14:00:04 #startmeeting cinder 14:00:04 Meeting started Wed May 3 14:00:04 2023 UTC and is due to finish in 60 minutes. The chair is whoami-rajat. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:05 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:05 The meeting name has been set to 'cinder' 14:00:06 #topic roll call 14:00:14 o/ 14:00:20 o/ 14:00:21 o/ 14:00:26 o/ 14:00:27 o/ 14:00:29 o/ 14:00:33 0/ 14:00:37 o/ 14:00:40 o/ 14:00:43 o/ 14:01:13 o/ 14:01:19 #link https://etherpad.opendev.org/p/cinder-bobcat-meetings 14:01:53 hi 14:01:56 \o 14:03:02 hello 14:03:23 o/ 14:04:15 always happy to see good number of people joining 14:04:20 let's get started 14:04:24 #topic announcements 14:04:34 Cinder 2023.2 Bobcat Midcycle-1 14:05:04 the first midcycle of 2023.2 cycle will be held on 02nd June 14:05:24 on wednesday between 1400-1600 UTC (our usual meeting time + 1 hour extra) 14:05:37 hello 14:06:03 I've prepared an etherpad to gather the topics, we still have a month but good to add them early since we only have 2 hours 14:06:05 #link https://etherpad.opendev.org/p/cinder-bobcat-midcycles 14:06:36 June 2nd is a Friday ... 14:06:58 wow, it should be 31st may ... 14:06:59 by 02 June, do you mean 31 May ? 14:07:02 * whoami-rajat checking again 14:07:13 :-) Cool. 14:07:51 it's correct on the release schedule but somehow in my personal notes the date is wrong 14:07:57 thanks for correcting it jungleboyj and rosmaita 14:08:01 it's actually 31st May 14:08:03 #link https://releases.openstack.org/bobcat/schedule.html#b-cinder-mid-cycle-ptg-1 14:08:36 so reiterating, we will be having our first midcycle of bobcat on 31st May (wednesday) 1400-1600 UTC 14:08:47 ++ 14:08:57 please add your topics to the etherpad 14:09:52 next, Cinderlib 2023.1 (cycle trailing) release 14:09:58 #link https://lists.openstack.org/pipermail/openstack-discuss/2023-May/033536.html 14:10:09 there was an email from release team regarding the cycle trailing release 14:10:18 and we've the cinderlib project which is cycle trailing 14:11:01 i checked the open patches in cinderlib master and nothing seemed urgent or in a merge-able state so i proposed the release early 14:11:07 #link https://review.opendev.org/c/openstack/releases/+/882132 14:12:13 if anyone has a urgent/required change in cinderlib that they want to include in 2023.1 release, they can propose it and let me know 14:12:28 but if there are no concerns, we will go ahead with releasing cinderlib 2023.1 (Antelope) 14:13:45 next, Stable releases 14:14:16 we recently released all active stable branches i.e. 2023.1, Zed, Yoga and Xena (final) 14:14:28 we also moved Xena to EM so there will be no more releases of xena 14:15:05 for the list of projects that are released with the recent stable releases, please go through the etherpad on L#101 14:16:07 thanks to jbernard for carrying out all the releases! 14:16:36 \o/ 14:17:00 Yeah. 14:17:14 == 14:17:15 ++ 14:17:17 ++ 14:19:19 that's all the announcements i had, anyone has anything else that the team should be aware about? 14:19:36 rocky and stein were tagged EOL about an hour ago 14:19:40 #link https://review.opendev.org/q/topic:cinder-eol 14:20:22 oh yes, forgot to provide an update on that 14:20:24 thanks rosmaita 14:21:04 so rocky and stein branches are not active anymore for cinder so we can't backport/propose patches to those branches now 14:23:05 i guess that's all for announcements, let's move to topics 14:23:09 #topic Cinder-Backup very slow / inefficient when using chunked drivers, e.g. S3 14:23:39 crohmann doesn't seem to be around ... 14:24:51 * whoami-rajat going through the topic 14:27:37 ok, so there are multiple things here 14:28:23 1. chunked backup driver is slow during backup of large volumes as it processes each chunk 14:28:27 #link https://bugs.launchpad.net/cinder/+bug/1918119 14:28:52 there was an attempt to perform parallel read/write of the chunks 14:29:01 #link https://review.opendev.org/c/openstack/cinder/+/779233 14:29:33 Gorka has a comment on it saying we already have config options to do the same thing the patch is trying to do (though i haven't gone through the implementation) 14:30:05 backup_native_threads_pool_size 14:30:08 backup_max_operations 14:30:24 the existing options probably parallelize across multiple backups instead of one backup? 14:30:51 so not really the same thing 14:31:04 that's a good question, I'm not sure but i think you're right 14:31:18 backup_max_operations is for the number of worker processes 14:31:33 backup_native_threads_pool_size is for native thread in a particular worker 14:32:06 but i guess they take different backup operations and don't do parallel read/write of a single backup as eharney said 14:32:08 In theory I was going to look into the performance of backup/restore, but I got sidetracked with Keystone. 14:33:52 Also, people seem to assume that backup is infinitely parallel, but it's not true, because we compress a whole segment. Even if your network performance is infinite, and the back-end has millions of drives, you have to consume blocks as they are compressed. 14:34:02 looks like backup discussions are more frequent than ever, but maybe it's from the same author 14:35:18 yeah, backup is a combination of steps so anything during the whole operation can become a bottleneck 14:35:55 s/anything/any step 14:36:45 anyway, not an expert on backups 14:37:01 the second point in this was regarding NFS backup driver 14:37:13 #link https://review.opendev.org/c/openstack/cinder/+/611079 14:37:39 maybe fix the merge conflict first... 14:37:52 i haven't worked on this one in a while 14:38:03 it fixed an issue where the c-bak service would be blocked by i/o operations iirc 14:39:04 there is a comment on it by Christian today, https://review.opendev.org/c/openstack/cinder/+/611079/5#message-3fc554786126198652ed152f52de89d5237d5368 14:41:28 anyway, that's all i could discuss about this from the agenda 14:41:49 anything else on this topic? 14:43:11 ok, let's move to open discussion 14:43:15 #topic open discussion 14:43:30 Just another call to get some reviews done by cores 14:43:34 we've 3 reviews in review request 14:43:56 Requesting some feedback on below patch (HPE 3PAR). Thanks. 14:43:58 4, i'm about to add another one 14:44:04 https://review.opendev.org/c/openstack/cinder/+/879836 14:45:19 Requesting some feedback on the ceph backup restore fix: https://review.opendev.org/c/openstack/cinder/+/880965 . Thanks! 14:45:34 The Pure patch has been hanging around for 4 weeks. I'd really like to get this merged 14:46:12 test-requirements bump here: https://review.opendev.org/c/openstack/cinder/+/879961 - have been trying to bump our mypy version up since March 14:49:38 O 9 14:56:29 started reviewing the pure patch, lost track of time 14:56:44 I think that's all for today 14:56:55 thanks everyone for joining 14:56:58 #endmeeting