15:00:06 #startmeeting manila 15:00:07 Meeting started Thu Aug 15 15:00:06 2019 UTC and is due to finish in 60 minutes. The chair is tbarron. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:10 The meeting name has been set to 'manila' 15:00:14 hi :) 15:00:14 o/ 15:00:15 o/ 15:00:20 hi :) 15:00:28 o/ 15:00:37 courtesy ping: gouthamr xyang toabctl bswartz ganso erlon tpsilva vkmc amito jgrosso dviroel lseki carloss 15:00:46 \o) 15:00:52 (o/ 15:00:55 vkmc: \o 15:00:58 vkmc: o/ 15:01:03 lol 15:01:15 o/ 15:01:20 it's cold in buenos aires I guess 15:01:31 yeah 15:01:38 "cold", our winter for you must be a joke 15:01:43 but I'm seriously suffering 15:01:57 I see you doing jumping jacks 15:01:59 53F 15:02:05 haha yes 15:02:14 keeps you warm and makes you fit, smart 15:02:38 Hi all! 15:02:55 Agenda: https://wiki.openstack.org/wiki/Manila/Meetings#Next_meeting 15:03:15 #topic Announcements 15:03:39 First, just our weekly schedule reminder, 15:03:57 This wee is Feature Proposal Freeze 15:04:01 This week 15:04:15 I think we're OK on that front -- let me know if 15:04:31 you have something cooking that we need special consideration for -- 15:04:39 but I need to clarify something. 15:05:00 Last week I said that this meant that doc for the feature was supposed to be ready. 15:05:03 But 15:05:30 it turns out that I had inadvertently copied that requirement into our schedule this release 15:05:37 We've never had it before 15:05:42 So mea culpa. 15:05:57 And we won't require that at this point. 15:06:34 Actual Feature Freeze is in threee weeks, 5 September. 15:06:53 "No featureful patch should be landed after this point." 15:07:00 tbarron: 5 september? 15:07:08 quoting from the release schedule 15:07:17 * tbarron looks again 15:07:22 =) 15:07:28 "Sep 09 - Sep 13" 15:07:29 Sep 09 - Sep 13 15:07:40 yeah 15:07:47 #link https://releases.openstack.org/train/schedule.html 15:07:48 i'm confused today, as usual 15:07:50 thanks 15:08:14 so that's four weeks, right? 15:08:31 yes 15:09:34 The only other announcement I have is that I was able to move the manila-csi talk in shanghai around 15:09:46 so that we don't have both manila talks at the same time 15:09:47 \o/ 15:10:02 tbarron++ 15:10:03 I'm looking forward to attending the talk the NetApp folks are doing 15:10:05 thanks tbarron 15:10:14 thanks tbarron 15:10:18 Any other announcments? 15:10:51 #topic Python 3 for Third Party CI 15:11:02 dviroel: I see you have an update 15:11:23 #link https://wiki.openstack.org/w/index.php?title=Manila/TrainCycle#Python3_Testing 15:12:19 tbarron: yes, not too much to add, we just follow the tips from cinder folks, and everything is running ok 15:12:37 So NetApp is running with 3.7, awesome 15:12:57 And there's a section there where vendors can report success. 15:13:28 Conversely, we'll move this section in the wiki to a quite public location at the end of Train or so :) 15:13:49 And list CI that is only testing with python2 still 15:13:55 Peer pressure :) 15:14:21 Dell-EMC is working on this stuff and thanks carloss for helping out 15:14:48 I'm not sure what's going on with https://bugs.launchpad.net/manila/+bug/1833160 15:14:49 Launchpad bug 1833160 in Manila "VNX: driver cannot start under py37 due to wrap_socket() got an unexpected keyword argument '_context'" [Undecided,New] 15:15:14 tbarron: yw 15:16:39 #topic Reviews 15:16:39 I saw his last comment, will answer soon 15:16:52 carloss: thanks 15:17:06 Last week I said I'd make a review focus etherpad and dropped the ball. 15:17:17 So I just started one :) 15:17:29 #link https://etherpad.openstack.org/p/manila-train-review-focus 15:17:36 Need to populate it 15:17:57 But let's distribute a bit of the work, the point of doing the etherpad rather 15:18:22 than just keeping track on our wiki is that people are more comfortable 15:18:32 just doodling into the etherpad 15:18:40 If you need a review, just list it there 15:19:12 I and/or others can do some formattting to make it look like our old review focus etherpads 15:19:46 We can update with people already reviewing and see gaps easier. 15:20:00 ack 15:20:15 Anything else on this ? 15:20:33 #topic bugs 15:20:41 jgrosso: what do you have for us today? 15:20:55 tbarron one bug 15:20:58 https://bugs.launchpad.net/manila/+bug/1631314 15:20:59 Launchpad bug 1631314 in Manila "Tempest tests "test_promote_out_of_sync_share_replica" and "test_resync_share_replica" are concurrency-prone" [Medium,Confirmed] 15:21:36 nothing big today :) 15:21:48 lowly ole temptest testsd 15:23:03 jgrosso: you asked whether we have a milestone when the issue will be addressed and 15:23:07 hmmm, one thing we can do, rather than chop off the tests is to retry the status 15:23:52 no chopping allowed gouthamr 15:23:54 dviroel: are you folks touching any part of the share replication tempest tests? ^ 15:24:08 the root cause analysis is https://bugs.launchpad.net/manila/+bug/1631314/comments/4 so far as I can tell 15:24:09 Launchpad bug 1631314 in Manila "Tempest tests "test_promote_out_of_sync_share_replica" and "test_resync_share_replica" are concurrency-prone" [Medium,Confirmed] 15:24:21 gouthamr: yes, at this moment. 15:24:28 I meant to point to comment #4 15:25:41 jgrosso: i'm with you :) hate to lose coverage because of something that shouldn't happen in a real cloud 15:26:05 gouthamr: didn't face that issue because we were not using too much threads 15:26:33 gouthamr agreed :) 15:27:11 dviroel: it's not something you can control (besides the replication update interval) 15:28:01 dviroel: it's been a recurring issue on the dummy driver job, and i've occasionally seen it with netapp and zfsonlinux jobs running those two tests.. 15:29:07 gouthamr: oh, now I saw the problem. 15:29:12 a maintenance thread in the share manager interferes with the test logic, so - the test should be more resilient to this imo 15:29:54 gouthamr: but how? 15:30:08 ^ given that the test is trying to set a perfectly healthy replica to unhealthy so it can test the API 15:30:58 it won't ever see it go "out of sync" because it really isn't, so what should it do? just assume that its reset state worked? 15:31:57 tbarron: nope, reset state again, perhaps before issuing the next API call? although that isn't going to be enough if the replication update interval is too low (i think the gate sets it to 60 seconds) 15:32:27 the real default is 300 seconds 15:32:29 gouthamr: oh I see 15:33:08 do we need to have the update interval that low for other tests? 15:33:28 "low" in quotes b/c normally I wouldn't call it that 15:33:51 yes, because the generic build timeout is 180 seconds 15:34:06 you wouldn't want to wait too long to create replicas of empty shares 15:34:08 https://github.com/openstack/manila/blob/8d03a2a46e233b3628ffe59af67c1feb789e1669/contrib/ci/pre_test_hook.sh#L103 15:34:35 #link https://github.com/openstack/manila/blob/8d03a2a46e233b3628ffe59af67c1feb789e1669/contrib/ci/pre_test_hook.sh#L162 15:35:02 so your suggestion is to have the test make several attempts to reset state since we're not testing resettign state, that's 15:35:08 so, the replica update interval is 10 seconds in the gate by default, and 60 seconds for the ZFSOnLinux driver 15:35:09 just a means to testing promotion 15:35:28 tbarron: yes, and "resync" 15:37:00 dviroel: do you know what value of "replica_state_update_interval" the NetApp CI uses, and if it's reliable? 15:37:34 gouthamr: maybe you can update the bug with that suggestion? it seems less work than your original idea in comment #2 to 15:37:49 tbarron: sure can 15:37:49 http://13.56.159.105/logs/43/671043/9/upstream-check/manila-cDOT-no-ss/c972cbe/logs/local.conf.txt.gz 15:37:51 introduce an api to disable/enable the periodic update 15:38:16 that takes us back to jgrosso's original question 15:38:23 dviroel: thanks, so you're not overriding the default 10 sec 15:38:26 jgrosso: no, we don't have a milestone target 15:38:28 yeah, 10 15:38:34 dviroel: that the pre-test hook sets up for all drivers 15:38:49 jgrosso: need volunteer to sign up, we can't just assign all hard problems 15:38:51 to gouthamr 15:39:06 why not ? 15:39:08 :) 15:39:09 or easy ones :) 15:39:21 tbarron: I can help with this one 15:39:26 dviroel: awesome 15:39:47 thanks dviroel 15:39:49 dviroel: ty! 15:39:59 that is all I had for today 15:40:05 jgrosso: dviroel: why don't you guys get together to figure miilestone, etc. 15:40:17 it doesn't have to be immediate, but great to get an ownere 15:40:19 owner 15:40:32 tbarron ack 15:40:41 thanks jgrosso 15:40:48 #topic open discussion 15:40:49 tbarron: ack 15:40:52 tbarron: welcome 15:41:10 I had something to check on 15:41:25 dviroel/netapp folks: what's the status on: 15:41:26 #link https://bugs.launchpad.net/manila/+bug/1699856 15:41:28 Launchpad bug 1699856 in Manila "Tempest tests missing adding Security service to share-network" [Medium,Triaged] - Assigned to Douglas Viroel (dviroel) 15:42:19 gouthamr: we are not focusing on it right now, but shouldn't not be that hard. 15:43:18 dviroel: ack, yep - please take it up whenever possible; since you guys have an AD already in your CI, this should be easy to get tested on your end 15:44:01 dviroel: i can help sync up with other CIFS vendors to do a similar test when you have a fix 15:44:16 dviroel: and buy you beer in Shanghai 15:44:31 just because we're friends, i'm not bribing you 15:44:40 gouthamr: ok, I'll take a look on that after replication 15:44:44 beer++ 15:45:13 dviroel: nice, thanks! 15:45:26 gouthamr: ty 15:45:52 Anything else today? 15:46:36 OK, see y'all in #openstack-manila 15:46:41 Thanks everyone!! 15:46:45 #endmeeting