14:00:29 #startmeeting tripleo 14:00:29 #topic agenda 14:00:29 * review past action items 14:00:29 * one off agenda items 14:00:29 * bugs 14:00:29 * Projects releases or stable backports 14:00:29 * CI 14:00:30 * Specs 14:00:30 Meeting started Tue Sep 19 14:00:29 2017 UTC and is due to finish in 60 minutes. The chair is mwhahaha. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:30 * open discussion 14:00:31 Anyone can use the #link, #action and #info commands, not just the moderatorǃ 14:00:31 Hi everyone! who is around today? 14:00:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:34 The meeting name has been set to 'tripleo' 14:00:35 o/ 14:00:43 o/ 14:00:43 o/ 14:00:45 o/ 14:00:49 hi 14:00:51 o/ 14:00:54 o/ 14:01:03 o/ 14:01:04 o/ 14:01:37 o/ 14:01:54 o/ 14:01:54 o/ 14:01:57 o/ 14:02:17 #topic review past action items 14:02:18 Janki Chhatbar proposed openstack/tripleo-common master: Add healthcheck for ODL container https://review.openstack.org/504864 14:02:23 drop alert on Bug 1713832 and ping storage/zaqar people to investigate 14:02:25 bug 1713832 in tripleo "Object PUT failed for zaqar_subscription" [Critical,In progress] https://launchpad.net/bugs/1713832 - Assigned to Marios Andreou (marios-b) 14:02:35 so we dropped the alert but I think it's still open 14:02:55 the last update points to https://bugs.launchpad.net/swift/+bug/1715177 14:02:56 Launchpad bug 1715177 in OpenStack Object Storage (swift) "Container cache management is racy" [Medium,Confirmed] 14:02:58 mwhahaha: yeah i had a look and therve had a lead at some point but we still didn't find the root 14:03:10 mwhahaha: i can revisit though i didn't look there all of last week 14:03:16 o/ 14:03:32 marios: it looks like it's a swift problem so we should poke the storage folks 14:03:32 mwhahaha: is it still relatively infrequent? (1/day?) /me fetches elastic recheck query 14:03:45 yea i don't think it's happening as much as it used to 14:03:56 o/ 14:03:57 http://status.openstack.org/elastic-recheck/#1713832 this one 14:04:41 looks like it's still a problem so we just need to track it in swift 14:04:45 moving on to the next item 14:04:51 shardy to look into version impacts in ci jobs related to Bug 1714361 14:04:52 bug 1714361 in tripleo "Mistral on gates is old and does not have the required patches" [Critical,Fix released] https://launchpad.net/bugs/1714361 - Assigned to Alex Schultz (alex-schultz) 14:05:08 shardy: i think we fixed that with the priority on repos 14:05:12 mwhahaha: Yeah I think we did that 14:05:37 mwhahaha to retarget rc2 bugs if not release critical - i believe this was done but I will double check today 14:05:43 shardy to look at how to reduce # of services deployed on ovb 14:05:51 mwhahaha, I found the issue, but it's stuck lacking tests 14:06:10 therve: ok let me know if there's anything we can do to help 14:06:14 mwhahaha: I made a start on that but need to revisit 14:06:21 shardy: ok 14:06:31 #action shardy to look at how to reduce # of services deployed on ovb (continued) 14:06:31 mwhahaha, https://review.openstack.org/500978 14:06:46 #link https://review.openstack.org/#/c/500942/ WIP ovb testing 14:07:11 that's it on the past action items 14:07:17 mwhahaha, Maybe they can relax the constraints so that we merge the fix 14:07:57 therve: since that's the upstream swift stuff they'd have to sign off on that unfortunately. 14:08:07 Yeah 14:08:15 cschwede: ^^ 14:08:30 mwhahaha, We could disable the cache middleware in swift, but it might makes things slow(er) 14:08:49 therve: since this is on the undercloud, how much perf impact? 14:09:04 mwhahaha, Hard to know really... 14:09:11 We use quite a bit between the templates and zaqar 14:09:14 o/ 14:09:38 shardy: will have another look later today, thx! 14:09:45 o/ 14:10:07 ok moving on 14:10:09 #topic one off agenda items 14:10:09 #link https://etherpad.openstack.org/p/tripleo-meeting-items 14:10:14 URGENT TRIPLEO TASKS NEED ATTENTION 14:10:14 https://bugs.launchpad.net/tripleo/+bug/1716914 14:10:15 Launchpad bug 1716914 in tripleo "CI: all periodic jobs fail to install undercloud: No resource and no name in property hash in nova_manage instance" [Critical,Triaged] 14:10:15 https://bugs.launchpad.net/tripleo/+bug/1717274 14:10:16 https://bugs.launchpad.net/tripleo/+bug/1717279 14:10:16 https://bugs.launchpad.net/tripleo/+bug/1717545 14:10:17 https://bugs.launchpad.net/tripleo/+bug/1717959 14:10:18 Launchpad bug 1717274 in tripleo "ocata2pike: broken package dependency during upgrade" [Critical,In progress] - Assigned to wes hayutin (weshayutin) 14:10:20 Launchpad bug 1717279 in tripleo "RDO registry denies images uploads" [Critical,Triaged] 14:10:21 Launchpad bug 1717545 in tripleo "nova populates incomplete cell_mappings" [Critical,Triaged] 14:10:22 Launchpad bug 1717959 in tripleo "RDO upstream-centos-7 nodepool image is out of date" [Critical,Triaged] - Assigned to Alan Pevec (apevec) 14:10:22 Michele Baldessari proposed openstack/puppet-pacemaker master: Parse pcs auth output with all pcs versions https://review.openstack.org/505284 14:10:30 mwhahaha, shardy: regarding the swift issue, couldn't the client be modified to make sure to create the container first? 14:10:54 mwhahaha: for https://bugs.launchpad.net/tripleo/+bug/1717274 weshay posted https://review.openstack.org/#/c/505268/ 14:10:59 tdasilva, It does create the container first :) 14:11:19 marios: ok i'll take a look after the meeting 14:11:25 mwhahaha: seems to just be missing base repo for the ethtool package which is a dependency (yum update fails) 14:11:34 tdasilva, I didn't get a nice workaround for it, but I'm opened to suggestions 14:11:50 mwhahaha: thanks (kudos jpena who called that last week already see comment #2 ) 14:12:03 therve: can you share the client side code? 14:12:32 tdasilva, https://github.com/openstack/zaqar/blob/master/zaqar/storage/swift/utils.py#L45 14:14:09 moving on 14:14:13 #topic bugs 14:14:13 #link https://launchpad.net/tripleo/+milestone/queens-1 14:14:32 we're into queens so please take a look at the bugs since we're going for stability this cycle 14:15:03 any specific bugs people wish to talk about (besides that upgrade bug) 14:16:56 i'll take that as a no 14:17:03 #topic projects releases or stable backports 14:17:30 Anyoen have any backports they need reviews on? 14:18:22 is infra still hosed wrt releases? 14:18:30 i believe so 14:18:37 i saw mention of that this morning 14:18:47 therve: ack, taking a look 14:19:18 speaking of things that are hosed 14:19:19 #topic CI 14:19:28 lol 14:19:45 the puppet-nova fix for the nova_manage command is in the gate 14:20:04 so that should merge soon (~30 mins) 14:20:13 so that should clear up some of the failures 14:20:21 any other outstanding issues that folks can talk about? 14:20:56 only other issue is in the periodic pipeline we are setting up on rdo infra 14:21:20 i saw there some bugs around images in the pipeline 14:21:23 has that been addressed? 14:21:38 needs a new infra image, but that is blocked by some new project that was created without a repo 14:21:50 https://git.openstack.org/cgit/openstack/networking-lagopus 14:21:59 nice 14:22:19 when that resolves we can rebuild the image in rdoproject and hopefully get the pipeline working 14:23:20 any other CI items? 14:24:17 sounds like no 14:24:23 #topic specs 14:24:23 #link https://review.openstack.org/#/q/project:openstack/tripleo-specs+status:open 14:24:31 please take some time to review any open specs 14:24:44 also retarget specs that might still be pointed to pike 14:25:18 additionally if there were any larger items from the PTG that probably should have a spec, now would be the time to propose it 14:25:38 anyone have a specific spec they would like to point out? 14:28:13 moving on 14:28:14 #topic open discussion 14:28:16 anything else? 14:28:20 if not i'll close in 2 mins 14:29:14 thoughts on moving logic of containerized deployments up to the puppet modules for the services specifically? 14:29:34 probably not 14:29:54 at the moment we only really using the providers for the config generation 14:30:11 so the actual docker container is created/deployed without puppet? 14:30:39 yea we're using kolla for the container build process but puppet for the config of the containers 14:31:11 and then we have a tool for running the containers themselves 14:31:14 mmkay. i was hoping we can work together so to have service_name => docker type of thing 14:31:41 i wanna see what a poc looks like first but it was just an idea :> 14:31:58 that might simplify what we have 14:32:06 but we don't actually use the serivce items anymore 14:33:07 https://github.com/openstack/tripleo-heat-templates/blob/b5c18ded6a6122053fb3a6557c063a44a90e41b9/docker/docker-puppet.py 14:33:32 mnaser: https://github.com/openstack/tripleo-heat-templates/blob/master/common/deploy-steps-tasks.yaml shows the overview of our deploy steps now 14:33:58 we use ansible to drive a tool called paunch, which consumes some json files and starts docker containers 14:34:20 so it's kind of moving in the direction of only using puppet for config generation at this point 14:34:46 (it is also used for some bootstrapping) 14:35:05 mnaser: always good to discuss any ideas tho :) 14:35:41 i think containers are here to stay (kolla or whatever) but there is many ways of orchestrating them, given you have a ton of existing puppet infra, it might be easier than you having to maintain one more tool 14:36:33 i'm sure it'll present a few challenges but i plan to take sometime to do a poc of a simpler puppet service and then show it and see what it looks like (and gather feedback) 14:36:59 ideally i'd love to see it adopted by tripleo so catering to that use case and avoiding duplicating effort would be nice :> 14:37:06 mnaser: yeah, understood - in this case paunch was derived from some code we already had in the heat docker-cmd hook 14:37:30 mnaser: seeing the poc would be good, but it does kinda sound like a different direction than where we're headed at this point 14:37:59 gotcha, i'll take feedback anyways "this is where we shot ourself in the foot when we were taking this approach" is helpful :) 14:39:53 good questions 14:39:55 anything else? 14:41:15 ok thanks everyone 14:41:17 #endmeeting