud https://review.openstack.org/271940 | 08:41 |
| *** devvesa has joined #tripleo | 08:46 | |
| openstackgerrit | xin wu proposed openstack/tripleo-heat-templates: Add extra config yaml files for big switch agents. https://review.openstack.org/271922 | 08:48 |
| *** mcornea has joined #tripleo | 08:50 | |
| openstackgerrit | Oded Shvartz proposed openstack/tripleo-common: overcloud-odl : add new image file definition https://review.openstack.org/266881 | 08:53 |
| *** Marga_ has joined #tripleo | 08:54 | |
| *** jhenner has quit IRC | 08:57 | |
| openstackgerrit | xin wu proposed openstack/tripleo-heat-templates: Include big switch puppet modules for deploying overcloud https://review.openstack.org/271953 | 08:58 |
| *** Marga_ has quit IRC | 09:00 | |
| *** mkovacik has joined #tripleo | 09:00 | |
| *** athomas has joined #tripleo | 09:02 | |
| *** akuznetsov has joined #tripleo | 09:02 | |
| *** bnemec has quit IRC | 09:12 | |
| *** hewbrocca has joined #tripleo | 09:16 | |
| *** mbound has joined #tripleo | 09:18 | |
| *** jhenner has joined #tripleo | 09:21 | |
| *** jistr has joined #tripleo | 09:23 | |
| jaosorior | Any +As for this CR? https://review.openstack.org/#/c/268647/ :D | 09:23 |
| *** nico_auv has joined #tripleo | 09:28 | |
| jaosorior | gfidente: Thanks dude! | 09:29 |
| *** stendulker has joined #tripleo | 09:30 | |
| *** akuznetsov has quit IRC | 09:31 | |
| openstackgerrit | Merged openstack/puppet-tripleo: SSL/Cinder: enable ssl_header_handler filter https://review.openstack.org/268647 | 09:32 |
| gfidente | jaosorior, ty | 09:32 |
| *** derekh has joined #tripleo | 09:34 | |
| *** tiswanso has joined #tripleo | 09:36 | |
| *** dtantsur|afk is now known as dtantsur | 09:44 | |
| *** tiswanso has quit IRC | 09:47 | |
| *** Marga_ has joined #tripleo | 09:56 | |
| *** Marga_ has quit IRC | 10:01 | |
| openstackgerrit | xin wu proposed openstack/python-tripleoclient: Install bigswitch networking agent by default https://review.openstack.org/271990 | 10:06 |
| openstackgerrit | xin wu proposed openstack/python-tripleoclient: Install bigswitch networking agent by default https://review.openstack.org/271990 | 10:08 |
| *** shadower has joined #tripleo | 10:09 | |
| openstackgerrit | Merged openstack/tripleo-common: Add openstack packages to heat-agents container https://review.openstack.org/266066 | 10:12 |
| openstackgerrit | xin wu proposed openstack/python-tripleoclient: Install bigswitch networking agent by default https://review.openstack.org/271990 | 10:12 |
| *** rebrego has joined #tripleo | 10:22 | |
| *** regebro is now known as Guest99008 | 10:22 | |
| *** rebrego is now known as regebro | 10:22 | |
| *** jaosorior has quit IRC | 10:22 | |
| *** jaosorior has joined #tripleo | 10:23 | |
| *** Guest99008 has quit IRC | 10:26 | |
| *** SlickN1k has joined #tripleo | 10:28 | |
| *** jaosorior has quit IRC | 10:29 | |
| *** SlickNik has quit IRC | 10:29 | |
| *** SlickN1k is now known as SlickNik | 10:29 | |
| *** cmyster has quit IRC | 10:31 | |
| *** jaosorior has joined #tripleo | 10:34 | |
| *** SlickNik has quit IRC | 10:34 | |
| *** SlickNik has joined #tripleo | 10:35 | |
| *** SlickN1k has joined #tripleo | 10:37 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/tripleo-heat-templates: Enable SSL middleware for cinder https://review.openstack.org/268651 | 10:39 |
| *** olap has joined #tripleo | 10:39 | |
| *** SlickNik has quit IRC | 10:39 | |
| *** SlickN1k is now known as SlickNik | 10:39 | |
| *** tosky has joined #tripleo | 10:43 | |
| *** tiswanso has joined #tripleo | 10:44 | |
| openstackgerrit | Jiri Stransky proposed openstack/tripleo-heat-templates: Enable upgrading the rest of the node types https://review.openstack.org/271359 | 10:45 |
| *** mgould has joined #tripleo | 10:47 | |
| *** paramite has joined #tripleo | 10:48 | |
| *** SlickN1k has joined #tripleo | 10:53 | |
| *** SlickNik has quit IRC | 10:54 | |
| *** SlickN1k is now known as SlickNik | 10:54 | |
| *** regebro has quit IRC | 10:54 | |
| *** Marga_ has joined #tripleo | 10:56 | |
| *** tiswanso has quit IRC | 10:57 | |
| *** stendulker_ has joined #tripleo | 10:59 | |
| openstackgerrit | Jiri Stransky proposed openstack/tripleo-heat-templates: Enable upgrading the rest of the node types https://review.openstack.org/271359 | 11:01 |
| *** Marga_ has quit IRC | 11:02 | |
| *** stendulker has quit IRC | 11:02 | |
| *** athomas has quit IRC | 11:04 | |
| *** regebro has joined #tripleo | 11:05 | |
| *** electrofelix has joined #tripleo | 11:05 | |
| *** stendulker_ has quit IRC | 11:06 | |
| *** athomas has joined #tripleo | 11:12 | |
| *** fgimenez has quit IRC | 11:16 | |
| *** fgimenez has joined #tripleo | 11:16 | |
| *** fgimenez has quit IRC | 11:16 | |
| *** fgimenez has joined #tripleo | 11:16 | |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: [NO MERGY] Test a update in trunk repository version https://review.openstack.org/229789 | 11:30 |
| shardy | Mysql::Server::Root_password/Exec[remove install pass]: Skipping because of failed dependencies | 11:34 |
| shardy | anyone know if we have a fix for that? | 11:34 |
| * shardy doesn't see a bug | 11:34 | |
| shardy | http://logs.openstack.org/80/270080/1/check-tripleo/gate-tripleo-ci-f22-ha/7f90f8e/console.html | 11:35 |
| *** bvandenh has quit IRC | 11:36 | |
| shardy | https://bugs.launchpad.net/tripleo/+bug/1537720 | 11:38 |
| openstack | Launchpad bug 1537720 in tripleo "mysql::server Skipping because of failed dependencies errors in CI" [Critical,Triaged] | 11:38 |
| derekh | shardy: is this the error (i.e. that failed dependency) | 11:39 |
| derekh | Execution of '/usr/libexec/mysqld -V' returned 1: 160125 | 11:39 |
| derekh | shardy: ^ | 11:39 |
| shardy | derekh: Hmm, maybe not, it's hard to parse the deploy_stderr wall-of-text | 11:40 |
| shardy | feel free to adjust the bug description :) | 11:41 |
| shardy | option '--wsrep_notify_cmd' requires an argument | 11:42 |
| shardy | maybe that's it | 11:42 |
| shardy | https://bugzilla.redhat.com/show_bug.cgi?id=1301290 | 11:44 |
| openstack | bugzilla.redhat.com bug 1301290 in openstack-puppet-modules "Mitaka - overcloud deploy gives: [ERROR] /usr/libexec/mysqld: option '--wsrep_notify_cmd' requires an argument" [Unspecified,New] - Assigned to emacchi | 11:44 |
| shardy | more information there I think | 11:44 |
| openstackgerrit | Marios Andreou proposed openstack/tripleo-heat-templates: Bump the pacemaker service op_params to 200s for start and stop https://review.openstack.org/272026 | 11:45 |
| * hewbrocca +1000 marios patch | 11:45 | |
| derekh | shardy: did the problem start on friday evening by any chance? Thats then a new puppet-modules package got into the liberty repository | 11:49 |
| shardy | derekh: Yeah, that sounds likely | 11:49 |
| shardy | and the downstream bug is wrt o-p-m in RDO | 11:49 |
| shardy | which we're still using in stable/liberty CI | 11:49 |
| marios | hewbrocca: heh... it is just a stable/liberty cherrypick (yeah it landed in master already) | 11:50 |
| derekh | shardy: yup | 11:50 |
| *** akrivoka has joined #tripleo | 11:50 | |
| shardy | I'll chat with EmilienM later, I wasn't sure if he was going to take a look at making CI build from stable/liberty source for the modules, if not I'll take a look | 11:50 |
| shardy | shouldn't be too hard, I just haven't got around to it | 11:51 |
| derekh | ok | 11:51 |
| *** bvandenh has joined #tripleo | 11:53 | |
| shardy | https://github.com/puppetlabs/puppetlabs-mysql/commit/e30e0bc958761890ea4f06cdd3f1fc7242a00fe2 is the commit referenced in the bz | 11:54 |
| shardy | I don't think everything is broken on master CI tho? | 11:54 |
| *** tiswanso has joined #tripleo | 11:54 | |
| *** absubram has quit IRC | 11:55 | |
| *** mgould has quit IRC | 12:01 | |
| *** teju has quit IRC | 12:03 | |
| *** tiswanso has quit IRC | 12:05 | |
| *** olap has quit IRC | 12:07 | |
| *** olap has joined #tripleo | 12:10 | |
| *** marcusvrn_ has joined #tripleo | 12:13 | |
| *** mburned_out is now known as mburned | 12:15 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-tripleoclient: Allow assigning profiles based on XXX_profile capabilities https://review.openstack.org/272054 | 12:21 |
| dtantsur | liberty backport ^^ | 12:22 |
| *** akuznetsov has joined #tripleo | 12:28 | |
| openstackgerrit | afazekas proposed openstack-infra/tripleo-ci: Add dstat https://review.openstack.org/271218 | 12:33 |
| EmilienM | shardy: hey | 12:36 |
| *** akuznetsov has quit IRC | 12:36 | |
| *** trown|outttypeww is now known as trown | 12:37 | |
| openstackgerrit | Emilien Macchi proposed openstack/puppet-tripleo: Enable X-Forwarded-Proto header for Heat and Nova https://review.openstack.org/271481 | 12:37 |
| *** Guest74617 is now known as beagles | 12:37 | |
| openstackgerrit | Emilien Macchi proposed openstack/puppet-tripleo: Enable X-Forwarded-Proto header for keystone_public https://review.openstack.org/271482 | 12:37 |
| openstackgerrit | Emilien Macchi proposed openstack/puppet-tripleo: Add tripleo_ipv6_interfaces fact https://review.openstack.org/271483 | 12:37 |
| openstackgerrit | Derek Higgins proposed openstack/tripleo-common: WIP: split overcloud setup out of pingtest https://review.openstack.org/270631 | 12:47 |
| *** olap has quit IRC | 12:50 | |
| *** lucasagomes is now known as lucas-hungry | 12:52 | |
| *** olap has joined #tripleo | 12:53 | |
| *** jhenner has quit IRC | 12:56 | |
| *** jhenner has joined #tripleo | 12:57 | |
| dtantsur | shardy, hi! is it possible you take a look at the profile matching docs? https://review.openstack.org/#/c/257867 the code was merged to master | 12:58 |
| *** tiswanso has joined #tripleo | 12:59 | |
| *** tiswanso has quit IRC | 12:59 | |
| *** Marga_ has joined #tripleo | 12:59 | |
| *** tiswanso has joined #tripleo | 13:00 | |
| *** Marga_ has quit IRC | 13:04 | |
| *** jdob has joined #tripleo | 13:06 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Add IPv6 Support to Isolated Networks https://review.openstack.org/235423 | 13:07 |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Support the deployment of Ceph over IPv6 https://review.openstack.org/272089 | 13:07 |
| gfidente | dsneddon, ^^ had to rebase | 13:07 |
| jistr | shardy, gfidente: do you folks have a minute to discuss upgrades a bit? | 13:08 |
| gfidente | jistr, yeah I would like to actually | 13:08 |
| jistr | s/minute/more than a minute/ | 13:08 |
| jistr | :) | 13:08 |
| gfidente | cause I am get lost there :( | 13:08 |
| *** tiswanso has quit IRC | 13:09 | |
| jistr | ok, yeah no wonder. I actually have a little bit of a hard time making them happen with Heat. There are some requirements which don't quite fit how THT are currently structured. | 13:10 |
| jistr | gfidente: i guess you're still kinda familiar with this one, right? https://review.openstack.org/#/c/253276 i did some changes but no big restructuring | 13:11 |
| gfidente | jistr, I was thinking if we have 'worded' the process somewhere? abstracting it completely from the tripleo implementation, just to discuss in which order things should happen | 13:12 |
| jistr | gfidente: yeah i think we have it on trello somewhere. Though that's the main problem, translating that into THT is becoming a bit of a nightmare :) | 13:13 |
| *** tzumainn has joined #tripleo | 13:13 | |
| gfidente | yeah I got that was the issue | 13:13 |
| jistr | i have one dead-end set of patches here https://review.openstack.org/#/c/270243/ | 13:14 |
| jistr | i tried to put it into the OS::TripleO::Tasks::PackageUpdate interface | 13:14 |
| *** hjensas has joined #tripleo | 13:14 | |
| *** hjensas has joined #tripleo | 13:14 | |
| jistr | but we need the controller upgrade to happen in orchestrated two steps | 13:15 |
| jistr | and the PackageUpdate hook only operates with knowledge of a single machine, so there's no way how to ensure on Heat side that Step1 runs on *all* controllers before Step2 is triggered | 13:15 |
| jistr | so i abandoned the PackageUpdate way | 13:15 |
| jistr | and i introduced a new UpdateWorkflow hook https://review.openstack.org/#/c/271358 | 13:16 |
| jistr | this one receives all machines by roles | 13:16 |
| jistr | meaning you can work with all controllers and sync operations in a Step1->Step2 manner if you want to | 13:17 |
| *** jayg|g0n3 is now known as jayg | 13:17 | |
| jistr | and then on top of that i added initial support for upgrading also other node types (tried just compute) here https://review.openstack.org/#/c/271359 | 13:18 |
| shardy | jistr: Hey, just getting back from lunch | 13:18 |
| jistr | shardy: cool :) i'll finish describing the situation and then if you could look at the patches if they're sane approach-wise, that would be most helpful :) | 13:19 |
| shardy | jistr: can't we use e.g | 13:19 |
| shardy | ControllerPrePuppet to run over all nodes? | 13:19 |
| *** dprince has joined #tripleo | 13:20 | |
| jistr | shardy: that's what we did initially, however there are additional requirements which don't play well with that | 13:20 |
| shardy | jistr: everything in -post.yaml runs over all nodes, so you could do steps inside a template mapped to ControllerPrePuppet | 13:20 |
| *** fgimenez has quit IRC | 13:20 | |
| shardy | jistr: ack | 13:21 |
| shardy | FWIW I agree that a step-by-step (non heat) description of the workflow would be really helpful | 13:21 |
| jistr | * we need to upgrade other node types too. Controllers need to go through the upgrade first. There shouldn't be a race between puppet applying on computes vs. upgrade being performed on controllers. | 13:21 |
| *** rhallisey has joined #tripleo | 13:21 | |
| shardy | jistr: I thought we decided compute upgrades were out of scope, at least initially | 13:22 |
| shardy | there's going to be operator intervention there to evacuate, so I assumed they'd probably evacuate a node, then exclude it and build a new one | 13:22 |
| jistr | ok perhaps i'm aiming this too high then. If we only upgraded controllers via Heat, then yes, we could do with ControllerPrePuppet | 13:23 |
| shardy | jistr: I think that's the first and most important step | 13:23 |
| *** fgimenez has joined #tripleo | 13:23 | |
| shardy | then we do have to consider e.g the *Storage nodes too | 13:23 |
| shardy | IMO the compute is the one we should probably consider last | 13:23 |
| shardy | as it is highly likely to require operator intervention anyway | 13:24 |
| gfidente | lukily storage, when ceph, does not depend on the status of the openstack services | 13:24 |
| jistr | ok... so whether we have a problem or not probably depends how much of this we want to do through Heat eventually | 13:24 |
| shardy | jistr: I personally think any workflow around e.g evaccuation of computes should never be done via heat | 13:25 |
| jistr | if we want to add the storage/compute upgrade support into Heat, then it might be valuable to start rethinking the ControllerPrePuppet approach right now | 13:25 |
| shardy | it's a workflow, perhaps a candidate for a mistral workflow or some other tool | 13:25 |
| jistr | yea | 13:25 |
| jistr | personally i'd like to do *all* updates/upgrades outside of Heat | 13:25 |
| shardy | jistr: that's also a possibility, although it seems like we're close enough with Heat-driven controller upgrades for that to be a reasonable first-pass? | 13:26 |
| shardy | even if in the long term that upgrade workflow is implemented elsewhere? | 13:26 |
| jistr | shardy: yes i think so | 13:26 |
| *** egafford has joined #tripleo | 13:26 | |
| jistr | i ran through controller upgrades several times upstream (though just no-op upgrades from master to master, just to test the "workflow"), and it worked fine | 13:27 |
| shardy | jistr: that's great! | 13:27 |
| shardy | I guess the next step is to test stable/liberty to master | 13:28 |
| shardy | and then to get that tested in CI | 13:28 |
| shardy | I think for now, having that work, and end with the controllers pinned to liberty computes would be perfectly fine | 13:28 |
| gfidente | shardy, yeah CI we have the job | 13:28 |
| gfidente | it just doesn't pass yet | 13:28 |
| jistr | gfidente: a new job? | 13:28 |
| shardy | then we have a documented process to manually upgrade/replace the computes, and finally upate to remove the pin | 13:28 |
| gfidente | jistr, no | 13:29 |
| gfidente | https://review.openstack.org/260466 | 13:29 |
| gfidente | this will add it for the ha job to both master and stable | 13:29 |
| shardy | gfidente: how does that handle updating the yum repo configuration between initial (e.g liberty) deployment, and the subsequent upgrade? | 13:30 |
| jistr | gfidente, shardy: ah ok... about that though -- i was wondering if we could perhaps have a new job for upgrades testing, and use pre-built stable images there | 13:30 |
| gfidente | shardy, it doesnt! | 13:30 |
| shardy | I assumed we'd need to export some additional variables to build a liberty image in both cases | 13:30 |
| shardy | then add logic (probably dprince's deploy artefacts) to update yum configs, then do the update | 13:30 |
| shardy | gfidente: Ok, cool, so we'll have to add that | 13:31 |
| jistr | shardy: yea, what about using pre-built images? do you think we could risk bugs falling through the cracks if using them? The reason i think pre-built might be the way to go is the CI job run time. | 13:32 |
| shardy | https://review.openstack.org/#/c/260144/ shows the steps to do the image build btw in the commit message | 13:32 |
| jistr | the upgrade run takes a while even when it doesn't do much | 13:32 |
| shardy | jistr: I think we do need some way to optimze the runtime, but what happens e.g when there's a Depends-On to puppet-* ? | 13:32 |
| shardy | it's a tough problem if we want to stay within the CI job timeout | 13:33 |
| shardy | again, maybe we could use the deploy artefacts stuff to always deploy the modules from source | 13:33 |
| shardy | and then use images built via the periodic job or something | 13:34 |
| shardy | derekh: ^^ interested in your thoughts here | 13:34 |
| jistr | hmm i guess if there's a depends-on to puppet, we have a problem anyway even when building the image, no? because you probably want the depends-on to be in effect for the upgrade itself, but not for the initial deploy, no? | 13:34 |
| shardy | jistr: If the Depends-On is proposed to both master and stable/liberty, we'd use it for the initial deploy, otherwise we'd do the initial deploy without it | 13:35 |
| shardy | although, atm that won't work because we're still using opm in stable/liberty CI | 13:35 |
| *** thrash|wknd is now known as thrah | 13:36 | |
| *** thrah is now known as thrash | 13:36 | |
| shardy | jistr: the other option wrt runtime is only test upgrades via the periodic job | 13:36 |
| shardy | but that has the disadvantage that invariably it'll constantly break | 13:36 |
| jistr | yea.. | 13:36 |
| gfidente | would a periodic job include a set of changes | 13:37 |
| gfidente | or would there be multiple periodic jobs running each with a specific change btw? | 13:38 |
| jistr | gfidente: probably a set of changes. i guess it would include anything that's been merged stable/liberty and master (for initial deploy and subsequent upgrade, respectively) by the time the periodic job runs | 13:39 |
| *** pradk has joined #tripleo | 13:39 | |
| jistr | s/merged/merged to/ | 13:39 |
| gfidente | oh so it would run against the merged changes, not the unmerged ones? | 13:39 |
| *** pradk has quit IRC | 13:39 | |
| jistr | yea, that's why it's not so helpful as check-CI | 13:39 |
| gfidente | it's also a nightmare to troubleshoot | 13:40 |
| jistr | only tells you when something already broke, not before | 13:40 |
| gfidente | it seems | 13:40 |
| jistr | yea | 13:40 |
| *** paramite is now known as paramite|afk | 13:40 | |
| gfidente | so to me, it'd be nice if we could get the upgrade workflow to pass without doing any yum upgrade first | 13:41 |
| gfidente | because that would validate the workflow, at least 'syntactically' | 13:41 |
| *** stendulker has joined #tripleo | 13:41 | |
| openstackgerrit | afazekas proposed openstack-infra/tripleo-ci: scp/ssh do not check for know host keys https://review.openstack.org/269660 | 13:45 |
| jistr | gfidente: that makes sense to me. However, do we want to have it at the end of the normal HA jobs? Perhaps the upgrade job should be a separate one, non-voting at first. And if we already have it as a separate job, it could be easier to transition from a "master->master" model to a "stable->master" model. For "stable->master" we'd need the separate job anyway. | 13:47 |
| gfidente | jistr so the problem was that currently that is meant to run the upgrade only in the pcmk scenario | 13:49 |
| gfidente | I think adding a single node /w pcmk job is interesting, but I am not sure it'd be catching actual upgrade issues then | 13:50 |
| jistr | and having another full pcmk job is too much for current CI resources? | 13:51 |
| *** coolsvap is now known as coolsvap|away | 13:51 | |
| *** dmacpher has joined #tripleo | 13:53 | |
| *** Marga_ has joined #tripleo | 13:54 | |
| derekh | shardy: Yup, I'd love to get the periodic job to start building images we can use the problem has been getting trunk to actually build (so the periodic job passes), | 13:57 |
| derekh | shardy: I had been trying to sort it out before christmas but started working on other stuff while waitong on patches to merge, getting back to it this week | 13:58 |
| shardy | derekh: ack, thanks for the update | 13:58 |
| *** lucas-hungry is now known as lucasagomes | 14:00 | |
| derekh | afazekas: are you using those instances I started on rh1 for testing ? wondering if either can be deleted | 14:00 |
| trown | derekh: I have a hash we should try for master | 14:01 |
| trown | derekh: it passed RDO CI | 14:01 |
| *** rlandy has joined #tripleo | 14:02 | |
| derekh | trown: cool, lets try it, I run the most recent earlier but got this error | 14:02 |
| derekh | trown: 2016-01-25 11:46:16.107 | ImportError: No module named positional | 14:02 |
| afazekas | derekh: yes | 14:02 |
| trown | derekh: ya I havent had anything pass RDO CI for a few days (after adjusting for our HA job not actually deploying HA and having to go backwards after resolving that) | 14:03 |
| afazekas | derekh: key changed on the 2th so I assumed it is not the same instance anymore, but 1th runs the gearman lockenv stat | 14:04 |
| derekh | afazekas: weird, I didn't change anything, looking | 14:04 |
| derekh | trown: ok, I'm going to see if I can help out here this week, weren't you keeping track of current issues somewhere | 14:06 |
| trown | derekh: do you have any review up for testing different delorean hashes? https://github.com/openstack-infra/tripleo-ci/blob/master/toci_gate_test.sh#L50 would be what we change? | 14:06 |
| trown | https://etherpad.openstack.org/p/delorean_master_current_issues | 14:06 |
| *** Marga_ has quit IRC | 14:07 | |
| trown | derekh: the top issue there is a puppet issue that is merged, so should be fine from source | 14:07 |
| *** julim has joined #tripleo | 14:07 | |
| *** Marga_ has joined #tripleo | 14:07 | |
| afazekas | derekh, what is missing for switching to OVB setup ? eta ? | 14:07 |
| derekh | trown: this is how I normally test a hash, https://review.openstack.org/#/c/229789/ | 14:07 |
| trown | derekh: for RDO I had to manually patch it in the images, because the rebase of OPM that picks up that change breaks something else | 14:07 |
| derekh | afazekas: mainly HW to test deploying and testing it on | 14:08 |
| *** akrivoka has quit IRC | 14:09 | |
| derekh | afazekas: I'd want to test deploying and using it on a cloud with at least 10 compute nodes to help prove nothing will fall over when if we redeploy rh1 | 14:10 |
| openstackgerrit | John Trowbridge proposed openstack-infra/tripleo-ci: [NO MERGY] Test a update in trunk repository version https://review.openstack.org/229789 | 14:10 |
| trown | derekh: do we pin puppet modules somewhere? | 14:11 |
| openstackgerrit | Ryan Hallisey proposed openstack/tripleo-docs: Docs for containerized compute node https://review.openstack.org/254743 | 14:12 |
| derekh | trown: we can | 14:13 |
| derekh | here https://git.openstack.org/cgit/openstack/tripleo-common/tree/scripts/tripleo.sh#n334 | 14:13 |
| derekh | and here https://git.openstack.org/cgit/openstack/tripleo-common/tree/scripts/tripleo.sh#n371 | 14:13 |
| derekh | jistr: any idea if that pin of puppet-mysql can be removed? | 14:14 |
| trown | derekh: ya puppet-mysql is the one we might need to move/remove for the top issue on that etherpad | 14:15 |
| jistr | derekh: don't know, i'll submit a revert and we'll see what CI will say, sounds ok? | 14:15 |
| openstackgerrit | Ryan Hallisey proposed openstack/tripleo-docs: Docs for containerized compute node https://review.openstack.org/254743 | 14:15 |
| derekh | jistr: sounds good | 14:15 |
| *** mgould has joined #tripleo | 14:15 | |
| derekh | afazekas: I can ssh to both instances, whats the IP of the one your using? I'll remove the other if thats ok | 14:17 |
| *** dtantsur is now known as dtantsur|brb | 14:17 | |
| *** Goneri has joined #tripleo | 14:17 | |
| *** masco has quit IRC | 14:18 | |
| afazekas | derekh, Do not remove: 66.187.229.83 | 14:20 |
| derekh | afazekas: ack, thanks | 14:20 |
| openstackgerrit | Ryota MIBU proposed openstack/os-cloud-config: add aodh service in keystone service list https://review.openstack.org/272110 | 14:21 |
| *** akrivoka has joined #tripleo | 14:21 | |
| *** lblanchard has joined #tripleo | 14:24 | |
| *** paramite|afk is now known as paramite | 14:24 | |
| *** panda_ has quit IRC | 14:25 | |
| *** panda_ has joined #tripleo | 14:26 | |
| *** rpothier has joined #tripleo | 14:34 | |
| *** tiswanso has joined #tripleo | 14:35 | |
| *** tiswanso has quit IRC | 14:36 | |
| *** tiswanso has joined #tripleo | 14:38 | |
| *** tiswanso has quit IRC | 14:40 | |
| *** absubram has joined #tripleo | 14:41 | |
| *** dmsimard has joined #tripleo | 14:41 | |
| *** tiswanso has joined #tripleo | 14:41 | |
| openstackgerrit | Matthew Thode proposed openstack/diskimage-builder: add support for package-installs to gentoo https://review.openstack.org/270597 | 14:45 |
| openstackgerrit | Merged openstack/python-tripleoclient: Enable TLS endpoints through os-cloud-config https://review.openstack.org/268160 | 14:47 |
| openstackgerrit | Jiri Stransky proposed openstack/tripleo-common: Revert "Pin puppetlabs-mysql to get CI going" https://review.openstack.org/272119 | 14:47 |
| jistr | shardy, gfidente: with the approach of doing controllers via Heat and the rest manually, we'll probably need to disable Puppet from running, as it's not safe in general (e.g. to re-run Puppet on computes, where old OPM would be present, but new THT manifest would be fed from the undercloud) | 14:51 |
| jistr | so i'm thinking i'll add an option no-op OS::TripleO::*PostDeployment | 14:52 |
| *** trozet has joined #tripleo | 14:52 | |
| jistr | so far i haven't hit any problem even though i was re-running puppet, but i don't think that's the general case | 14:52 |
| jistr | heh if i no-op OS::TripleO::ControllerPostDeployment, then the update wouldn't run at all if we used the ControllerPrePuppet way :) | 14:57 |
| jistr | perhaps we could keep the UpdateWorkflow hook then | 14:58 |
| openstackgerrit | John Trowbridge proposed openstack-infra/tripleo-ci: [NO MERGY] Test a update in trunk repository version https://review.openstack.org/229789 | 15:00 |
| *** jprovazn has quit IRC | 15:00 | |
| *** pradk has joined #tripleo | 15:02 | |
| openstackgerrit | Merged openstack/tripleo-puppet-elements: Trove Integration https://review.openstack.org/233236 | 15:02 |
| shardy | jistr: yeah, you could just map ComputePostDeployment to OS::Heat::None | 15:03 |
| shardy | or we could split DeployIdentifier so we have ComputeDeployIdentifier etc | 15:03 |
| * jistr searches if OS:Heat::None is in liberty Heat | 15:04 | |
| shardy | jistr: it is, I wrote it ;) | 15:04 |
| jistr | shardy: alright, thanks :) | 15:04 |
| shardy | It basically does the same as a noop empty nested stack, but it's a bit more convenient (and faster) | 15:04 |
| shardy | because you don't need to know anything about the parameters/properties or outputs | 15:05 |
| gfidente | shardy, finally! | 15:05 |
| gfidente | :P | 15:05 |
| gfidente | time to use it | 15:05 |
| jistr | yeah not having to put effort into matching the interface when no-opping sounds cool | 15:06 |
| *** mbound has quit IRC | 15:07 | |
| shardy | jistr: I've been meaning to try it for all out various noop templates | 15:07 |
| shardy | I think it may make things a bit faster than creating lots of noop templates | 15:07 |
| shardy | although some we can't replace because we mangle parameters into outputs with no resources in the template | 15:08 |
| * shardy personally hates that | 15:08 | |
| shardy | jistr: FYI, to enable moving towards split-stack, I've been looking at moving the hiera deployments, and UpdateDeployment to *-post.yaml | 15:10 |
| shardy | https://github.com/openstack/tripleo-heat-templates/blob/master/puppet/compute.yaml#L652 | 15:10 |
| shardy | If we do that, just disabling e.g ComputePostDeployment would also be enough to stop the yum update, even if we did use UpdateIdentifier | 15:10 |
| shardy | (I know we're not using that for the pacemaker upgrades atm) | 15:11 |
| shardy | The other options would be to remove UpdateDeployment completely and instead always do updates via *PrePuppet | 15:11 |
| *** athomas has quit IRC | 15:11 | |
| shardy | e.g like we're doing in https://review.openstack.org/#/c/253276/11/environments/major_upgrade_pacemaker.yaml | 15:11 |
| jistr | shardy: we'd need to introduce *PrePuppet for other node types then | 15:13 |
| shardy | jistr: Yeah | 15:13 |
| shardy | Not something to worry about now, I was just thinking how we might rework things | 15:13 |
| jistr | shardy: btw i think that moving all those "software deployment/config/update" things into *-post.yaml is good. I can't wait for the split stack :) | 15:13 |
| openstackgerrit | Merged openstack/puppet-tripleo: loadbalancer: add Gnocchi API support https://review.openstack.org/252015 | 15:14 |
| *** cmyster has joined #tripleo | 15:14 | |
| *** cmyster has quit IRC | 15:14 | |
| *** cmyster has joined #tripleo | 15:14 | |
| openstackgerrit | Lennart Regebro proposed openstack/python-tripleoclient: More tests for overcloud deploy https://review.openstack.org/254767 | 15:15 |
| jistr | if we had UpdateDeployment in -post.yaml, we could do multi-step orchestrated operations in it, but only within a single role (not orchestrating controllers vs. computes for example), and we couldn't disable Puppet without disabling the UpdateDeployment too | 15:15 |
| *** athomas has joined #tripleo | 15:16 | |
| jistr | that's not a problem in general, but it doesn't replace UpdateWorkflow functionality this way | 15:16 |
| *** fgimenez has quit IRC | 15:20 | |
| *** fgimenez has joined #tripleo | 15:23 | |
| *** fgimenez has joined #tripleo | 15:23 | |
| openstackgerrit | Ben Nemec proposed openstack/tripleo-docs: Document deploying the overcloud with ssl https://review.openstack.org/265006 | 15:24 |
| *** trown is now known as trown|brb | 15:24 | |
| *** mcornea has quit IRC | 15:24 | |
| *** mcornea has joined #tripleo | 15:25 | |
| *** fgimenez has quit IRC | 15:25 | |
| *** egafford has quit IRC | 15:25 | |
| *** fgimenez has joined #tripleo | 15:26 | |
| *** fgimenez has quit IRC | 15:26 | |
| *** fgimenez has joined #tripleo | 15:26 | |
| *** chlong has joined #tripleo | 15:27 | |
| openstackgerrit | Brent Eagles proposed openstack/tripleo-heat-templates: Enable configuration of Neutron QoS https://review.openstack.org/272131 | 15:29 |
| *** trown|brb is now known as trown | 15:33 | |
| *** dtantsur|brb is now known as dtantsur | 15:37 | |
| *** masco has joined #tripleo | 15:38 | |
| openstackgerrit | Steven Hardy proposed openstack/tripleo-heat-templates: Show os-net-config applied via script, not oac https://review.openstack.org/271450 | 15:38 |
| *** chlong has quit IRC | 15:40 | |
| *** chlong has joined #tripleo | 15:40 | |
| *** chlong has quit IRC | 15:43 | |
| *** mgould has quit IRC | 15:43 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Support the deployment of Ceph over IPv6 https://review.openstack.org/272089 | 15:44 |
| *** chlong has joined #tripleo | 15:44 | |
| a2hill | Can we please get some eyes on this: https://review.openstack.org/#/c/271190/ | 15:44 |
| *** paramite is now known as paramite|afk | 15:44 | |
| a2hill | https://bugs.launchpad.net/tripleo/+bug/1536844 | 15:44 |
| openstack | Launchpad bug 1536844 in tripleo "get-pip.py failing, using an outdated location for get-pip.py script" [High,In progress] - Assigned to Phillip Toohill (phillip-toohill) | 15:44 |
| a2hill | This is a critical bug for us. | 15:45 |
| *** dshulyak_ has quit IRC | 15:48 | |
| openstackgerrit | Matthew Thode proposed openstack/diskimage-builder: add support for package-installs to gentoo https://review.openstack.org/270597 | 15:49 |
| openstackgerrit | Merged openstack/tripleo-image-elements: Drop pip-and-virtualenv element https://review.openstack.org/271190 | 15:52 |
| *** mgould has joined #tripleo | 15:54 | |
| *** liverpooler has quit IRC | 15:55 | |
| *** cmyster has quit IRC | 15:58 | |
| *** cmyster has joined #tripleo | 15:59 | |
| *** cmyster has quit IRC | 15:59 | |
| *** cmyster has joined #tripleo | 15:59 | |
| *** cwolferh has joined #tripleo | 16:02 | |
| openstackgerrit | John Trowbridge proposed openstack/tripleo-heat-templates: Remove empty value for wsrep_notify_cmd https://review.openstack.org/272149 | 16:04 |
| trown | bandini: I think ^ fixes the latest issue with puppet-mysql | 16:05 |
| *** stendulker has quit IRC | 16:05 | |
| *** mbound has joined #tripleo | 16:05 | |
| openstackgerrit | John Trowbridge proposed openstack-infra/tripleo-ci: [NO MERGY] Test a update in trunk repository version https://review.openstack.org/229789 | 16:06 |
| bandini | trown: I have just launched a full ha deploy job with that change. I will comment on the review (~1h) | 16:10 |
| *** tosky has quit IRC | 16:12 | |
| trown | bandini: nice, thanks! your BZ helped me quickly figure that out | 16:12 |
| *** lazy_prince has joined #tripleo | 16:12 | |
| *** chlong has quit IRC | 16:15 | |
| marios | jistr: the 'heat' way is that now , since you don't have the requirement to do *all* nodes, you can use the config pre? | 16:15 |
| *** chlong has joined #tripleo | 16:15 | |
| *** mkovacik has quit IRC | 16:15 | |
| *** tiswanso has quit IRC | 16:16 | |
| *** tosky has joined #tripleo | 16:16 | |
| marios | jistr: (*all* nodes via _heat_ I mean) | 16:16 |
| jistr | marios: i think that's still a bit unclear. Using ControllerPrePuppet would mean that you have to run Puppet too, at least on the controllers. I'm not sure ATM whether that's a good idea. | 16:17 |
| *** paramite|afk is now known as paramite | 16:17 | |
| jistr | we might be better off delaying all Puppet runs until the whole deployment is upgraded | 16:17 |
| *** egafford has joined #tripleo | 16:18 | |
| marios | jistr: ok thanks... i just caught a bit of that earlier | 16:18 |
| *** tiswanso has joined #tripleo | 16:20 | |
| trown | shardy: confirmed https://review.openstack.org/229789 fixes stable/liberty HA job, and it will be needed on master to move the current-tripleo delorean pin | 16:20 |
| trown | rather it fixes stable/liberty's ability to deploy HA, I did not test it with tripleoci | 16:21 |
| *** mbound has quit IRC | 16:22 | |
| *** sbalukoff has quit IRC | 16:24 | |
| shardy | trown: thanks! I didn't spot the pin earlier so was confused as to why master was still passing | 16:24 |
| trown | shardy: ya I started looking at moving the delorean pin and saw we had old puppet-mysql | 16:25 |
| trown | pinning the puppet modules can put us in a weird spot where stable/liberty is using newer puppet code than master | 16:26 |
| shardy | trown: Yeah, also using opm for stable CI is confusing things | 16:27 |
| shardy | I was discussing that with EmilienM earlier, I'll try to send a patch later switching that over to build from source like the master job | 16:27 |
| trown | ah cool, that would at least allow the same pin for master and stable/liberty | 16:27 |
| trown | ideally we dont need to pin at all though | 16:28 |
| *** rcernin has quit IRC | 16:31 | |
| marios | jistr: :) pick it up tomorrow? if we have some time, would also be nice to briefly chat about that presentation that gfidente will give next friday (joking aside perhaps some of this upgrades design/tradeoff/issues will go into those upgrades slides too? | 16:34 |
| jistr | marios: yes, ++ on mentioning upgrades a bit more | 16:34 |
| gfidente | this talk was a trap | 16:34 |
| gfidente | I told you | 16:34 |
| * marios looks forward to hearing gfidente's presentation | 16:35 | |
| jistr | marios: and sure thing, let's follow up in the morning, i'll get the no-op stuff sorted out this evening | 16:35 |
| marios | jistr: ok othanks man | 16:35 |
| jistr | it worked but i need to put it into a different place in the patch chain | 16:35 |
| *** jeh has joined #tripleo | 16:35 | |
| jistr | so i'll just retest and post it | 16:35 |
| jistr | gfidente: +1 re trap :D | 16:36 |
| marios | jistr: gfidente: don't hate me | 16:36 |
| * marios looks at hewbrocca | 16:36 | |
| jistr | :D no worries guys, we're gonna make it work somehow | 16:36 |
| * hewbrocca ducks | 16:37 | |
| jistr | we have a nice video thanks to marios | 16:37 |
| marios | jistr: hope so :) (yah i'm sure will be fine) | 16:37 |
| jistr | and the rest is just typey typey^W^Wtalkey talkey | 16:37 |
| marios | jistr: esp. since gfidente is presenting most of it | 16:37 |
| marios | so jo worries | 16:37 |
| marios | s/jo/no | 16:37 |
| *** jeh has quit IRC | 16:37 | |
| *** paramite is now known as paramite|afk | 16:37 | |
| gfidente | we can use jesse pinkman words | 16:38 |
| gfidente | to describe it better | 16:38 |
| marios | gfidente: haha. /me wondering how long will take you to swear at me | 16:38 |
| gfidente | why at you? | 16:38 |
| gfidente | marios, that means you know breaking bad? | 16:38 |
| marios | gfidente: cos i keep implying that you will give the presentation by yourself | 16:38 |
| marios | gfidente: breaking bad was great | 16:39 |
| gfidente | so I'll bring the voice recordings | 16:39 |
| *** paramite|afk is now known as paramite | 16:42 | |
| *** bnemec has joined #tripleo | 16:45 | |
| *** oshvartz has quit IRC | 16:47 | |
| *** lazy_prince has quit IRC | 16:54 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Support the deployment of Ceph over IPv6 https://review.openstack.org/272089 | 16:56 |
| *** jprovazn has joined #tripleo | 16:56 | |
| *** Madkiss has quit IRC | 17:00 | |
| *** chlong has quit IRC | 17:02 | |
| *** chlong has joined #tripleo | 17:02 | |
| *** jpeeler has quit IRC | 17:04 | |
| *** chlong has quit IRC | 17:04 | |
| *** chlong has joined #tripleo | 17:05 | |
| *** mkovacik has joined #tripleo | 17:05 | |
| *** masco has quit IRC | 17:07 | |
| openstackgerrit | Jiri Stransky proposed openstack/tripleo-heat-templates: Introduce update/upgrade workflow https://review.openstack.org/271358 | 17:09 |
| *** fgimenez has quit IRC | 17:13 | |
| jaosorior | jistr, bnemec: Do you guys have a chance to check out this CR? https://review.openstack.org/#/c/268651/ it is already being used downstream, but it's part of the stuff that's missing to merge for TLS | 17:15 |
| jaosorior | bnemec: Thanks dude | 17:16 |
| trozet | EmilienM: i dont know if michchap contacted you, but we have added the aodh support to tht for stable/liberty | 17:17 |
| jistr | jaosorior: +2+A | 17:17 |
| jaosorior | jistr: yay :D | 17:18 |
| jistr | jaosorior: though it's probably not going to merge until https://review.openstack.org/#/c/268653/ is in :/ | 17:18 |
| EmilienM | trozet: it's not even merged in master, how did you do? | 17:18 |
| EmilienM | https://review.openstack.org/241408 | 17:18 |
| openstackgerrit | Derek Higgins proposed openstack/tripleo-common: WIP: split overcloud setup out of pingtest https://review.openstack.org/270631 | 17:18 |
| trozet | EmilienM: we have a local copy we use with OPNFV: https://github.com/trozet/opnfv-tht/commit/dc6bbb7b3d46096925b1fb2ea65be7ab2114d714 | 17:19 |
| trozet | EmilienM: the only issue right now, is for some reason the keystone endpoints do not get created | 17:19 |
| trozet | EmilienM: even though they are defined in that patch | 17:19 |
| jaosorior | jistr: Is that the case? That's in stable/liberty. The patch for master has already merged | 17:19 |
| bnemec | tripleo.sh: unrecognized option '--overcloud-pingtest' | 17:20 |
| jistr | jaosorior: yea but i think gerrit doesn't have that smarts built in (yet) unfortunately | 17:20 |
| bnemec | Umm, did we break stable CI with the ping test? | 17:20 |
| jaosorior | bummer | 17:20 |
| *** mburned has quit IRC | 17:20 | |
| *** egafford has quit IRC | 17:20 | |
| *** rhallisey has quit IRC | 17:20 | |
| *** julim has quit IRC | 17:20 | |
| *** cwolferh has quit IRC | 17:20 | |
| *** jpeeler has joined #tripleo | 17:20 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Support the deployment of Ceph over IPv6 https://review.openstack.org/272089 | 17:21 |
| *** ifarkas has quit IRC | 17:22 | |
| EmilienM | trozet: you'll need https://review.openstack.org/#/c/244162/14 | 17:23 |
| *** devvesa has quit IRC | 17:23 | |
| *** mbound has joined #tripleo | 17:23 | |
| trozet | EmilienM: ah | 17:24 |
| trozet | EmilienM: I think its also missing a keystone-then-aodh pacemaker constraint | 17:26 |
| trozet | EmilienM: thanks for the link, will try it | 17:26 |
| openstackgerrit | Ben Nemec proposed openstack/tripleo-common: Adds a simple overcloud tenant vm ping test to tripleo.sh https://review.openstack.org/272191 | 17:27 |
| *** mbound has quit IRC | 17:27 | |
| bnemec | shardy: jistr: We need ^ to unblock stable CI. | 17:27 |
| shardy | bnemec: thanks, we also need a backport of https://review.openstack.org/#/c/272149/ before CI will actually pass | 17:28 |
| shardy | I'll propose that now so we can get CI running it | 17:29 |
| jistr | shardy: is that breaking the CI too? i think we still have a pin for puppet-mysql | 17:30 |
| bnemec | shardy: Sounds good. | 17:30 |
| shardy | jistr: the pin doesn't work for stable/liberty because we're still using opm | 17:30 |
| jistr | ah ok | 17:31 |
| openstackgerrit | Steven Hardy proposed openstack/tripleo-heat-templates: Remove empty value for wsrep_notify_cmd https://review.openstack.org/272194 | 17:31 |
| bnemec | Although strangely I saw a CI job that deployed the overcloud, and didn't fail until it tried to run the ping test. | 17:32 |
| bnemec | I would have expected the deploy to fail based on that commit message. | 17:32 |
| shardy | bnemec: was it an HA job? | 17:32 |
| bnemec | shardy: Ah, no. | 17:32 |
| *** julim has joined #tripleo | 17:33 | |
| *** mbound has joined #tripleo | 17:33 | |
| trown | bnemec: ya it is only the pacemaker env that is affected | 17:34 |
| *** trown is now known as trown|lunch | 17:35 | |
| gfidente | and isn't the cmd empty by default anyway? | 17:36 |
| *** rhallisey has joined #tripleo | 17:36 | |
| jistr | shardy, bnemec: should we base one of the changes on top of the other, to see CI passing at least on the topmost one before merging? | 17:36 |
| jistr | i mean the 2 backports which should fix stable/liberty | 17:37 |
| *** aufi has quit IRC | 17:39 | |
| *** mburned has joined #tripleo | 17:39 | |
| shardy | jistr: oh wow, we actually landed the pingtest to toci_instack without the -common backport | 17:39 |
| shardy | I assumed bnemec just meant the policy on having that pass before approval | 17:39 |
| * bnemec did not even remember that we had a stable branch for tripleo-common | 17:40 | |
| shardy | we should reall move tripleo.sh out of tripleo-common into tripleo-ci I think | 17:40 |
| bnemec | We need to get stable jobs running on toci proposed changes. | 17:40 |
| shardy | we discussed it a while back IIRC | 17:40 |
| shardy | I can't think of any patch that hasn't been backported, which seems like a bit of a waste of effort | 17:40 |
| bnemec | Yeah, I think tripleo.sh was intended to be branchless. | 17:41 |
| bnemec | Hence the separate stable job settings in the script. | 17:41 |
| derekh | +1 to moving it to tripleo-ci | 17:42 |
| shardy | Ok, let me rebase https://review.openstack.org/272194 on your change bnemec | 17:42 |
| shardy | then if that all passes, including the ha job, we can land them both | 17:42 |
| jistr | i'd be +1 as well, just wondering -- do we need tripleo.sh packaged for some reason? (trown?) | 17:42 |
| bnemec | tripleo.sh should not be packaged | 17:43 |
| derekh | jistr: nope, we shouldn't | 17:43 |
| bnemec | It's a dev and test tool. | 17:43 |
| shardy | after that we can move tripleo.sh and fix stable CI so it uses source not opm | 17:43 |
| derekh | it should be a DEV only script for upstead | 17:43 |
| derekh | *upstream | 17:43 |
| *** hewbrocca has quit IRC | 17:46 | |
| *** nico_auv has quit IRC | 17:46 | |
| jistr | makes sense :) | 17:50 |
| jistr | ttyt | 17:51 |
| *** jistr has quit IRC | 17:51 | |
| trown|lunch | ya +1 to not packaging tripleo.sh | 17:51 |
| *** derekh has quit IRC | 17:52 | |
| trown|lunch | shardy: derekh, I forgot about https://review.openstack.org/#/c/270890/2 as well... we will need something like that to move the master delorean pin, because there is a newer heatclient | 17:52 |
| openstackgerrit | Miles Gould proposed openstack/python-tripleoclient: Remove tripleoclient.baremetal wrapper https://review.openstack.org/265336 | 17:53 |
| openstackgerrit | Miles Gould proposed openstack/python-tripleoclient: Use Ironic API v1.11 to support ENROLL state https://review.openstack.org/272206 | 17:53 |
| openstackgerrit | Emilien Macchi proposed openstack/tripleo-heat-templates: Manage keystone initialization directly in t-h-t manifests https://review.openstack.org/244162 | 17:53 |
| ayoung | running tripleo.sh Gives me CREATE_FAILED ResourceInError: resources.Controller: Went to status ERROR due to "Message: No valid host was found. There are not enough hosts available., Code: 500" | 17:55 |
| *** mbound has quit IRC | 17:55 | |
| ayoung | there are 2 baremetal VMs, but neither are turned on. | 17:55 |
| *** julim_ has joined #tripleo | 17:57 | |
| *** Goneri has quit IRC | 17:57 | |
| *** dshulyak has joined #tripleo | 17:58 | |
| *** julim has quit IRC | 17:59 | |
| *** dtantsur is now known as dtantsur|afk | 18:01 | |
| *** lucasagomes is now known as lucas-dinner | 18:01 | |
| openstackgerrit | Steven Hardy proposed openstack-infra/tripleo-ci: WIP move tripleo.sh into tripleo-ci repo https://review.openstack.org/272210 | 18:02 |
| openstackgerrit | Steven Hardy proposed openstack-infra/tripleo-ci: WIP move tripleo.sh into tripleo-ci repo https://review.openstack.org/272210 | 18:03 |
| *** ukalifon has quit IRC | 18:04 | |
| EmilienM | shardy: I was about sending this patch https://review.openstack.org/#/c/272194/ | 18:04 |
| EmilienM | but yu did it :-) | 18:04 |
| shardy | ayoung: check out http://docs.openstack.org/developer/tripleo-docs/troubleshooting/troubleshooting.html#no-valid-host-found-error | 18:04 |
| EmilienM | shardy: before merging it, I would like to see a puppet run without OPM package (specially because it is a backport) | 18:05 |
| shardy | ayoung: it normally means there's an issue with ironic or nova, where a node isn't in the expected state so the scheduler thinks there are no nodes | 18:05 |
| shardy | check via nova hypervisor-stats | 18:05 |
| shardy | EmilienM: Ok, I'm going to work on that asap, but do you really think we must block the change on that? | 18:06 |
| shardy | our CI is completely blocked until that fix lands | 18:06 |
| EmilienM | shardy: no | 18:06 |
| EmilienM | do not block it | 18:06 |
| EmilienM | just saying it won't be actually tested, since we use OPM | 18:06 |
| shardy | EmilienM: Yeah, I agree it's not ideal, I'm working on gettting the from-source thing working | 18:07 |
| openstackgerrit | Merged openstack/tripleo-heat-templates: Allow container template to recognize an update https://review.openstack.org/265368 | 18:07 |
| shardy | that move-tripleo.sh patch was the first step, so I don't have to do it in two places | 18:07 |
| EmilienM | makes sense | 18:08 |
| *** AJaeger has joined #tripleo | 18:09 | |
| *** jcoufal has quit IRC | 18:10 | |
| AJaeger | tripleo cores, could you approve a change to remove argparse, please? You don't support python 2.6 anymore... https://review.openstack.org/#/c/270376/ https://review.openstack.org/#/c/270377/ https://review.openstack.org/#/c/270378/ | 18:11 |
| AJaeger | thanks, shardy! | 18:12 |
| *** dshulyak has quit IRC | 18:12 | |
| shardy | AJaeger: np! | 18:13 |
| *** shivrao has joined #tripleo | 18:13 | |
| openstackgerrit | Ryan Hallisey proposed openstack/tripleo-heat-templates: Allow container template to recognize an update https://review.openstack.org/272212 | 18:13 |
| *** Goneri has joined #tripleo | 18:14 | |
| mgould | can anyone help me with why http://logs.openstack.org/06/272206/1/check/check-osc-plugins/e06fea2/console.html is failing? It looks related to the removal of tuskar | 18:15 |
| mgould | it dies when it tries to run `pip install` in a python-tuskarclient directory | 18:15 |
| AJaeger | mgould: indeed, we need to first merge the infra change. Let me ask on #openstack-infra for that | 18:16 |
| mgould | AJaeger, thanks! | 18:16 |
| *** akrivoka has quit IRC | 18:16 | |
| mgould | once that's gone through, do I need to recheck all my builds? | 18:16 |
| openstackgerrit | Zane Bitter proposed openstack/tripleo-common: Catch NotFound when listing events during updates https://review.openstack.org/272213 | 18:17 |
| AJaeger | mgould: yes, once https://review.openstack.org/270869 is in and the jobs are updated (usually 30 mins after merge), you can recheck | 18:18 |
| *** trown|lunch is now known as trown | 18:19 | |
| mgould | AJaeger, thanks! | 18:19 |
| mgould | I'll have a look tomorrow morning, since it's nearly dinner time here :-) | 18:19 |
| AJaeger | mgould: if the patch does not get reviewed soonish, please ping again yourself in case I'm not around. | 18:20 |
| AJaeger | mgould: enjoy! | 18:20 |
| mgould | AJaeger, thanks :-) | 18:20 |
| *** mgould has quit IRC | 18:20 | |
| ayoung | shardy, I have two that look like this | 18:22 |
| ayoung | | UUID | Name | Instance UUID | Power State | Provisioning State | Maintenance | | 18:22 |
| ayoung | +--------------------------------------+------+---------------+-------------+--------------------+-------------+ | 18:22 |
| ayoung | | bd1b3c32-afce-473c-91df-130cba50106f | None | None | power off | available | False | | 18:22 |
| *** jaosorior has quit IRC | 18:23 | |
| ayoung | shardy, what state does it expect them to be in? I would guess power off before it does anything to them, no? | 18:23 |
| shardy | ayoung: Yeah that looks OK so it may be some other issue - what count does nova hypervisor-stats give you? | 18:24 |
| *** jaosorior has joined #tripleo | 18:24 | |
| ayoung | shardy, count | 2 | 18:24 |
| *** yamahata has joined #tripleo | 18:24 | |
| ayoung | shardy, I'm taking the defaults on everything for the undercloud. Is it possible I need to up the resources on something>? | 18:25 |
| shardy | ayoung: What you have should be OK for a basic two node deployment | 18:26 |
| trown | ayoung: 'No valid host was found' is a very generic nova error, I always start by looking in /var/log/nova/nova-scheduler.log to find what filter returned 0 hosts | 18:26 |
| trown | ayoung: that usually points directly to the problem | 18:26 |
| ayoung | shardy, trown is Pacemaker and the HA stuff part of Tripleo upstream? It is right? | 18:26 |
| trown | ayoung: ya... did you try to do HA? | 18:27 |
| shardy | ayoung: it is, but it's not used for simple 2-node nonha deployments | 18:27 |
| ayoung | trown, not yet, but there is an issue. I need to get Keystone running in HTTPD | 18:27 |
| ayoung | and that is getting messed up by... | 18:28 |
| shardy | downstream IIRC pacemaker is enabled even for single controllers | 18:28 |
| ayoung | https://bugs.launchpad.net/puppet-keystone/+bug/1524927 | 18:28 |
| openstack | Launchpad bug 1524927 in puppet-keystone "Puppet keystone cannot restart HTTPD managed by Pacemaker" [Critical,Confirmed] | 18:28 |
| ayoung | as well as the fact that many of the other services depend on Keystone (I think) in an HA world | 18:28 |
| jaosorior | shardy: that is correct | 18:29 |
| ayoung | We had someone file an internal bug "Remove keystone constraints and add the openstack-core dummy resource in its place" | 18:29 |
| *** paramite has quit IRC | 18:29 | |
| ayoung | but I think that is an upstream issue to solve first, no? | 18:29 |
| *** chlong has quit IRC | 18:29 | |
| *** chlong has joined #tripleo | 18:29 | |
| trown | ayoung: you rapidly moved out of my area of expertise :p | 18:29 |
| ayoung | trown, see, all of this is out of my area....I am first trying to get tripleo to runm, and athen going to try and get EmilienM 's patch for Keystone HTTPD to apply | 18:30 |
| ayoung | but I know all this is futile if the HA stuff falls apart with Keystone in HTTPD | 18:30 |
| gfidente | ayoung, I think you should get in touch with bandini or fabbione first to see how the ha arch should change to host keystone in httpd | 18:31 |
| trown | I have no idea about the second bit, but the nova error is pretty straight forward to track down | 18:31 |
| ayoung | trown, OK, let's start there. I'm in the nova log. | 18:32 |
| *** chlong has quit IRC | 18:32 | |
| ayoung | What am I looking for? | 18:32 |
| *** chlong has joined #tripleo | 18:32 | |
| trown | ayoung: `grep filter` will get you close | 18:32 |
| ayoung | 2016-01-23 16:25:46.171 18859 DEBUG nova.scheduler.filter_scheduler [req-9bdfcb42-a13b-44d1-81d3-65f62b7bd729 b9fe43dab743443eb01c9a09d683e18f f93f6d78b5ab4930980ac905fce579b9 - - -] There are 0 hosts available but 1 instances requested to build. select_destinations /usr/lib/python2.7/site-packages/nova/scheduler/filter_scheduler.py:93 | 18:32 |
| ayoung | something like that? | 18:32 |
| shardy | ayoung: did you by any chance attempt an overcloud deploy immediately after registering the nodes? | 18:33 |
| ayoung | shardy, I was using tripleo.sh. So, I think that is exactly what it does. Is there a race condition? | 18:33 |
| ayoung | I can try it step by step | 18:33 |
| shardy | ayoung: but how did you use it, the --all mode? | 18:33 |
| ayoung | shardy, yes | 18:33 |
| ayoung | shardy, this time --all; | 18:33 |
| shardy | ayoung: yes, there is a possible race condition, between --register-nodes and --overcloud-deploy | 18:34 |
| shardy | because it takes a few seconds for nova to pick up the new nodes in ironic | 18:34 |
| ayoung | shardy, OK, how do I get to a good state here? can I rerun the overloud-deploy, or are there things I need to clean up first? | 18:34 |
| shardy | ayoung: heat stack-delete overcloud | 18:35 |
| shardy | then heat stack-list until it's gone | 18:35 |
| ayoung | stack_status CREATE_FAILED | 18:35 |
| shardy | ayoung: then you can either run tripleo.sh --overcloud-deploy again | 18:35 |
| shardy | or run openstack overcloud deploy --templates | 18:35 |
| ayoung | stack_status DELETE_IN_PROGRESS | 18:35 |
| ayoung | OK...looks good. Let me stick with tripleo.sh for the moment | 18:36 |
| ayoung | shardy, I seee this | 18:36 |
| ayoung | There are 2 ironic nodes with no profile that will not be used: bd1b3c32-afce-473c-91df-130cba50106f, 79705cd9-7dbb-4bee-99fe-2a69f38fe871 | 18:36 |
| shardy | Hmm, the profile should not be mandatory, I never run introspection or assign profiles locally | 18:38 |
| shardy | what does ironic node-show <id> show ? | 18:38 |
| fabbione | gfidente: we already provided the steps in the BZ | 18:39 |
| fabbione | gfidente: the change based on our input | 18:39 |
| shardy | ayoung: also nova flavor-show baremetal | 18:41 |
| ayoung | shardy, the tripleo.sh --overcloud-deploy is still running | 18:44 |
| shardy | ayoung: Ah, that's good I guess! :) | 18:44 |
| ayoung | let me leave that and try a second log in | 18:44 |
| shardy | I assumed your no profile was an error | 18:44 |
| *** akuznetsov has joined #tripleo | 18:46 | |
| ayoung | shardy, http://paste.openstack.org/show/484917/ | 18:46 |
| shardy | https://github.com/openstack/python-tripleoclient/blob/master/tripleoclient/utils.py#L742 | 18:47 |
| shardy | ayoung: I think that's an over-zealous warning | 18:47 |
| shardy | it means they're not tagged to be matched to a profile/role, but that shouldn't matter unless you've set up flavors to match those profiles | 18:48 |
| shardy | your paste looks OK to me | 18:48 |
| shardy | hopefully you just hit the race where nova lags ironic nodes getting added | 18:48 |
| *** pcaruana has joined #tripleo | 18:51 | |
| *** cwolferh has joined #tripleo | 18:52 | |
| *** jaosorior has quit IRC | 18:57 | |
| *** akuznetsov has quit IRC | 18:57 | |
| *** jhenner has quit IRC | 18:58 | |
| openstackgerrit | Steven Hardy proposed openstack-infra/tripleo-ci: WIP move tripleo.sh into tripleo-ci repo https://review.openstack.org/272210 | 18:59 |
| *** sthillma has joined #tripleo | 19:00 | |
| *** shardy is now known as shardy_afk | 19:00 | |
| *** olap has quit IRC | 19:01 | |
| *** openstackgerrit has quit IRC | 19:02 | |
| *** openstackgerrit has joined #tripleo | 19:02 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/tripleo-common: Updated from global requirements https://review.openstack.org/268549 | 19:04 |
| *** akuznetsov has joined #tripleo | 19:07 | |
| *** sbalukoff has joined #tripleo | 19:10 | |
| ayoung | shardy_afk, CREATE_FAILED ResourceInError: resources.Controller: Went to status ERROR due to "Message: No valid host was found. There are not enough hosts available., Code: 500" | 19:10 |
| *** tosky has quit IRC | 19:12 | |
| *** mbound has joined #tripleo | 19:21 | |
| *** electrofelix has quit IRC | 19:22 | |
| AJaeger | once mgloud is back, feel free to tell him that tuskar repos are retried and he can recheck his changes | 19:24 |
| openstackgerrit | Emilien Macchi proposed openstack/tripleo-heat-templates: WIP - Deploy Aodh services, replacing Ceilometer Alarm https://review.openstack.org/241408 | 19:25 |
| *** jhenner has joined #tripleo | 19:25 | |
| *** AJaeger has left #tripleo | 19:28 | |
| *** akuznetsov has quit IRC | 19:30 | |
| *** pcaruana has quit IRC | 19:32 | |
| *** pcaruana has joined #tripleo | 19:36 | |
| *** pcaruana has quit IRC | 19:40 | |
| openstackgerrit | Merged openstack/tripleo-common: Rename stack_update to _stack_update https://review.openstack.org/235976 | 19:41 |
| *** pcaruana has joined #tripleo | 19:43 | |
| *** rcernin has joined #tripleo | 19:44 | |
| *** mbound has quit IRC | 19:45 | |
| *** pcaruana has quit IRC | 19:56 | |
| *** dprince has quit IRC | 19:57 | |
| *** jprovazn has quit IRC | 19:58 | |
| *** yamahata has quit IRC | 20:05 | |
| *** dmacpher has quit IRC | 20:15 | |
| *** mbound has joined #tripleo | 20:22 | |
| *** yamahata has joined #tripleo | 20:22 | |
| *** dmacpher has joined #tripleo | 20:28 | |
| openstackgerrit | Matthew Thode proposed openstack/diskimage-builder: add support for package-installs to gentoo https://review.openstack.org/270597 | 20:31 |
| *** dprince has joined #tripleo | 20:39 | |
| openstackgerrit | ayoung proposed openstack/tripleo-heat-templates: puppet: run keystone in wsgi https://review.openstack.org/213175 | 20:40 |
| ayoung | EmilienM, please +1 that if it meets your approval | 20:40 |
| EmilienM | ayoung: I'll look ! thanks | 20:40 |
| *** devananda has joined #tripleo | 20:54 | |
| *** devananda has quit IRC | 20:54 | |
| *** devananda has joined #tripleo | 20:56 | |
| *** tserong has quit IRC | 20:56 | |
| *** devananda has quit IRC | 20:57 | |
| EmilienM | ayoung: syntax is not passing | 21:03 |
| EmilienM | ayoung: http://logs.openstack.org/75/213175/13/check/gate-tripleo-heat-templates-puppet-syntax-3-dsvm-centos7/6f61096/console.html#_2016-01-25_20_40_58_136 | 21:04 |
| *** chlong has quit IRC | 21:07 | |
| *** chlong has joined #tripleo | 21:08 | |
| openstackgerrit | ayoung proposed openstack/tripleo-heat-templates: puppet: run keystone in wsgi https://review.openstack.org/213175 | 21:16 |
| *** davidlenwell has quit IRC | 21:16 | |
| *** davidlenwell has joined #tripleo | 21:18 | |
| *** ChanServ sets mode: +v davidlenwell | 21:18 | |
| *** devananda has joined #tripleo | 21:19 | |
| *** gfidente has quit IRC | 21:23 | |
| EmilienM | dprince, slagle: https://review.openstack.org/#/c/244162/ is here for long time and blocking some important patches... I'm wondering what to do | 21:23 |
| slagle | EmilienM: i'd like to test updates with it | 21:26 |
| slagle | EmilienM: deploy a cloud with out that. deploy with it, see if anything breaks | 21:26 |
| slagle | *update with it | 21:26 |
| slagle | that's my opinion anyway, given it's a bit invasive | 21:27 |
| slagle | i can try and do that this week | 21:27 |
| *** julim_ has quit IRC | 21:27 | |
| ayoung | EmilienM, https://review.openstack.org/#/c/213175/ now passes | 21:32 |
| dprince | EmilienM: I think we've actually solved the previous upgrades issue, which was the same as the undercloud's | 21:32 |
| EmilienM | ayoung: well, only syntax/lint is passing, let's see functional jobs | 21:33 |
| dprince | EmilienM: I would like to see someone manually test it as well, but I think it is probably okay myself | 21:33 |
| *** tserong has joined #tripleo | 21:33 | |
| dprince | EmilienM: I guess what I'm saying is having tested the upgrade on the undercloud I'm fairly sure the same issue applies in the overcloud too | 21:33 |
| *** weshay has quit IRC | 21:36 | |
| *** weshay_xchat has joined #tripleo | 21:37 | |
| *** tserong has quit IRC | 21:38 | |
| *** chlong has quit IRC | 21:39 | |
| *** chlong has joined #tripleo | 21:39 | |
| *** tserong has joined #tripleo | 21:42 | |
| *** weshay_xchat is now known as weshay | 21:45 | |
| EmilienM | dprince, slagle: I have the same question for https://review.openstack.org/#/c/235423/ - a lot of patches are rebased everyday because this one is not merged. Do we know what is blocking? | 21:46 |
| *** trown is now known as trown|outttypeww | 21:46 | |
| *** bvandenh has quit IRC | 21:46 | |
| dprince | EmilienM: ack, I would like to pull in shardy on this tomorrow morning perhaps | 21:48 |
| EmilienM | I hate asking for reviews | 21:49 |
| dprince | EmilienM: no, thanks for poking about these things. It helps highlight them | 21:50 |
| dprince | EmilienM: but to be fair there are some patches that have been up even longer :/ | 21:50 |
| *** NobodyCa1 has joined #tripleo | 21:51 | |
| *** NobodyCa1 has quit IRC | 21:52 | |
| slagle | EmilienM: i think the last review comments explain what's going on with it | 21:52 |
| slagle | seems there isnt agreement on if shardy's suggestion should be implemented now, or later | 21:53 |
| EmilienM | slagle: my bad :( sorry I missed it | 21:53 |
| *** davidlenwell has quit IRC | 21:55 | |
| EmilienM | slagle: about https://review.openstack.org/#/c/235423/ I would suggest next time we do smaller chuncks | 21:55 |
| EmilienM | it would help in reviews & testing | 21:55 |
| EmilienM | and probably help to merge patches safer & faster | 21:56 |
| *** davidlenwell has joined #tripleo | 21:56 | |
| *** ChanServ sets mode: +v davidlenwell | 21:56 | |
| slagle | i can't disagree there :) | 21:57 |
| *** rhallisey has quit IRC | 22:07 | |
| *** rpothier has quit IRC | 22:11 | |
| *** shardy_afk has quit IRC | 22:11 | |
| *** lblanchard has quit IRC | 22:19 | |
| *** jayg is now known as jayg|g0n3 | 22:19 | |
| *** Goneri has quit IRC | 22:30 | |
| ayoung | bnemec, on https://review.openstack.org/#/c/213175/14/puppet/manifests/overcloud_controller_pacemaker.pp how to deal withe the pacemaker comment? The other one is straight forward enought | 22:31 |
| *** rcernin has quit IRC | 22:31 | |
| ayoung | should it be just the "if $enable_load_balancer {" stanza, but with keystone replaced by apache, like we did above? | 22:32 |
| bnemec | ayoung: I'm not super-familiar with how this all works, but I _think_ it should be enough just to wrap the haproxy-then-keystone-constraint with the if $enable_load_balancer. | 22:32 |
| bnemec | The main thing is that we can't wait on haproxy when we're not deploying haproxy. | 22:32 |
| *** Marga__ has joined #tripleo | 22:33 | |
| ayoung | bnemec, like this | 22:33 |
| ayoung | http://paste.openstack.org/show/484937/ | 22:33 |
| *** Marga__ has quit IRC | 22:33 | |
| *** Marga__ has joined #tripleo | 22:34 | |
| *** Marga__ has quit IRC | 22:34 | |
| *** davidlenwell has quit IRC | 22:34 | |
| *** Marga_ has quit IRC | 22:35 | |
| *** Marga_ has joined #tripleo | 22:35 | |
| bnemec | ayoung: I was thinking this: http://paste.openstack.org/show/484938/ | 22:35 |
| ayoung | bnemec, I'm not seeing the difference between yours and mine | 22:36 |
| *** davidlenwell has joined #tripleo | 22:36 | |
| *** ChanServ sets mode: +v davidlenwell | 22:36 | |
| ayoung | I had droppped the if, but other than that ,they are identical | 22:37 |
| bnemec | ayoung: Mine has an if around the block. The only thing I see changed in yours is s/keystone/apache/ | 22:37 |
| ayoung | OK | 22:37 |
| bnemec | The if is the thing I'm concerned about. | 22:37 |
| ayoung | OK. I got that. let me resubmit | 22:37 |
| bnemec | The name changes all look fine to me. | 22:37 |
| bnemec | Modulo my lack of expertise with pacemaker constraints, of course. ;-) | 22:38 |
| openstackgerrit | ayoung proposed openstack/tripleo-heat-templates: puppet: run keystone in wsgi https://review.openstack.org/213175 | 22:38 |
| ayoung | bnemec, I think that is what you were saying | 22:38 |
| ayoung | https://review.openstack.org/#/c/213175/14..15/puppet/manifests/overcloud_controller_pacemaker.pp | 22:39 |
| bnemec | ayoung: Yes, but I think there are two copies of that block now. The one at 1057 needs to be removed. | 22:39 |
| ayoung | bnemec, the one inside the if $pacemaker_master { block, but leave the if block itself, right? | 22:41 |
| bnemec | ayoung: Yeah, I think you just need to remove lines 1057-1065. | 22:42 |
| *** tiswanso has quit IRC | 22:42 | |
| bnemec | I'm pretty sure puppet wouldn't allow two haproxy-then-keystone-constraint objects anyway. | 22:43 |
| openstackgerrit | ayoung proposed openstack/tripleo-heat-templates: puppet: run keystone in wsgi https://review.openstack.org/213175 | 22:43 |
| bnemec | We might have to rename those once we have more than just keystone running in httpd. | 22:43 |
| ayoung | bnemec, there ya go | 22:43 |
| bnemec | /aside | 22:43 |
| * bnemec looks | 22:43 | |
| bnemec | ayoung: lgtm | 22:44 |
| ayoung | coo | 22:45 |
| bnemec | Although, there are a bunch of other ifs that got dropped later in the file too. :-/ | 22:46 |
| bnemec | Makes me wonder if it was intentional, or there was a rebase issue at some point. | 22:48 |
| ayoung | bnemec, were they in the origianl patch? | 22:52 |
| ayoung | If they were there, then the droppage is mine, and probably due to me being too eager to get rid of all the ssl-isms from Keystone | 22:52 |
| bnemec | ayoung: No, looks like they were dropped before your first patch set. | 22:52 |
| ayoung | Keystone eventlet SSL does a bunch of cert generation that should not be done | 22:52 |
| ayoung | as you will end up with self signed certs. | 22:53 |
| bnemec | Oh wait, you pushed more patches than I realized. | 22:53 |
| * bnemec looks further back | 22:53 | |
| ayoung | bnemec, yeah. only the first version was EmilienM 's | 22:54 |
| *** mcornea has quit IRC | 22:55 | |
| ayoung | er up to 9 was EmilienM 's....That is what happends when I get involved on a patch where I don't know the tech. | 22:56 |
| bnemec | ayoung: It looks like a rebase issue. The ifs were added in https://github.com/openstack/tripleo-heat-templates/commit/e82c50d8f964bb7eb6ce84248b8c1cc78413a431 , which merged after this was initially proposed. | 22:59 |
| bnemec | I think we do need to keep them then. | 22:59 |
| *** dprince has quit IRC | 23:02 | |
| *** chlong has quit IRC | 23:02 | |
| ayoung | bnemec, OK... | 23:02 |
| *** chlong has joined #tripleo | 23:03 | |
| *** chlong has quit IRC | 23:04 | |
| *** chlong has joined #tripleo | 23:05 | |
| openstackgerrit | Zane Bitter proposed openstack/tripleo-common: Catch NotFound when listing events during updates https://review.openstack.org/272213 | 23:10 |
| openstackgerrit | ayoung proposed openstack/tripleo-heat-templates: puppet: run keystone in wsgi https://review.openstack.org/213175 | 23:11 |
| *** absubram has quit IRC | 23:13 | |
| ayoung | bnemec, can we remove things like keystone-to-neutron-server-constraint | 23:14 |
| bnemec | ayoung: I don't think so. We still need to ensure that keystone starts before neutron. | 23:15 |
| ayoung | why] | 23:15 |
| ayoung | bnemec, It needs to be running before someone makes a query, but not before starting it | 23:16 |
| bnemec | ayoung: I think it's our way of ensuring nobody queries neutron before keystone is up. | 23:16 |
| ayoung | bnemec, I think that might be what is messing up Keystone in HTTPD. It s causing restarts of HTTPD | 23:17 |
| * ayoung does not know this from anything other than reports | 23:17 | |
| bnemec | Hmm, I wouldn't think that it should be. | 23:18 |
| bnemec | We shouldn't be triggering a bunch of restarts of keystone, and restarting services that depend on keystone shouldn't cause keystone to restart. | 23:18 |
| bnemec | You may have reached the end of my knowledge of Pacemaker though. ;-) | 23:19 |
| *** trozet has quit IRC | 23:22 | |
| ayoung | bnemec, mine was passed long ago | 23:25 |
| *** chlong has quit IRC | 23:29 | |
| *** chlong has joined #tripleo | 23:29 | |
| *** jtomasek has quit IRC | 23:32 | |
| *** sthillma has quit IRC | 23:42 | |
| *** jtomasek has joined #tripleo | 23:44 | |
| *** david-lyle has quit IRC | 23:44 | |
| *** sthillma has joined #tripleo | 23:46 | |
| *** mbound has quit IRC | 23:49 | |
| *** chlong has quit IRC | 23:58 | |
| *** chlong has joined #tripleo | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!