17:00:16 <devkulkarni> #startmeeting Solum Team Meeting
17:00:17 <openstack> Meeting started Tue Feb 23 17:00:16 2016 UTC and is due to finish in 60 minutes.  The chair is devkulkarni. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:21 <openstack> The meeting name has been set to 'solum_team_meeting'
17:00:23 <devkulkarni> #topic Roll Call
17:00:28 <devkulkarni> Devdatta Kulkarni
17:03:35 <vijendar> o/
17:03:43 <devkulkarni> hi vijendar
17:03:55 <vijendar> Hi devkulkarni
17:04:06 <devkulkarni> right now its just you and me here
17:04:21 <devkulkarni> lets wait for few minutes to see if others are joining us
17:04:37 <vijendar> ok
17:04:40 <dimtruck> o/
17:04:41 <devkulkarni> I haven't seen ashishjain or pt_15 on irc channel in last few days
17:04:45 <devkulkarni> hi dimtruck
17:04:47 <dimtruck> hi!
17:04:49 <devkulkarni> good to see you
17:05:02 <devkulkarni> here is the agenda for today: https://wiki.openstack.org/wiki/Meetings/Solum#Agenda_for_2016-02-23_1700_UTC
17:05:18 <muralia1> o/
17:05:23 <devkulkarni> hi muralia1
17:05:28 <devkulkarni> nice to see you
17:05:32 <muralia1> hi everyone
17:05:41 <devkulkarni> we have dimtruck and vijendar here today
17:06:04 <devkulkarni> alright.. I think we have a quorum so lets get started
17:06:13 <devkulkarni> the agenda for today is: https://wiki.openstack.org/wiki/Meetings/Solum#Agenda_for_2016-02-23_1700_UTC
17:06:23 <devkulkarni> take a moment to look through it
17:06:48 <devkulkarni> we are light on things to cover today, so probably we should be able to finish bit early
17:06:59 <devkulkarni> #topic Announcements
17:07:05 <devkulkarni> I have none prepared
17:07:26 <devkulkarni> do any of you have any announcements for the team?
17:07:53 <devkulkarni> alright..
17:08:18 <dimtruck> no sir
17:08:22 <devkulkarni> let me directly go to reviews as there were no action items from last time
17:08:29 <devkulkarni> thanks dimtruck
17:08:44 <devkulkarni> #topic Blueprint/Bug Review and Discussion
17:08:55 <devkulkarni> We have three items in this topic today
17:09:06 <devkulkarni> 1) Replace oslo incubation package for log with oslo log library
17:09:12 <devkulkarni> #link https://review.openstack.org/#/c/273566/
17:09:19 <devkulkarni> let me give background about this patch
17:10:03 <devkulkarni> Basically it replaces the log module from oslo incubator, which we have been using till now, with the corresponding module from the oslo library
17:10:15 <devkulkarni> I tried the patch yesterday in my devstack setup
17:10:45 <devkulkarni> it doesn't affect end-to-end functionality, but it is polluting the screens with an error message
17:10:51 <devkulkarni> I have added my comments to the patch
17:11:19 <devkulkarni> I discussed with rbradfor about a path to resolve this
17:11:56 <devkulkarni> basically the plan is to remove the support_trace formatting from plugin.sh and try the patch again
17:12:23 <devkulkarni> if it works, we will know that we need to fix the support_trace issue somehow
17:12:41 <devkulkarni> do any of you remember why we had added support_trace in our log formatting line?
17:13:30 <devkulkarni> you will find the discussion about the issue in the patch review history
17:13:35 <devkulkarni> https://review.openstack.org/#/c/273566/
17:14:09 <dimtruck> i believe most of our logs are trace logs
17:14:17 <dimtruck> not sure if we need them at all
17:14:22 <dimtruck> but they've been around as long as i can remember
17:14:26 <devkulkarni> you mean, the log level is trace?
17:15:17 <devkulkarni> dimtruck: hmm.. same here.. it has been there since the beginning. I am going to remove it and try the patch
17:15:25 <dimtruck> ok
17:15:27 <dimtruck> yes devkulkarni
17:15:32 <devkulkarni> will update the patch with the result
17:15:35 <devkulkarni> cool
17:15:44 <dimtruck> the downside is we won't see a lot of the internal flows
17:15:53 <dimtruck> but not sure if we need them on by default
17:15:58 <devkulkarni> ah I see
17:16:14 <devkulkarni> we would want to see internal flows.. at least in the devstack setup
17:16:56 <devkulkarni> if we are not going to see internal flows if we remove support_trace then I would be reluctant to merge this patch as it is
17:17:14 <dimtruck> well, we can also validate it by removing support_trace
17:17:17 <devkulkarni> in any case, it is right now polluting our screens with error message
17:17:24 <devkulkarni> dimtruck: yep
17:17:39 <devkulkarni> so before reaching any conclusion, let me try that first
17:17:58 <devkulkarni> if we remove support_trace and are still able to see all the internal flows, then we are fine
17:18:15 <dimtruck> right
17:18:31 <devkulkarni> good discussion.. any more thoughts on this topic? will update the patch after my experimentation
17:18:50 <devkulkarni> ok, lets move on to the second item
17:18:55 <devkulkarni> 2) Solum horizon plugin
17:19:14 <devkulkarni> recently we received a patch to enable the plugin in our devstack env
17:19:18 <devkulkarni> https://review.openstack.org/#/c/277197/
17:19:32 <devkulkarni> please don't approve the patch yet, as I want to try it out in my devstack setup
17:20:08 <dimtruck> ok
17:20:09 <devkulkarni> but while looking at the patch I thought of checking out solum-dashboard repository and enabling it in my devstack setup
17:20:27 <devkulkarni> and while doing that I ran into an issue, which I have captured in this bug:
17:20:29 <devkulkarni> https://bugs.launchpad.net/solum/+bug/1546596
17:20:29 <openstack> Launchpad bug 1546596 in Solum "integration-of-solum-dashboard-in-horizon-fails" [High,New] - Assigned to Han Manjong (aksmj8855)
17:21:02 <devkulkarni> have any of you worked on horizon before? or have basic understanding of how it works, and how the plugins work?
17:21:17 <dimtruck> looking
17:21:35 <devkulkarni> if you have any thoughts/comments regarding this, please add it to the bug description
17:22:01 <dimtruck> strange
17:22:03 <devkulkarni> dimtruck: thanks.. on the bug, there was a comment towards the end by Hans.. I need to go back and reply to his questions
17:22:18 <devkulkarni> dimtruck: listening..
17:22:23 <dimtruck> cool...but i'd think that'd be provided by devstack
17:22:42 <devkulkarni> what would be provided by devstack?
17:22:48 <dimtruck> https://github.com/openstack/solum-dashboard/blob/master/requirements.txt#L3
17:22:55 <dimtruck> yeah, so it's <1.8
17:23:29 <devkulkarni> I see.. so we need to bump it to 1.8?
17:23:29 <dimtruck> we'd need to update these since they haven't been touched since 2014
17:23:39 <devkulkarni> that is correct
17:23:57 <dimtruck> there had to be a reason it was set to those specific verisons...
17:24:04 <dimtruck> we'd probably want to bump everything and see what breaks
17:24:22 <dimtruck> and everything = pbr and django
17:24:23 <devkulkarni> I was thinking of resuscitating this repo now that we have a working end-to-end cli experience
17:24:29 <dimtruck> devkulkarni: +1
17:24:47 <devkulkarni> it will be awesome if we can get the horizon plugin working before the austin summit
17:25:10 <dimtruck> i can take a look and see what needs to be done to do that
17:25:22 <dimtruck> if you don't mind creating a bug and assigning it to me
17:25:32 <devkulkarni> dimtruck: that would be super helpful.. you are awesome.. sure I can do that
17:25:59 <devkulkarni> dimtruck: no pressure though.. I am also trying to see if pt_15 or hans can help us on this
17:26:07 <devkulkarni> but haven't seen them around in a while
17:26:23 <dimtruck> awesome! thanks
17:26:36 <devkulkarni> https://bugs.launchpad.net/solum/+bug/1546596
17:26:36 <openstack> Launchpad bug 1546596 in Solum "integration-of-solum-dashboard-in-horizon-fails" [High,New] - Assigned to Han Manjong (aksmj8855)
17:27:20 <devkulkarni> thanks to you dimtruck.. lets try to see if we can get the plugin in good shape before the summit
17:27:28 <devkulkarni> I will also start looking into it
17:27:48 <devkulkarni> cool.. let me go to the third item on the list
17:27:55 <devkulkarni> 3) Cutting a release
17:28:00 <devkulkarni> http://lists.openstack.org/pipermail/openstack-dev/2016-February/086152.html
17:28:11 <devkulkarni> this needs to happen this week
17:28:30 <devkulkarni> I will submit patches for this in next couple of days
17:28:50 <devkulkarni> there is one patch related to scaling which is failing gate test
17:29:11 <devkulkarni> I will try to get it fixed
17:29:23 <devkulkarni> and then submit patches to the release repo
17:29:34 <devkulkarni> just wanted to give update regarding that
17:29:42 <dimtruck> devkulkarni: is this the last release before the summit?
17:29:51 <devkulkarni> dimtruck: no, this is not
17:30:11 <devkulkarni> if you check the email, what has been suggested is all the projects that are following the intermediate release cycle
17:30:17 <dimtruck> understood
17:30:18 <dimtruck> thank you
17:30:27 <devkulkarni> do a release in three weeks (announced two weeks back)
17:30:55 <devkulkarni> the idea is that the release team wants to have some release done before April 1st so that they can use it to cut mitaka
17:31:31 <devkulkarni> they were suggesting that, just as a precaution we do a release now so that there will be something available, should we end up not cutting a release before April 1st again
17:31:39 <dimtruck> understood
17:31:43 <devkulkarni> in case we do cut a release before April 1st, they will use the latest
17:31:44 <devkulkarni> cool
17:32:17 <devkulkarni> alright.. so those were the patches/bugs that I wanted to discuss
17:32:29 <devkulkarni> are there any bugs/patches that you all want to discuss?
17:33:05 <devkulkarni> alright.. lets move on to open discussion
17:33:09 <devkulkarni> #topic Open Discussion
17:33:16 <devkulkarni> 1) Proposal to separate design summit from OpenStack conference
17:33:22 <devkulkarni> #link http://lists.openstack.org/pipermail/openstack-dev/2016-February/087161.html
17:33:31 <devkulkarni> I don't know if you have been following this thread
17:33:55 <devkulkarni> the discussion here is to separate the design summit from the main conference
17:34:17 <devkulkarni> whenever you get a chance, I would encourage to read it
17:34:44 <dimtruck> wow!
17:34:45 <devkulkarni> it will affect all of us as far as attending openstack conferences is concerned
17:35:08 <devkulkarni> dimtruck: yep
17:35:14 <devkulkarni> there are pros and cons
17:35:30 <devkulkarni> but mostly ppl seem to be in favor of the proposal
17:35:43 <dimtruck> right
17:35:45 <dimtruck> seems legit
17:36:08 <devkulkarni> the main issue that I feel is it won't allow developers to meet users
17:36:28 <devkulkarni> at least not all developers would get to meet the users
17:36:43 <devkulkarni> lets see how this pans out
17:36:56 <dimtruck> right
17:36:56 <devkulkarni> I am expecting major discussion about this in the Austin summit
17:37:22 <devkulkarni> The second upstream update is
17:37:28 <devkulkarni> 2) Lock fencing for addressing race conditions/mutual exclusion with distributed services
17:37:32 <devkulkarni> http://lists.openstack.org/pipermail/openstack-dev/2016-February/087032.html
17:37:37 <devkulkarni> An interesting thread..
17:38:02 <devkulkarni> basically it talks about issue of how to ensure that concurrent services don't end up getting into race condition
17:38:43 <devkulkarni> I feel it is relevant to us as we have seen situations where two solum-deployer run into race condition when updating the db corresponding to two successive app deploy actions
17:39:29 <devkulkarni> the solution that is proposed, which actually comes from the discussion of redis's redlock, is to use a monotonically increasing id with each db call
17:40:11 <dimtruck> right
17:40:11 <devkulkarni> then the storage layer needs to reject calls which it receives that contain an id which is less than the id that it currently knows as the latest
17:40:45 <devkulkarni> our approach is somewhat similar .. we use timestamps
17:40:55 <devkulkarni> to reject updates from an earlier timestamp
17:41:10 <devkulkarni> but this works only when we have multiple deployer processes running on the same node
17:41:24 <devkulkarni> it won't work when we have deployers on different nodes
17:41:43 <devkulkarni> so, thought we should be aware of what others are thinking about this
17:41:49 <dimtruck> right
17:41:55 <devkulkarni> would definitely recommend checking the thread and the references there in
17:42:02 <devkulkarni> cool
17:42:11 <devkulkarni> third update from upstream is:
17:42:16 <devkulkarni> 3) Non-blocking IO in Nova -- discussion that may be relevant for solum-deployer
17:42:20 <devkulkarni> http://lists.openstack.org/pipermail/openstack-dev/2016-February/087169.html
17:42:28 <devkulkarni> one more interesting thread
17:43:00 <devkulkarni> the issue that this is discussing is, the main thread of nova seems to making blocking io calls
17:43:41 <devkulkarni> the thread has suggestions on how to solve it
17:44:02 <devkulkarni> I thought we can check the mechanisms suggested there for our deployer
17:44:23 <devkulkarni> in the deployer we are busy waiting on heat status check..
17:44:43 <dimtruck> +1
17:44:44 <dimtruck> definitely!
17:44:55 <devkulkarni> while we are not making blocking calls, the issue of busy waiting has similar effect on our code
17:44:59 <devkulkarni> that the deployer is blocked
17:45:17 <devkulkarni> may be we can find some ideas from that thread to be used in our code
17:45:34 <devkulkarni> dimtruck: yep
17:45:57 <devkulkarni> and final item on upstream updates is related to documentation generation
17:46:05 <devkulkarni> 4) Should not use 'git log' in generating documentation
17:46:09 <devkulkarni> http://lists.openstack.org/pipermail/openstack-dev/2016-February/086917.html
17:46:51 <devkulkarni> this talks about the problem that packagers run into if the documentation for a package is defined to be generated using 'git log'
17:47:01 <devkulkarni> an interesting thread to ready
17:47:07 <devkulkarni> s/ready/read/
17:47:34 <devkulkarni> I thought those were the main updates from upstream in the last week
17:47:51 <devkulkarni> anything that caught your attention on the mailing list?
17:48:21 <devkulkarni> I know that there was a magnum midcycle.. would be interested to know how it went and what is the latest on /container endpoint?
17:49:11 <dimtruck> sounds like we're keeping it
17:49:18 <dimtruck> there's still some conversation on that
17:49:27 <devkulkarni> cool
17:49:31 <dimtruck> but the cores are split 50/50
17:49:37 <devkulkarni> I see
17:49:59 <devkulkarni> what were the main take aways?
17:50:09 <dimtruck> it seems like what magnum /may/ become is its own COE on top of swarm, k8s, and mesos
17:50:15 <dimtruck> with the /containers endpoint
17:50:29 <dimtruck> so you'd essentially have a choice of 4 COEs
17:50:36 <dimtruck> a la AWS
17:50:36 <devkulkarni> I see
17:50:38 <devkulkarni> interesting
17:51:04 <dimtruck> adrian will put together notes about the meetup sometime this week and i'll be sure to update the team with them
17:51:06 <devkulkarni> I remember reading about such a proposal from hongbin
17:51:10 <dimtruck> right
17:51:24 <devkulkarni> that would be very helpful dimtruck
17:51:25 <dimtruck> he's the main proponent of it
17:51:38 <devkulkarni> gotcha
17:52:09 <devkulkarni> dimtruck: do you know when we will hear about our submissions?
17:52:15 <dimtruck> no sir
17:52:35 <devkulkarni> hmm.. I have also not come across the date when it will be announced
17:52:47 <devkulkarni> I am thinking it may be sometime in mid-march
17:53:18 <dimtruck> i believe so as well
17:53:40 <dimtruck> i'm sure we'll see an email come out
17:53:52 <devkulkarni> yep..
17:54:07 <devkulkarni> anything else to discuss today? If not, we can end.. its almost top of the hour anyways
17:54:29 <devkulkarni> ok then
17:54:38 <devkulkarni> thanks dimtruck muralia1 vijendar for joining today
17:54:47 <dimtruck> thank you!
17:54:48 <devkulkarni> hopefully see you next week too
17:54:53 <devkulkarni> #endmeeting