16:00:15 <Jeffrey4l> #startmeeting kolla
16:00:16 <openstack> Meeting started Wed Apr 11 16:00:15 2018 UTC and is due to finish in 60 minutes.  The chair is Jeffrey4l. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:20 <openstack> The meeting name has been set to 'kolla'
16:01:06 <Jeffrey4l> #topic rollcall
16:01:21 <spsurya__> \0
16:01:21 <duonghq> o/
16:01:29 <gbuteau> o/
16:01:34 <mgiles> o/
16:01:39 <caoyuan> o/
16:02:08 <egonzalez> O/
16:02:09 <chason> o/
16:02:13 <bmace> o/
16:02:31 <Jeffrey4l> #topic announcements
16:02:55 <Jeffrey4l> stable-policy tag has been removed in kolla
16:03:03 <Jeffrey4l> more info please check http://lists.openstack.org/pipermail/openstack-dev/2018-April/129265.html
16:03:17 <Jeffrey4l> this is a big change for us.
16:03:47 <Jeffrey4l> with this, we could backport bp
16:04:06 <Jeffrey4l> but we need care about this . and should not break anything.
16:04:13 <Jeffrey4l> any other from community?
16:05:05 <Jeffrey4l> ok. seems we have no agenda today.
16:05:13 <Jeffrey4l> #topic Open Discussion
16:05:22 <Jeffrey4l> i have one for open discussion.
16:05:59 <Jeffrey4l> we have a long review queue now. some develop start complain that patches have no reviewers to check.
16:07:00 <Jeffrey4l> so please check the review queue we you have times.
16:07:28 <gema> sgoel: you around?
16:07:32 <gema> Jeffrey4l: we had a question
16:07:38 <Jeffrey4l> moreover, there are two tools may helpful to find which patch is ready to ber reviewed
16:07:41 <Jeffrey4l> gema, yep
16:07:58 <Jeffrey4l> #link https://github.com/openstack/gertty
16:08:13 <gema> Jeffrey4l: trying to get sgoel to explain, but essentially we are seeing a race condition when deploying with the kolla tools container
16:08:22 <gema> and were wondering if this is something someone else has seen
16:08:36 <Jeffrey4l> #link https://github.com/openstack/gerrit-dash-creator
16:08:45 <Jeffrey4l> check these interesting tools :D
16:09:11 <sgoel> hey
16:09:14 <Jeffrey4l> gema, could you explain the details?
16:09:21 <gema> sgoel: go for it, please
16:09:25 <Jeffrey4l> what kind of race condition?
16:09:48 <Jeffrey4l> hello sgoel
16:10:24 <sgoel> I am seeing a connectivity error caught in kolla_toolbox.py
16:10:40 <Jeffrey4l> sgoel, better provides some error logs.
16:10:55 <sgoel> this is seen when exec_start is called
16:10:57 <Jeffrey4l> i don't see the error
16:11:20 <Jeffrey4l> sgoel, could you file a bug with some logs ?
16:11:26 <sgoel> If I put in a delay or a retry the deployment works fine
16:11:37 <sgoel> Ok, I can post logs.
16:11:58 <gema> Jeffrey4l: so this is something you think may be worth looking into
16:12:02 <gema> Jeffrey4l: we'll raise a bug then
16:13:10 <Jeffrey4l> this issue may be possible. i guss some former container isn't ready during calling kolla_toolbox.py
16:13:33 <gema> ack
16:13:54 <Jeffrey4l> if so, adding a task in kolla-ansible to wait until the container is really ready will be helpful
16:14:09 <Jeffrey4l> anyway, need logs to check what's really going on.
16:14:12 <gema> Jeffrey4l: we have no better way of blocking than delays?
16:14:41 <Jeffrey4l> gema, yes. adding a wait task in ansible is the only way.
16:14:52 <gema> Jeffrey4l: ok
16:15:02 <Jeffrey4l> i don't think docker itself have better solution.
16:15:14 <gema> Jeffrey4l: no, kubernetes however.. ;)
16:15:36 <Jeffrey4l> another solution is: adding the wait logical in the later container.
16:15:38 <gema> Jeffrey4l: either way, we will raise it and investigate/try to fix
16:15:43 <Jeffrey4l> kubernetes just retry many times.
16:15:55 <gema> yup
16:15:58 <Jeffrey4l> which is not a good solution i think
16:16:04 <gema> good point
16:16:20 <Jeffrey4l> we could use the same method: use "until" in the ansible task.
16:16:27 <Jeffrey4l> then retry until it works. :D
16:16:53 <gema> that may be better than random delay tasks, tbh, but until we know why it is failing, we don't really know
16:17:11 <gema> Jeffrey4l: you gave us enough to get started, thanks
16:17:15 <Jeffrey4l> yeah
16:17:20 <sgoel> Jeffrey4l: agree that a retry framework will be good. i'll raise a bug today with logs
16:17:51 <Jeffrey4l> cool. thanks gema sgoel
16:18:06 <rwellum> Kubernetes, you'd create an init container that does a readiness probe :)
16:18:44 <Jeffrey4l> rwellum, it should not related to init container. just adding readiness probe for pod(container) is OK.
16:18:44 <rwellum> Pretty much no-one just allows the containers to crash and retry anymore
16:18:57 <rwellum> Yeah - so something similar then?
16:19:43 <Jeffrey4l> rwellum, but it is not good. say it as "endless retry". :/
16:20:22 <Jeffrey4l> i still think we need some to do the orchestration works on k8s. k8s don't handle this properly.
16:21:10 <Jeffrey4l> then any voluteer for next open topic?
16:21:25 <gbuteau> o/
16:21:33 <Jeffrey4l> gbuteau please
16:21:40 <gbuteau> hi all, wrt blueprint neutron-ipam-driver-infoblox - thank you for accepting the kolla changes, although I notice that the kolla-ansible changes are still pending
16:21:49 <gbuteau> #link https://review.openstack.org/#/c/555032/
16:22:05 <gbuteau> i'm happy to provide more information or updates if needed
16:23:10 <Jeffrey4l> gbuteau cool. i added to my review queue, and left a "recheck" to rerun the CI jobs again.
16:23:26 <Jeffrey4l> make sure it won't break anything.
16:23:27 <gbuteau> thank you Jeffrey41
16:23:34 <Jeffrey4l> i will review it tomorrow
16:24:25 <Jeffrey4l> anything esle? gbuteau
16:24:43 <gbuteau> That's it, thx again Jeffrey41
16:25:25 <Jeffrey4l> thanks gbuteau.
16:25:32 <Jeffrey4l> any other topics?
16:27:39 <Jeffrey4l> seems no
16:27:53 <Jeffrey4l> let us end the meeting now.
16:28:01 <Jeffrey4l> thanks guys for coming.
16:28:06 <Jeffrey4l> have a good day :D
16:28:09 <Jeffrey4l> #endmeeting