14:01:11 <whoami-rajat_> #startmeeting cinder 14:01:11 <opendevmeet> Meeting started Wed Jul 17 14:01:11 2024 UTC and is due to finish in 60 minutes. The chair is whoami-rajat_. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:11 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:11 <opendevmeet> The meeting name has been set to 'cinder' 14:01:17 <whoami-rajat_> #topic roll call 14:01:24 <Sai> o/ 14:01:57 <rosmaita> o/ 14:01:58 <whoami-rajat_> Jon has some conflicts so I will be chairing today's meeting 14:02:02 <eharney> hi 14:02:21 <kpdev> hi 14:02:42 <msaravan> hi 14:03:18 <whoami-rajat_> #link https://etherpad.opendev.org/p/cinder-dalmatian-meetings 14:03:39 <jungleboyj> o/ 14:04:47 <tosky> hi 14:04:51 <ccokeke[m]> hello 14:05:16 <akawai> o/ 14:05:20 <whoami-rajat_> hello everyone 14:05:32 <whoami-rajat_> let's get started 14:05:45 <whoami-rajat_> I haven't prepared much for the announcements 14:05:51 <whoami-rajat_> just few upcoming deadlines 14:05:58 <whoami-rajat_> 1. New feature status checkpoint (R-9, Aug-02) 14:06:04 <whoami-rajat_> #link https://releases.openstack.org/dalmatian/schedule.html#d-cinder-feature-checkpoint 14:06:27 <whoami-rajat_> if you are planning to implement a feature, this is a reminder to start preparing for it 14:06:54 <yuval> o/ 14:07:20 <whoami-rajat_> a highlight of this would be to keep in mind the client release is earlier (M3) than the final project release 14:07:53 <whoami-rajat_> so it's better to keep your feature ready beforehand so reviewers have plenty time to test the client + feature changes 14:08:17 <whoami-rajat_> 2. Midcycle-2 (R-7 Aug-14) 14:08:21 <whoami-rajat_> #link https://releases.openstack.org/dalmatian/schedule.html#d-cinder-mid-cycle-ptg-2 14:08:43 <whoami-rajat_> we will be conducting another session of midcycle before the feature freeze to ensure everything is on track 14:09:36 <whoami-rajat_> 3. M3 (R-5, Aug-30) 14:09:42 <whoami-rajat_> #link https://releases.openstack.org/dalmatian/schedule.html#d-3 14:10:27 <whoami-rajat_> this has a bunch of deadlines like feature freeze, client library release, requirement freeze and a few more things 14:12:19 <yuval> "implement a feature" - this is also vendor feature? or pure cinder features? 14:12:24 <whoami-rajat_> other than that i couldn't find anything interesting related to our project 14:12:30 <whoami-rajat_> yuval, both 14:12:45 <yuval> I have a patch up - not sure if it counts as a feature 14:13:15 <yuval> its enabling a behavior for our driver 14:13:29 <yuval> thats a "feature"? 14:13:36 <whoami-rajat_> do you have a link? 14:14:10 <yuval> https://review.opendev.org/c/openstack/cinder/+/924323 14:15:31 <whoami-rajat_> looks like a feature to me 14:16:28 <yuval> I see but it does not need any client support 14:16:55 <yuval> anyway I would like it to be merge before aug 26 14:17:20 <yuval> sorry for interrupt please continue 14:18:17 <whoami-rajat_> yes it doesn't, i was just stating that for awareness if anyone has a feature that requires client support, it's better to get that support added on time 14:18:22 <whoami-rajat_> it doesn't apply to your case 14:18:47 <whoami-rajat_> I'm done with announcements, anyone has anything else to announce 14:22:28 <whoami-rajat_> ok looks like not 14:22:33 <whoami-rajat_> let's proceed with topics 14:22:46 <whoami-rajat_> #topic reminder: vendors (particularly quobyte, virtuozzo, and nfs-based drivers) should verify that the recent CVE-2024-32498 fix has not caused a regression 14:22:48 <whoami-rajat_> rosmaita, that's you 14:22:56 <rosmaita> thanks 14:23:03 <rosmaita> the topic pretty much says it all 14:23:24 <whoami-rajat_> i think netapp reported a scenario where it caused a regression 14:23:33 <rosmaita> #link https://review.opendev.org/c/openstack/cinder/+/923244 14:23:40 <rosmaita> oh, i did not see that yet 14:23:47 <msaravan> https://bugs.launchpad.net/cinder/+bug/2073146 14:24:17 <msaravan> this is the bug we logged..and hitting this issue when glance backend is based on Cinder NFS 14:24:21 <whoami-rajat_> i was looking into it and it might need to leverage the format information we store in admin metadata 14:24:29 <whoami-rajat_> but i didn't have an nfs env to confirm or work on it 14:24:38 <whoami-rajat_> but looks like a real issue 14:25:30 <msaravan> whoami-rajat_: I can share my environment, if you want to get into a call to see this 14:26:09 <yuval> "real size is 1" maybe thats in GB and the qcow side is in kb? 14:26:11 <rosmaita> msaravan: i may take you up on that 14:26:28 <msaravan> rosmaita: Sure, we can work on that. 14:26:30 <rosmaita> yuval: yes, looks like there's a unit mismatch 14:27:20 <rosmaita> msaravan: let's talk after this meeting and set something up 14:27:41 <msaravan> rosmaita: sure 14:28:45 <rosmaita> ok, that's all from me 14:29:01 <whoami-rajat_> thanks rosmaita 14:29:11 <whoami-rajat_> #topic Storpool clone-across-pools 14:29:15 <whoami-rajat_> rosmaita, that's you again 14:29:17 <kpdev> For storpool, i saw recent comment from Brian. @rosmaita, are you suggesting you ok with this feature but it does not need entry in support matrix ? regarding documentation, yes we can add that via separate PR under this blue-print 14:29:51 <rosmaita> yeah, i just wanted to give a pointer to the discussion of this Storpool feature 14:30:16 <rosmaita> (because i had completely forgotten it) 14:30:22 <whoami-rajat_> but i think it's not storpool specific, other driver vendors could also leverage it who support cross pool cloning? 14:30:31 <rosmaita> yes, exactly 14:30:34 <kpdev> +1 14:30:38 <rosmaita> #link https://www.youtube.com/watch?v=yvRVS9aic5o 14:30:49 <rosmaita> discussion starts at 2:18 and ends at 27:49 14:30:49 <whoami-rajat_> yeah we should move that forward 14:31:04 <yuval> simon here? 14:31:07 <whoami-rajat_> I was also planning to have similar support in case when cinder is glance backend 14:31:21 <rosmaita> kpdev: yes, i think as a team we are fine with the feature, but maybe this does not need to go into the support matrix 14:31:50 <kpdev> ok, i will remove support matrix entry 14:32:11 <jungleboyj> rosmaita: We would only have a couple of drivers implement this? 14:32:17 <rosmaita> kpdev: you can mention it prominently in the storpool driver docs 14:32:28 <kpdev> yes. sure will add documentation 14:32:38 <rosmaita> jungleboyj: don't know ... storpool at first, maybe some others 14:33:02 <rosmaita> i think we've had massive turnover in driver maintainers since may 2022 14:33:05 <jungleboyj> Ok. Then I don't think we need to add it to the matrix. Can re-address if more drivers start adding it. 14:33:24 <kpdev> ack 14:33:26 <rosmaita> so that's why i wanted to flag the video, so maybe other drivers that can do this too will know about it 14:33:44 <jungleboyj> The matrix would was intended to show features that consumers would be expecting. 14:34:07 <jungleboyj> kpdev: I think it makes sense for you to highlight it in your documentation. 14:34:46 <kpdev> yes, sure will add in doc after this gets added. 14:35:36 <yuval> regarding the bug: this is probably the offending line: if info.virtual_size != volume.size * units.Gi: 14:35:59 <yuval> in the nfs driver 14:37:38 <whoami-rajat_> rosmaita, thanks for bringing this up, i think we are good on this topic then? 14:37:50 <rosmaita> yeah, that's all from me 14:37:55 <whoami-rajat_> thanks 14:38:05 <whoami-rajat_> we don't have any more topics for today 14:38:20 <whoami-rajat_> we have a bunch of review requests from different driver vendors 14:38:28 <whoami-rajat_> so please take a look at them 14:38:36 <whoami-rajat_> let's move to open discussion 14:38:39 <whoami-rajat_> #topic open discussion 14:39:04 <yuval> I had an issue today 14:39:43 <yuval> In one of my patches there was a request to add the "Depends-on" (that was a good comment) but when I change the commit message gerrit created a new patch 14:39:56 <yuval> how should I continue with that close the old one? 14:40:18 <yuval> or I can somehow force my changes on the same patch 14:40:35 <yuval> this is the old one:https://review.opendev.org/c/openstack/cinder/+/903573 new one: https://review.opendev.org/c/openstack/cinder/+/924323 14:41:18 <eharney> you should be able to edit the Depends-On into the old patch, just make sure the Change-Id is in the footer of the commit message and not moved 14:41:49 <yuval> this what caused the change: I added "os_brick patch: https://review.opendev.org/c/openstack/os-brick/+/903574" 14:41:55 <yuval> (not the Depends-on) 14:42:41 <whoami-rajat_> did you do ``git commit`` instead of ``git commit --amend`` ? 14:43:00 <eharney> the new patch has two Change-Ids, probably added an empty line at the end or so 14:43:11 <yuval> no, it was added to the commit 14:43:18 <yuval> the commit msg 14:44:04 <yuval> maybe I should put the "os-brick" line before the new line? 14:45:14 <rosmaita> well, if you have the change-id by itself as the very last thing in the commit message, the commit-msg hook should see it and not add a new one 14:45:18 <rosmaita> (in theory) 14:46:33 <yuval> ah I see a new change-ID was added 14:46:56 <rosmaita> yeah, and tbh, i am not sure why that happened 14:47:30 <yuval> ok, never mind - I dont want to add any more patches 14:47:53 <yuval> I am leaving the new and closing the rest 14:50:00 <yuval> I would really appreciate someone going over these patches, they are tested and ready 14:52:42 <rosmaita> ok, just abandon the ones that we don't need to look at 14:53:21 <yuval> yes just done that, thanks 14:54:25 <yuval> regarding the nfs bug - I dont think there is a need to bring up the whole system, but through the unittest - test_initialize_connection - it could be tested 14:54:41 <yuval> there is a negative unittest test_initialize_connection_raise_on_wrong_size 15:00:18 <whoami-rajat_> yuval, i think the problem is before that logic, we are trying to do a qemu-img info on a qcow2 volume that is actually a raw volume 15:00:31 <whoami-rajat_> i don't have all the details but it will need a real configuration with glance using cinder 15:00:38 <whoami-rajat_> we are out of time 15:00:44 <whoami-rajat_> thanks everyone for joining 15:00:46 <whoami-rajat_> #endmeeting