14:00:09 <zhaochao> #startmeeting trove
14:00:10 <openstack> Meeting started Wed May  2 14:00:09 2018 UTC and is due to finish in 60 minutes.  The chair is zhaochao. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:13 <openstack> The meeting name has been set to 'trove'
14:00:27 <zhaochao> #topic roll call
14:01:44 <songjian> o/
14:02:04 <zhaochao> hi
14:02:25 <gang> o/
14:03:01 <zhaochao> hi, let's wait for some more minutes for others to show up
14:03:09 <fanzhang> zhaochao o/
14:03:21 <zhaochao> fanzhang: :)
14:04:32 <songjian> hi
14:05:04 <zhaochao> OK, let's start
14:05:20 <zhaochao> #topic Rocky goal updates
14:06:11 <zhaochao> In fact, still nothing new from me in the last week, though I opened another patch in trove-dashboard today
14:06:34 <zhaochao> #link https://review.openstack.org/#/c/565686/
14:07:01 <zhaochao> the issue about unittest blocked the patches of pike branch for a long time
14:07:25 <zhaochao> however I finnaly found it's been fixed in the master, thanks to Samuel
14:08:09 <zhaochao> and the left patches in pike branch of trove-dashboard should now be able to get merged
14:08:14 <zhaochao> #link https://review.openstack.org/#/q/status:open+project:openstack/trove-dashboard+branch:stable/pike
14:08:59 <zhaochao> this one may also need to be rebased, I'll do it later
14:10:17 <fanzhang> Nice job :) Honestly, I just start coding for bp 'adapting file injection deprecation in nova' today.  So far so good, will commit the patches as soon as possible in order to let your guys review sooner.
14:10:39 <zhaochao> fanzhang: great news, thanks :)
14:11:40 <fanzhang> Just finishing adding InjectedFiles object and refactoring get_injected_files, I have not fixed my test trove env, so maybe it'll cost some time.
14:12:58 <fanzhang> That's all from my side. Oh, because of the holiday, never got the chance to review the spec gang brought up earlier and the patches committed by zhaochao , sorry for that, will do this week
14:13:06 <gang> I have updated my spec and fixed some format errors. Please review when you have time.
14:13:30 <zhaochao> building up an dev env for trove may need some more documents to speed up development and easy new contribution
14:13:31 <fanzhang> yeah, that's what I am intended to do.
14:15:06 <zhaochao> gang: though I may still don't have much focusing on Trove, but I started reading the spec today, I think I'll finish the first reading through this week
14:16:28 <gang> zhaochao: thanks.
14:16:35 <songjian1994> Still busy for gnocchi, recently idle time watched some mongodb and redis cluster restart, I think it can be discussed in the spec
14:16:43 <zhaochao> gang: no problem, :)
14:17:58 <gang> Looking forward advice of all you guys.
14:18:19 <gang> songjian: looks good!
14:19:20 <zhaochao> songjian1994: do we have issues about mongodb and redis cluster restarting? I didn't look into that part of code before
14:20:14 <gang> I think it's part of cluster supporting.
14:20:21 <songjian1994> In fact we haven't implemented these two types of cluster restart yet.
14:20:46 <songjian1994> We only have mariadb with roll restart
14:20:58 <gang> yes, rolling_xxx is lack.
14:21:16 <songjian1994> But restart seems to be necessary for resize-volume and resize-flavor
14:22:08 <zhaochao> gang: songjian1994 got that, these could be noted in the comments during review, and then we could discuss more about those topics
14:23:13 <songjian1994> yes,I think so
14:23:18 <gang> zhaochao: OK
14:24:47 <songjian1994> If the reserved data is necessary for the restart action, then we may need some discussion.I will comment it during review
14:27:51 <zhaochao> I think restarting is necessary during resize-volume and resize-flavor
14:28:11 <zhaochao> what's reserved data? songjian1994
14:31:41 <songjian1994> For example, if the user decides to resize after using it for a period of time, then it is necessary for the user to retain the data in the existing database, but it seems that for redis and mongo，this requires some additional commands to operate. The specific content we can trace and discuss in spec
14:31:50 <gang> Currently, rolling_restart_cluster is defined in taskmanager/models.py as a common method, if we want support different type of datastore.Maybe should more disscuss.
14:33:00 <songjian1994> For example, for mongo, we may need to take the initiative to elect the primary, like using rs.stepDown()
14:34:37 <fanzhang> I don't know a lot about resizing, but I think songjian1994 is right. The operation for cluster can be tough in order to maintain the availability of database service
14:34:44 <gang> Different situations for different datastore.
14:34:54 <fanzhang> right
14:35:42 <songjian1994> Yes, I think this may be one of the reasons why only galera's roll restart was implemented.
14:38:01 <zhaochao> I got the point, let's leave these topics for review then we could do more investigation
14:38:11 <fanzhang> agreed.
14:38:35 <songjian1994> agreed
14:39:00 <zhaochao> looks like indeed we have lots of work to do if we want to improve cluster supporting
14:39:50 <gang> support cluster step by step, restart seems a good start compare with resize_volume or resize_flavor.
14:40:47 <zhaochao> yes, agree
14:41:18 <zhaochao> so do we have more for this rocky update topic?
14:41:42 <fanzhang> One topic I want to bring up here is that we may need some energy to focus on helping developers those who ask question in channel trove
14:41:56 <fanzhang> but it's tough I know
14:42:47 <fanzhang> I tried but looks like I cannot help with all these questions which a lot I'm not farmiliar with :(
14:43:23 <fanzhang> so I'm just saying if you guys have a little bit time, please stay in openstack-trove :)
14:44:02 <zhaochao> yes, for a open source community, the users are helping each other is the ideal situation, but we don't have a large user base for trove
14:44:11 <fanzhang> Looks like they're struggling with building image, booting an instance with no guest_info or can't reading guest_info in mongodb
14:45:07 <fanzhang> I've suggested they should report a bug in launchpad, hope they would do that.
14:45:34 <fanzhang> Oh, the summit is coming, who will do the project update?
14:46:45 <fanzhang> zhaochao I'm just curious :)
14:48:09 <zhaochao> last time we talked about this, but it seems none of us will go to the summit, don't know whether maciej will go
14:48:49 <fanzhang> ok, I saw amrith was on the speaker list
14:49:04 <fanzhang> nothing more here.
14:50:56 <zhaochao> O, then maybe kumarmn have talked with Amrith
14:51:36 <zhaochao> OK, I think that's all today
14:51:57 <zhaochao> gang: songjian1994 ?
14:52:06 <gang> no more for me
14:52:13 <songjian1994> no more for me
14:52:45 <zhaochao> OK, thanks, everyone, Goodnight!
14:52:54 <gang> good night
14:53:07 <zhaochao> #endmeeting