13:59:26 <pdeore> #startmeeting glance
13:59:26 <opendevmeet> Meeting started Thu May 11 13:59:26 2023 UTC and is due to finish in 60 minutes.  The chair is pdeore. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:59:26 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:59:26 <opendevmeet> The meeting name has been set to 'glance'
13:59:27 <pdeore> #topic roll call
13:59:27 <pdeore> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
13:59:31 <pdeore> o/
14:00:06 <pdeore> croelandt, abhishekk I was thinking to wait for today, if there are no objections then I will set the +W
14:01:06 <pdeore> shall we start? I think everyone is here :)
14:01:07 <mrjoshi> o/
14:01:19 <croelandt> o/
14:01:40 <pdeore> ok, let's start
14:01:42 <pdeore> #topic release/periodic jobs updates
14:01:52 <rosmaita> o/
14:01:58 <pdeore> We are in M1 release week, and i think we are good to tag M1
14:02:11 <pdeore> waiting for the config refresh patch to merge, then I will submit the release patch today
14:02:18 <abhishekk> o/
14:02:22 <pdeore> for glance_store release we need to wait for os-brick release
14:02:50 <abhishekk> rosmaita, could you please keep watch on it?
14:02:53 <pdeore> rosmaita, can we update the glance_store hash after os-brick release
14:03:25 <abhishekk> you can update the hash now but release should be done after os-brick release
14:03:46 <rosmaita> ok, so what i would like to do is this
14:03:55 <rosmaita> the os-bricks have been released
14:04:00 <dansmith> o/
14:04:11 <rosmaita> but the upper constraints have not been updated
14:04:19 <rosmaita> due to a pep8 problem
14:04:41 <rosmaita> anyway, the glance-store change has merged, but i think we should update the min version of os-brick in setup.cfg
14:04:51 <rosmaita> before doing the new glance_store releases
14:05:03 <rosmaita> i have patches up for that
14:05:32 <rosmaita> gimme a sec to find the link
14:05:50 <rosmaita> https://review.opendev.org/q/topic:bug/2004555+project:openstack/glance_store
14:05:50 <pdeore> ack,
14:06:22 <rosmaita> another problem is that as far as i can tell, the stable/2023.1 upper constraints are not being updated by the bot
14:06:51 <rosmaita> but i can push a patch for that by hand
14:07:06 <rosmaita> after the pep8 stuff is fixed in requirements
14:07:27 <pdeore> the master patch is already merged, so we can update the hash, right?
14:07:59 <rosmaita> we haven't released os-brick from master yet
14:08:44 <pdeore> yeah, https://review.opendev.org/c/openstack/releases/+/882580
14:08:52 <pdeore> we will have to wiat for this
14:09:32 <rosmaita> or, not, if we don't update the 'extras' in setup.cfg
14:10:15 <rosmaita> pdeore: thanks for pointing out that patch, didn't realize elod had updated it
14:10:30 <dansmith> so one option would be to just release it without the dep and then follow up later with a bump right?
14:10:45 <dansmith> so that people *can* get the fix if they put the versions together
14:10:52 <dansmith> which right now they can't easily
14:11:36 <rosmaita> there is that
14:13:43 <pdeore> ok, so we can release the store after os-brick release without the dep?
14:14:32 <rosmaita> well, if we don't change the req, you can release glance_store now
14:14:34 <dansmith> that would be my preference, but on the gradient of concern, glance's exposure to this is probably the lowest of the three projects, so it's certainly less concerning than nova
14:15:09 <pdeore> ok
14:15:52 <rosmaita> pdeore: the situation is that the change in glance_store adds some parameters to a os-brick call that haven't changed, so using an old os-brick won't break anything with the new glance_store code
14:16:10 <rosmaita> (not sure i said that clearly)
14:16:24 <pdeore> yeah i got your point
14:16:43 <rosmaita> anyway, new glance_store + old os-brick should be fine, just doesn't give you full protection for the CVE
14:17:24 <pdeore> rosmaita, Is it possible for you to keep watch on this ? I will try to as much late as possible though
14:17:27 <rosmaita> but it's the "accidental part" anyway, so pretty low probability
14:17:47 <rosmaita> i can do it, my time zone is good for this
14:17:56 <pdeore> rosmaita, gr8 Thanks !
14:18:11 <pdeore> ok, let's move to next topic
14:18:19 <pdeore> Periodic job all green except oslo-tips failure,  may be again version conflict issue
14:18:28 <rosmaita> what we can do is, if the upper constraint hasn't changed by say 1800UTC, we can release without the req change
14:18:42 <pdeore> rosmaita, ok
14:18:44 <rosmaita> (sorry, i know you wanted to move on)
14:18:48 <rosmaita> one more question
14:19:09 <rosmaita> actually, not ...
14:19:10 <pdeore> np, I thought we are done :)
14:19:32 <rosmaita> we will  release all the glance_store versions at 1800 UTC
14:19:41 <pdeore> rosmaita, ack
14:19:47 <rosmaita> i can ping dansmith to verify the hashes on the release patch
14:19:53 <dansmith> sure
14:20:05 <rosmaita> ok, that's all from me, thank you
14:20:05 <pdeore> cool
14:20:13 <pdeore> rosmaita, Thanks
14:20:20 <pdeore> shall we move ahead ?
14:21:49 <dansmith> I think it's safe :)
14:21:58 <pdeore> ok, let me continue with the next peridoc job updates :)
14:22:09 <pdeore> so as I said 2 oslo-tips jobs are still failing with version conflict issue again I think
14:23:03 <pdeore> but i still wonder how the jobs which were failing till last week started passing without merging the changes submitted by croelandt ?
14:23:50 <pdeore> or I'm missing some context on this?
14:23:56 <croelandt> hm
14:24:05 <rosmaita> it's magic
14:24:09 <croelandt> yeah
14:24:10 <pdeore> :D
14:24:11 <croelandt> everything is magic
14:24:51 <pdeore> then I think we should wait everytime for this magic to be happened :P
14:25:21 <pdeore> croelandt, you have abandon the py38 to p310 patch right?
14:25:39 <croelandt> hm not really
14:25:43 <pdeore> so we need only the one where nodeset is changed
14:25:47 <croelandt> we wanted to wait & see the results ont he tips jobs
14:25:54 <pdeore> ohh ok
14:26:02 <croelandt> but we never test py310 with tips then?
14:26:05 * croelandt is confused
14:26:45 <pdeore> i think we did
14:27:01 <pdeore> ohh no, not tested
14:27:30 <croelandt> damn it feels like this should be tested somehow
14:27:51 <rosmaita> "this" == ?
14:28:10 <pdeore> ok, let's test that and check if these recently failing ones are passing or not
14:28:19 <pdeore> https://review.opendev.org/c/openstack/glance/+/882185
14:28:25 <abhishekk> this == py310 I think
14:28:29 <croelandt> rosmaita: the tips
14:29:46 <pdeore> move to next ?
14:30:23 <pdeore> moving ahead :)
14:30:24 <pdeore> #topic Important reviews
14:30:31 <pdeore> #link https://review.opendev.org/c/openstack/glance/+/881940 - Add new add loc api
14:30:31 <pdeore> #link https://review.opendev.org/c/openstack/glance/+/882498 - Add get loc api
14:30:42 <pdeore> I have submitted the new location apis changes, kindly please have a look
14:30:51 <croelandt> Would we like these merged for M1 or was it just the spec?
14:31:01 <dansmith> pdeore: I really want to review that I have just been slammed lately
14:31:26 <pdeore> croelandt, it would be nice if we can get those in m1 :)
14:31:38 * croelandt stretches fingers
14:32:00 <pdeore> #link https://review.opendev.org/c/openstack/glance-specs/+/880030 - Reproposed New Location API spec
14:32:07 <pdeore> on spec we have 2 +2s
14:32:22 <pdeore> I will go ahead and will set +W if there are no objections on the design until tommorw.
14:33:34 <pdeore> the sdk part for these new loc api, I've started working on, will submit the patch tmrw
14:33:46 <pdeore> any questions?
14:34:31 * croelandt has none
14:34:39 <pdeore> cool, moving to next
14:34:41 <pdeore> #topic Specs
14:34:49 <pdeore> we have 2 more specs for review, so kindly please have a look
14:34:55 <pdeore> #link https://review.opendev.org/c/openstack/glance-specs/+/880627 - Reproposed Image Encryption Spec
14:34:55 <pdeore> #link https://review.opendev.org/c/openstack/glance-specs/+/881951 - Spec-lite Add new Location Strategy
14:35:12 <rosmaita> ack
14:35:42 <pdeore> moving to next
14:35:44 <pdeore> #topic Can we delete multiple images from the same store
14:35:48 <pdeore> mrjoshi_ ^
14:36:22 <pdeore> #link https://review.opendev.org/c/openstack/python-openstackclient/+/882086/3/openstackclient/image/v2/image.py#1854
14:36:31 <mrjoshi_> this is regarding the ``glance stores-delete`` which deletes image from store
14:36:38 <mrjoshi_> I recently added a patch for adding support of ``glance stores-delete`` equivalent in  the openstack client .
14:37:22 <mrjoshi_> There's a comment from stephen regarding it ,Currently we can only delete a single image from the store , what if user's want to delete multiple images from the same store?
14:37:42 <rosmaita> they make multiple requests
14:38:13 <mrjoshi_> Can't we have it in a single request
14:38:31 <rosmaita> we don't have it for image-delete, do we?
14:38:42 <mrjoshi_> we do
14:38:52 <croelandt> we can do openstack image delete IMG1 IMG2 IMG3... ?
14:38:59 <mrjoshi_> yes
14:39:11 <abhishekk> we can certainly do glance image-delete 1 2 3
14:39:16 <croelandt> for the sake of consistency, it would make sense to have this behaviour everywhere
14:39:30 <croelandt> as a user it infuriates me when commands are not consistent
14:39:59 <rosmaita> that's not in the API, though, is it?  (batch delete, i mean)
14:40:07 <abhishekk> The proposal is to deleted different images from same store at once (we can decide this in next PTG)
14:40:14 <abhishekk> no that is not in API
14:40:22 <croelandt> but this could be in the CLI, right?
14:40:26 <abhishekk> yes
14:40:58 <abhishekk> its easy to do it in CLI
14:41:01 <rosmaita> so it wouldn't require any glance-side changes, not sure what the issue is here
14:41:15 <abhishekk> let me tell you
14:41:26 <abhishekk> we are adding support to delete image from a store in SDK
14:41:57 <abhishekk> the maintainer suggested rather than adding new command like image store delete we should pass --store option to existing delete command
14:42:32 <abhishekk> and then he asked us whether we want to add support to delete multiple images from same store
14:42:55 <abhishekk> So we can have image delete store image1, 2, 3 and so on
14:44:04 <rosmaita> so, that's an SDK design decision, that is, whether they want the SDK to diverge from the actual API
14:44:20 <croelandt> seems fine to me to have the SDK do 3 calls to the API to delete 3 images
14:44:35 <croelandt> rather than have the user type 3 commands to do the same 3 calls
14:46:04 <pdeore> +1
14:47:11 <pdeore> we have only few mins left
14:47:52 <mrjoshi_> so if everyone agrees can we implement this in the patch or should we have a separate patch for it?
14:48:47 <croelandt> mostly an SDK decision to be honest
14:49:08 <abhishekk> you can tell them that we can work on it later
14:49:35 <mrjoshi_> ok
14:50:14 <pdeore> shall we move to open discussions now?
14:51:06 <croelandt> yes
14:51:10 <pdeore> #Open Discussions
14:51:36 <pdeore> I will be on PTO from 18th-25th May
14:52:29 <abhishekk> I don't have anything to for Open discussion
14:52:32 <croelandt> Would be nice to fix this before M1 as well :) https://review.opendev.org/c/openstack/python-glanceclient/+/880696
14:52:32 <pdeore> so won't be there for next 2 weekly meetings, any volunteer to chair the meeting for next 2 week?
14:52:48 <croelandt> 18th is a public holiday in France
14:52:59 <croelandt> and I might take some PTO after that, gotta check my balance :p
14:53:11 <abhishekk> we can call off 18th
14:53:32 <abhishekk> and 25th if anything comes urgent then I will chair it
14:53:49 <pdeore> ok, I will send the mail accordingly
14:54:03 <pdeore> abhishekk, Thanks ! :)
14:54:14 <pdeore> That's it from me
14:54:29 <pdeore> anyone has anything else to discuss?
14:55:42 <pdeore> ok, assuming nothing left.. let's conclude for today !
14:56:03 <pdeore> Thanks everyone for joining !!
14:56:17 <pdeore> #endmeeting