15:06:02 <enriquetaso> #startmeeting cinder_bs 15:06:02 <opendevmeet> Meeting started Wed Jul 7 15:06:02 2021 UTC and is due to finish in 60 minutes. The chair is enriquetaso. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:06:02 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:06:02 <opendevmeet> The meeting name has been set to 'cinder_bs' 15:06:10 <enriquetaso> hello! 15:06:21 <rosmaita> o/ 15:06:29 <enriquetaso> Hey Brian, we have 5 new bugs reported this week. 15:06:42 <enriquetaso> #topic LVM calls - CI failure status 15:06:47 <enriquetaso> As Brian wrote early in the cinder meeting, Wallaby backport https://review.opendev.org/c/openstack/cinder/+/798671 is waiting for https://review.opendev.org/c/openstack/cinder/+/798696/ to be merged. 15:06:52 <enriquetaso> Please core reviews if you have some time please check those 2 ^ 15:06:54 <enriquetaso> :) 15:07:14 <enriquetaso> #topic bug_1: Cinder matrix - clarification of snapshot attachment feature 15:07:21 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1934168 15:07:25 <enriquetaso> Thanks hemna for reporting this. 15:07:43 <enriquetaso> The question is if this should be pitched as an optimization rather than missing functionality in the driver and/or backend. 15:08:52 <enriquetaso> Summary: In most arrays, snapshots are not attachable entities. The default behavior is that a volume is created from the snapshot and that volume is then attached. 15:09:04 <enriquetaso> Option 1: remove this feature from the matrix. 15:09:04 <enriquetaso> Option 2. Call out in 'Notes' that this is an optimization in backup support for drivers that can attach snapshots and is not a gap in functionality for those who don't. 15:09:37 <rosmaita> i seem to remember discussing this -- was it at last week's meeting? 15:09:48 <rosmaita> cinder meeting, not bug squad, i mean 15:09:52 <enriquetaso> oh, sorry about that 15:09:59 <walshh_> Yes, I brought up it 15:10:03 <geguileo> yes, I think we agreed on the note, right? 15:10:36 <enriquetaso> great, sorry! I'll update the launchpad but all this then 15:10:57 <walshh_> I am happy to make that change if there are no objections. Thank you 15:11:06 <enriquetaso> :) cool 15:11:10 <rosmaita> even better! 15:11:15 <geguileo> walshh_: thanks! 15:11:22 <rosmaita> yes, please put up a patch and we can discuss there 15:11:35 <walshh_> will do 15:11:42 <enriquetaso> thanks! 15:11:52 <enriquetaso> #topic bug_2: Cinder can not use system scope due to endpoint 15:11:57 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1934480 15:12:03 <enriquetaso> Summary: Cinder cannot use system scope since its endpoint has project_id and will get a empty url in catalog in system scope. 15:12:11 <enriquetaso> I'm not completely sure if this is a feature or a bug. 15:12:12 <enriquetaso> Does cinder use the other scopes already (project-scope, domain-scope)? 15:12:12 <enriquetaso> I've marked it as Medium importance but I should move it to the wishlist if it's a feature. 15:12:18 <eharney> as written, this seems to be a bug about how keystone endpoints were deployed and not a cinder bug 15:12:37 <rosmaita> i think also it is a duplicate 15:12:52 <geguileo> didn't jokke report that one too? 15:12:54 <rosmaita> pretty sure lbragstad filed a bug around this 15:13:11 <rosmaita> i think jokke's was a bit different 15:13:15 <geguileo> ah, ok 15:13:51 <enriquetaso> so, looks like a keystone bug then 15:14:04 <rosmaita> well, it's also a cinder problem 15:14:05 <eharney> no 15:14:26 <rosmaita> a system-scoped token is not related to any project 15:14:37 <enriquetaso> make sense 15:14:44 <rosmaita> but we assume that every user , including an admin, is in a project 15:14:53 <rosmaita> and that project-id must be in the URL 15:15:11 <rosmaita> (and we check somewhere that the URL project and token project match) 15:15:18 <eharney> so do we want this to be an RFE for "support system scope" in general? 15:15:40 <rosmaita> something like that 15:16:19 <rosmaita> there's no good status in launchpad for that 15:16:21 <enriquetaso> but, shouldn't mark it as duplicate of lbragstad bug ? and continue the work on that one? 15:16:44 <rosmaita> yes, if there is a bug 15:16:57 <rosmaita> not sure if there is though, i may be thinking of a patch 15:17:49 <rosmaita> maybe this one: https://bugs.launchpad.net/cinder/+bug/1926336 15:18:50 <opendevreview> jinyuanliu proposed openstack/cinder master: Clean up extra spaces https://review.opendev.org/c/openstack/cinder/+/799848 15:18:51 <hemna> rosmaita fwiw, I haven't had time to look at the ceph-iscsi CI job failures yet 15:18:54 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1926336 15:20:07 <enriquetaso> oh, well, then I can make 1934480 and RFE for "support system scope" as Eric mentioned? 15:21:25 <rosmaita> hemna: that's ok, i did get the github mirroring working though 15:21:35 <hemna> ok perfect thanks 15:23:00 <enriquetaso> #action (enriquetaso): make 1934480 RFE for "support system scope" :) 15:23:07 <enriquetaso> OK, last one 15:23:09 <rosmaita> enriquetaso: sorry, i got pulled away for a minute 15:23:46 <enriquetaso> sure, don't worry, i think it was an obvious question doh :P 15:24:08 <enriquetaso> #bug_3: Solidfire with OSprofiller and ElementOs 15:24:14 <enriquetaso> There are two bugs that sort of confused me. Maybe the reporter already discussed it in the cinder meeting and I forgot. :s 15:24:29 <enriquetaso> On the one hand, we have (1)https://bugs.launchpad.net/cinder/+bug/1934459 "Enabling OSprofiler in a Openstack cluster with Element OS , runs into an infinite loop recursion". 15:24:29 <enriquetaso> On the other hand (2) https://bugs.launchpad.net/cinder/+bug/1934435 I left a question asking if it's a feature or a bug report and asking for more info. 15:24:48 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1934459 15:24:54 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1934435 15:25:40 <enriquetaso> The first one looks like solidfire config isn't correct 15:25:40 <eharney> bug 1934435 looks like the backend is throwing errors, someone from solidfire will need to look at that one 15:26:32 <eharney> for 1934459, the solidfire driver should probably just stop overriding __getattr__, dunno why it does that 15:26:43 <eharney> unsurprising that it causes weird issues 15:28:20 <enriquetaso> eharney++ 15:28:35 <enriquetaso> thanks Eric 15:28:50 <enriquetaso> OK, that's all I have for today's meeting 15:28:57 <enriquetaso> #topic open discussion 15:29:15 <enriquetaso> do we have any bug to discuss in this 2 min left 15:30:11 <enriquetaso> OK! Thanks everyone! 15:30:24 <rosmaita> thanks, sofia! 15:30:27 <enriquetaso> sorry for the misunderstanding in the first bugs 15:30:32 <enriquetaso> my head was a little clueless 15:30:40 <enriquetaso> #endmeeting