15:00:06 <carloss> #startmeeting manila 15:00:06 <opendevmeet> Meeting started Thu Jun 6 15:00:06 2024 UTC and is due to finish in 60 minutes. The chair is carloss. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:06 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:07 <opendevmeet> The meeting name has been set to 'manila' 15:00:24 <carloss> courtesy ping: dviroel vhari gouthamr carthaca msaravan pulluri ashrodri 15:00:32 <jayaanand> hi 15:00:51 <ccokeke[m]> Hello 15:00:58 <carthaca> Hi 15:01:21 <haixin> o/ 15:02:20 <msaravan> Hi 15:03:15 <gouthamr> o/ 15:03:19 <kpdev> hi 15:03:20 <ashrodri> o/ 15:03:29 <vhari> hi 15:03:59 <carloss> o/ hello everyone 15:04:05 <carloss> good quorum, let's get started 15:04:14 <carloss> today's meeting agenda: 15:04:17 <carloss> #link https://wiki.openstack.org/wiki/Manila/Meetings#Next_meeting 15:04:36 <carloss> #topic Announcements 15:04:55 <carloss> first, thank you ashrodri for running the last week's meeting 15:06:16 <carloss> in the past month or so I was away during various meetings due to PTO, holidays, so thanks for holding the fort while I was away :) 15:06:35 <carloss> Schedule and Deadlines 15:06:46 <carloss> #link https://releases.openstack.org/dalmatian/schedule.html (Dalmatian release schedule) 15:07:31 <carloss> next milestone to be looking at would be our spec freeze, which will happen in 3 weeks from now 15:08:11 <carloss> it means that all Manila specs targeted to 2024.2 Dalmatian must be approved by the end of the week R-14 15:08:33 <carloss> we have two specs being worked on and there has been some chatter in both of them, but we'll get to that soon 15:09:11 <carloss> the second and last thing I have for announcements (which is not quite an announcement) is: 15:09:19 <carloss> Possible dates for our mid cycle 15:10:04 <carloss> this was discussed at the last PTG and we talked about meeting halfway through the cycle to check how things we discussed at the PTG are going 15:10:41 <carloss> I am initially thinking about the dates within: Jul 15 - Jul 19 15:11:10 <carloss> because we will be past spec freeze and we will be two weeks from feature proposal freeze 15:11:32 <carloss> and I think we will have a good idea of how things are going in general 15:12:20 <carloss> I am initially considering a single day for the discussions and a 2 hour meeting for it. Do you have any thoughts or suggestions? 15:13:53 <ashrodri> Those dates works well. Is this an extended manila weekly meeting, or a block of time during that week outside of this meeting? 15:14:38 <carloss> I think we can reuse the original manila meeting slot and extend it for another hour... it would be a bit less disruptive for everyone's weeks 15:15:02 <ashrodri> sounds good thanks for clarifying 15:17:28 <carloss> do we have an agreement on date + format then? :) 15:18:59 <carloss> taking silence as a yes 15:19:19 <carloss> but please let me know in case you'd like some changes to it 15:19:29 <carloss> I'll remind us when time comes 15:20:02 <carloss> that's all the announcements I had for today. Do you have an announcement to share with us? 15:21:01 <carloss> moving on... 15:21:11 <carloss> #topic Deferred deletion periodic task issues (kpdev) 15:21:27 <carloss> #link https://etherpad.opendev.org/p/manila-dalmatian-review-focus#L12 (Deferred deletion issues) 15:21:51 <carloss> kpdev, floor is yours 15:23:13 <kpdev> as mentioned on etherpad.. I am not sure what way would be best.. but seems like periodic task are not reliable 15:23:40 <kpdev> so i was suggesting option 2. unless we have definite way to fix 1 15:25:55 <carloss> thank you for bringing this up 15:27:10 <carloss> I think fix 1 would be a bit more complicated as the slowness might be related to a very busy cluster 15:27:48 <gouthamr> maybe we have a limit on the number of shares to delete per iteration 15:28:10 <kpdev> but what is task is not executed on time. 15:28:33 <gouthamr> ^ that's probably because the previous periodic tasks outlasted their interval 15:28:39 <kpdev> i observed sometime even it took 10 minutes.. i.e. delayed by 10 times of designated time. 15:29:18 <kpdev> >maybe we have a limit on the number of shares to delete per iteration 15:29:34 <kpdev> its difficult to decide this limit. as number of shares keep changing.. 15:30:01 <kpdev> i mean max_gigabytes/max_shares might decide that number.. 15:32:27 <carthaca> https://bugs.launchpad.net/oslo.service/+bug/2011587 is part of the problem - the tasks are running sequentially. Could we somehow stop the task instead of letting it run longer? 15:33:00 <carthaca> But I'm in favor of fix 2 anyway - avoid using periodic tasks as much as possible ;) 15:33:25 <gouthamr> +1 15:34:58 <carloss> yep 15:35:11 <gouthamr> this was my initial suggestion on the spec :) 15:35:44 <kpdev> ok, then with updated consideration.. deferred deletion would be immediate deletion with new states i.e. deferred_deleting instead of deleting. and error scenario only considered in periodic task 15:37:26 <gouthamr> makes sense 15:37:57 <kpdev> thanks for consideration. 15:38:24 <carloss> np, thanks for bringing it up 15:38:32 <carloss> that will be a good enhancement 15:39:11 <gouthamr> so iiuc; you'd release quota immediately, and attempt deletion right away.. if the deletion errors out, you'd set the appropriate error state ("error_deleting" or "error_deferred_deleting") based on the prior state 15:39:31 <gouthamr> and a periodic task will reap shares in error_deferred_deleting 15:40:23 <kpdev> yes.. 15:40:41 <kpdev> do i need to release quota immediately.. ? 15:41:16 <gouthamr> if deferred deletion is enabled 15:41:34 <kpdev> now since deletion happen at same place, the quota release and deletion is not much different. 15:42:00 <kpdev> what we achieve is once deleted, end-user need not to look into share. 15:42:10 <kpdev> deferred deleting is hidden from user 15:42:33 <kpdev> so share is deleted from their point of view.. and error_deferred_deleting will try to delete again until deleted 15:43:22 <gouthamr> ack; think we can take this discussion to gerrit.. 15:43:35 <kpdev> ok 15:43:37 <gouthamr> thanks for highlighting this issue kpdev carthaca 15:43:47 <carloss> ++ 15:44:29 <carloss> thanks kpdev carthaca gouthamr 15:44:35 <carloss> #topic Review Focus 15:44:45 <carloss> I'll be brief here :) 15:45:01 <carloss> next thing we have in mind is the spec freeze 15:45:07 <carloss> currently, two specs are being worked on: 15:45:54 <carloss> #link https://review.opendev.org/c/openstack/manila-specs/+/915480 (Add spec for ensure shares) 15:46:01 <carloss> and 15:46:19 <carloss> #link https://review.opendev.org/c/openstack/manila-specs/+/916595 (Add spec lite for resource metadata backend operations) 15:46:37 <carloss> both already got some reviewer feedback 15:47:03 <carloss> and they have been updated, so let's continue with the discussions on the specs. I believe both of them might be close at this point 15:47:21 <carloss> I have managed to convert the ensure shares spec from a lite spec to a regular spec 15:47:52 <carloss> kpdev ashrodri: can I have your eyes on the ensure shares spec as well? 15:48:00 <kpdev> ack 15:48:08 <ashrodri> yes, can take a look 15:49:31 <carloss> tyvm 15:50:04 <carloss> that's all the changes I wanted to highlight for today's meeting. Is there something else you would like to bring up? 15:51:03 <carloss> #topic Bug Triage (vhari) 15:51:16 <carloss> o/ vhari - I think we have some minutes for some bug triage :) 15:51:25 <carloss> #link https://etherpad.openstack.org/p/manila-bug-triage-pad-new (Bug Triage etherpad) 15:51:26 <vhari> ty carloss :) 15:51:50 <vhari> there were 2 new bugs this week .. both have fixes proposed 15:52:00 <vhari> thanks to folks who got a quick start on these 15:52:15 <vhari> so there are no bugs to triage 15:52:36 <vhari> we could wrap the session and pick up next week 15:52:53 <vhari> floor is back to you carloss :) 15:53:06 <carloss> ah, thank you very much vhari 15:53:11 <carloss> #topic Open Discussion 15:55:27 * carloss hear crickets 15:55:32 <carloss> let's wrap up 15:55:42 <carloss> thank you for participating, let's get back to #openstack-manila :) 15:55:45 <carloss> have a great day 15:55:48 <carloss> #endmeeting