18:00:20 #startmeeting trove 18:00:21 Meeting started Wed May 25 18:00:20 2016 UTC and is due to finish in 60 minutes. The chair is amrith. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:23 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:00:25 The meeting name has been set to 'trove' 18:00:43 o/ 18:00:52 hello peterstac 18:00:55 o/ 18:01:03 hello tellesnobrega 18:01:48 howdy yall 18:01:56 o/ 18:01:59 hello cp16net 18:02:05 howdy cp16net 18:02:12 hello SlickNik 18:02:27 * cp16net tips hat 18:02:30 hi hi 18:02:43 hi amrith 18:02:53 hi amrith 18:03:07 hi stewie925 18:03:12 o/ 18:03:15 o/ 18:03:22 hello dougshelley66 twm2016 18:03:30 o/ 18:03:54 hello vgnbkr 18:03:58 let's get started 18:04:05 #agenda https://wiki.openstack.org/wiki/Meetings/TroveMeeting 18:04:16 #topic Action items from last week's meeting 18:04:30 We had one last week, to review https://review.openstack.org/#/c/298994/5 18:04:39 unfortunately, it remains undone 18:04:53 #action [all] review https://review.openstack.org/#/c/298994/5 18:05:04 nothing else from last week 18:05:43 ./ 18:06:01 anyone have anything else to add re: last weeks action items 18:06:07 amrith, 298994 merged 18:06:37 good morning ... time to wake up 18:06:41 what day is it today? 18:06:47 sorry, not sure what I was looking at 18:07:04 so it merged, great ... 18:07:13 anything else to add re: action items 18:07:27 #topic Trove pulse update 18:07:28 with the /5 on the end it takes you to a 'not current' page (something I don't really like about gerrit) 18:07:44 aha, that's what it was 18:07:51 ok, thanks for clearing that up 18:07:54 #link http://bit.ly/1VQyg00 18:08:10 #link https://gist.github.com/amrith/7d0f3122b0132d30f6a4fa912c1e9a31 18:08:30 so, the momentum we had on reviews seemed to go away last week 18:08:45 the queue dropped but that's largely because there were not many new changes 18:08:53 (fewer) 18:09:10 we have a number of open pages that could benefit from reviews 18:09:22 One area to focus is specs 18:09:24 #link https://review.openstack.org/#/q/project:openstack/trove-specs+NOT+label:Workflow%253C%253D-1++NOT+label:Verified%253C%253D-1%252Cjenkins+NOT+label:Code-Review%253C%253D-2+status:open++%28project:openstack/trove+OR+project:openstack/trove-integration+OR+project:openstack/trove-specs+OR+project:openstack/python-troveclient+OR+project:openstack/trove-dashboard%29 18:10:11 anyone have anything to add on the subject of momentum, reviews, 18:10:43 Ok, moving along 18:10:45 #topic https://review.openstack.org/#/q/project:openstack/trove-specs+NOT+label:Workflow%253C%253D-1++NOT+label:Verified%253C%253D-1%252Cjenkins+NOT+label:Code-Review%253C%253D-2+status:open++%28project:openstack/trove+OR+project:openstack/trove-integration+OR+project:openstack/trove-specs+OR+project:openstack/python-troveclient+OR+project:openstack/trove-dashboard%29 18:10:51 #topic Announcements 18:11:00 * amrith damn cut and paste gone wild 18:11:26 reminder: we have a spec proposal deadline in a couple of weeks 18:11:28 R-16 18:11:32 June 13-17 18:11:47 does anyone else have specs that you'll be submitting (not yet up for review) 18:12:52 OK, next announcement 18:13:04 Please take a look at the email about mentoring that hit the ML earlier today 18:13:10 http://markmail.org/message/nec7ez6ocgjcjxxf 18:13:50 That's all I had for announcements 18:13:56 anyone else have anything? 18:15:15 ok, moving along 18:15:22 there's nothing else on the agenda 18:15:25 so let's move to 18:15:29 #topic open discussion 18:15:50 stewie925, you had a question earlier 18:15:58 was it about rebase? 18:16:19 hi amrith - yes it is about rebase - https://review.openstack.org/#/c/312611/5 18:16:43 patch 4 was a rebase showing dependencies to another gerrit 18:17:07 but when I submitted patch 5, it seems the dependency is gone 18:17:59 stewie925, what do you mean by 'dependency to another gerrit'? 18:18:29 I'm not seeing any dependencies called out in the commit message 18:19:07 if you look at patch 4 https://review.openstack.org/#/c/312611/4 it shows a parent id that points to a code fix in gerrit #308707 18:19:41 I think that just lists related changes 18:19:54 wait a second, you said parent id 18:20:14 yes parent id 18:20:26 pmalik helped me with it yesterday 18:20:54 pmalik, yt? 18:21:29 yes 18:21:38 hi pmalik, I have made the mocking changes you suggested for the guest_id 18:21:48 I think you somehow detached yourself from the parent patch set. :/ 18:22:08 pmalik, is there a need for this change to be dependent on the other? 18:22:10 We can look at it offline. 18:22:18 Yes. It need it. 18:22:27 He needs to mock a config property. 18:22:29 yeah good idea, pmalik, thank you appreciate it 18:22:40 ok, in that case the two of you can chat about it in #openstack-trove 18:22:46 thank you amrith 18:22:46 anything else of open discussion 18:22:48 Looks like there was a rebase that (maybe) got rid of the dependent change? 18:22:48 ? 18:23:08 Not from me. 18:23:17 cp16net, anything? 18:23:21 peterstac? 18:23:35 vgnbkr, tellesnobrega, dougshelley66, pmalik anything to add? 18:23:39 i have the cinder snapshot spec up for review, would be good to get some eyes there, as well as a API change patch, so we can agree on the structure, if you guys can take a look 18:23:54 tellesnobrega, I have it on my list for today 18:23:55 #link https://review.openstack.org/#/c/320512/ 18:24:03 amrith, awesome 18:24:13 seeing some problems from the mysql systemd vendor packages; anyone else been down this path? 18:24:39 i could bring up when i thought about earlier 18:24:40 what problem(s) pmackinn 18:25:34 also, what vendor? 18:25:43 mysqld seems to start up again in the prepare before the new conf has been set 18:25:43 I know that I've had no trouble with the percona packages 18:25:57 thus writing a new db to /var/lib/mysql 18:26:21 but we've already copied the original to /mnt/volume/data 18:26:51 don't you mount /mnt/volume/data to /var/lib/mysql 18:27:24 that indeed happens 18:27:28 we copy it first tho 18:27:47 the database should be stopped at that point 18:27:51 but before that mysqld runs again with datadir still at /var/lib/mysql 18:28:00 but if systemd magically starts it again 18:28:05 hilarity ensues 18:28:21 so is restart set in systemd? 18:28:34 might need to set a configuration to not auto restart? 18:28:35 i.e. the watchman in systemd potentially treating it as a down mysql? 18:28:43 * amrith is guessing 18:28:47 sounds like it 18:29:05 yes but if we've stopped via systemctl it shouldn't restart 18:29:16 on its own 18:29:36 and yet it is? 18:29:49 trove is restarting it 18:29:55 its in the logs 18:30:15 percona btw 18:30:31 anywho, will keep digging 18:30:33 this is before we mount the volume to /var/lib/mysql? 18:30:38 that shouldnt be the case 18:31:11 after but before the tmp conf has come back in with datadir=/var/lib/mysql/data 18:31:31 oh... 18:31:42 i c now 18:33:13 sounds bad, very bad 18:33:32 makes me mad, very mad 18:33:38 https://github.com/openstack/trove/blob/master/trove/guestagent/datastore/mysql_common/manager.py#L229 18:33:49 pmackinn, in that case it is a feature 18:34:25 * pmackinn can't wait for SUPERCONDUCTOR 18:34:41 cp16net, the mount happens earlier https://github.com/openstack/trove/blob/master/trove/guestagent/datastore/mysql_common/manager.py#L218 18:34:56 and the datadir is reset https://github.com/openstack/trove/blob/master/trove/guestagent/datastore/mysql_common/manager.py#L228 18:35:00 right but thats after the mount happened it restarts 18:35:06 without the updates configs 18:35:37 one way we could "fix" this is in the image set the mysql dir to /var/lib/mysql/data 18:36:01 but wait, this works ... 18:36:10 we set datadir 18:36:13 that's the important part 18:36:18 before restarting 18:36:23 or I'm missing something 18:36:59 i can assure you that it restarts with the old config 18:37:00 oh yeah i see what you mean 18:37:17 old == original 18:37:26 like coke 18:38:10 baking /var/lib/mysql/data into the image CAN work but it leaves a mark 18:38:18 yeah 18:38:23 /var/lib/mysql/data/data 18:38:51 app.set_data_dir(mount_point + '/data') 18:39:08 shoudl set the dir right before we start it again 18:39:27 any logs aroudn that? 18:39:41 might be helpful 18:39:47 you and your logs...jas 18:40:13 anyways its something to look into 18:40:39 meh, that instance is gone...i'll hit the channel up later 18:40:42 * amrith TIMECHECK 18:40:44 k 18:40:52 so is this something we can pick up in #openstack-trove 18:41:00 right after we adjourn here? 18:41:11 thx all for the input! 18:41:19 does anyone else have any other things for open discussion ... 18:41:32 #action pmalik and stewie925 to look into the rebase issue 18:41:33 i could bring up what this lead me back to thinking about 18:41:50 #action amrith pmackinn cp16net to see what this mysql issue is 18:41:54 anything else? 18:42:19 so in pxc i had some weird issue that i hacked around 18:42:30 https://github.com/openstack/trove/blob/master/trove/guestagent/datastore/experimental/pxc/service.py#L70 18:42:30 #action review https://review.openstack.org/#/c/320512/ ( tellesnobrega's spec on cinder snapshots ) 18:43:00 go ahead cp16net we have about 18m here before the bot kicks us out 18:43:04 amrith, this one is not the spec, this is the API structure change 18:43:06 we're done for the meeting it appears. 18:43:13 since we have never seen this issue with single instances and only in pxc cluster 18:43:24 amrith, https://review.openstack.org/#/c/306620/ this one is the spec 18:43:44 tellesnobrega, thanks for clarification 18:43:45 yes 18:43:50 np 18:44:23 my theory is maybe we cant set the admin password until the cluster has quarum 18:44:45 and the sync state is up. 18:45:11 wondered if working on maria clustering or others if anyone saw this same issue? 18:45:55 cp16net, haven't seen that but sounds plausible 18:46:45 yeah because writes are not allowed until quarum is reached i believe 18:47:06 it was an idea i had after talking with pmackinn about his issue 18:47:23 i thought they might be related but it doesnt seem like it now 18:48:04 * pmackinn has pxc working with systemd btw, ugly but working 18:48:18 i just wanted to throw that idea out there and see if it sticks 18:48:29 thats all i had 18:48:38 ok 18:48:54 let's take our stuff and mosey over to #openstack-trove and continue there 18:48:58 last call ... 18:49:05 anybody have anything for the meeting? 18:49:14 objections to #endmeeting speak now 18:49:47 #endmeeting