13:59:26 <pdeore> #startmeeting glance 13:59:26 <opendevmeet> Meeting started Thu Aug 10 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 13:59:32 <pdeore> o/ 13:59:35 <mrjoshi> o/ 14:00:39 <pdeore> lets wait few minutes for others to join 14:00:58 <pdeore> abhishekk, croelandt , dansmith 14:01:06 <dansmith> o/ 14:01:07 <dansmith> sorry 14:01:10 <abhishekk> o/ 14:01:24 <pdeore> ohh np :) 14:01:34 <pdeore> let's start 14:01:37 <pdeore> #topic Release/periodic jobs updates 14:01:48 <pdeore> M3 is just 3 weeks from now 14:02:39 <pdeore> I just hope there would not be any surprise on location api :P 14:02:52 <pdeore> Periodic jobs all green except cursive-tips jobs failed on py38 & py39 during last run 14:03:17 <pdeore> moving to next 14:03:20 <pdeore> #topic New Location API 14:03:31 <croelandt> o/ 14:03:44 <pdeore> As per the discussion during last meeting, Cinder POC to test new location api as been submitted by Rajat 14:03:45 <pdeore> #link https://review.opendev.org/c/openstack/cinder/+/890655/2 14:03:49 <pdeore> croelandt, hey 14:04:18 <dansmith> so last I saw, whoami-rajat had some issue when trying to use it 14:04:20 <dansmith> was that resolved? 14:04:24 <pdeore> so while testing the location api via cinder, we had noticed a issues from glance side while cinder adding the location to an image but it's resolved now 14:04:32 <dansmith> okay 14:04:33 <pdeore> yeah it's resolved 14:05:01 <pdeore> It would be helpful if someone from nova team verifies these new apis from nova side 14:05:06 <abhishekk> I also verified it locally (cinder + glance) 14:05:10 <dansmith> so I was just looking, 14:05:38 <dansmith> we need to use a new method in the glanceclient to do this right? we need to make sure there's some way to identify that glance is too old to use the new api, so we fall back to the old one 14:05:49 <dansmith> is that not going to be handled in the glanceclient itself? 14:05:56 <dansmith> seems like the cinder poc doesn't handle that case in any way 14:06:10 <abhishekk> I think that should be part of client 14:06:21 <dansmith> me too, unless there's some reason it can't be 14:06:44 <abhishekk> no reason, it should be 14:07:15 <dansmith> okay so if that happens, then I think testing nova is as simple as a DNM depends-on patch with a zuul conf change to build glanceclient from git 14:07:38 <dansmith> I can put that up and confirm it fails (as expected) currently and then we can recheck that when it's fixed in the client 14:07:51 <abhishekk> +1 14:07:56 <pdeore> ack 14:08:09 <pdeore> Thanks ! 14:08:38 <pdeore> so, the tempest tests for add location api are still inprogress , I've just submitted the tests with single store, multistore tests are pending, I will try to submit the remaining tests by tommorow EOD 14:08:46 <abhishekk> So dansmith I have one question, when did nova snapshot calls add location 14:09:08 <abhishekk> is it when nova and glance uses ceph 14:09:08 <dansmith> pdeore: looks like the tempest patch is failing everything currently 14:09:08 <pdeore> Till then it would be really helpful if you have a look at glance side changes, as functional test i've added are covering most of the scenarios 14:09:18 <abhishekk> or nova and glance use same pool of ceph? 14:09:22 <dansmith> abhishekk: yeah, so the nova-ceph-multistore job 14:09:27 <dansmith> abhishekk: only when they use the same pool 14:09:41 <pdeore> dansmith, yeah i just saw that, will check it after meeting 14:09:50 <abhishekk> means if glance is using images pool, nova should use images pool as well? 14:10:07 <dansmith> abhishekk: right, and if not, we just upload the snapshot like normal 14:10:21 <abhishekk> when I see confs generated by nova multistore I don't see nova and glance using same ceph pool 14:10:35 <abhishekk> yeah, that upload process I know 14:11:10 <dansmith> that's the point of that job... so news to me, but I can look 14:11:26 <abhishekk> I was just trying to figure out how nova and glance should access same pool, I think keyring need to be modified for same? 14:11:54 <abhishekk> Ack, I might be wrong but there is one parmaeter in nova which points to glance rbd store but the pool is different 14:12:16 <dansmith> abhishekk: look at the rbd options in [libvirt] here: https://78a70965f904893ce65b-6333727e8e4cab8f6cba85bb2fbb6639.ssl.cf5.rackcdn.com/890509/3/check/nova-ceph-multistore/d2d7dd4/controller/logs/etc/nova/nova-cpu_conf.txt 14:12:51 <abhishekk> images_rbd_pool = vms 14:12:59 <abhishekk> and in glance it is images 14:13:21 <abhishekk> images_rbd_glance_store_name = robust 14:13:22 <dansmith> right but does that means glance can't see things in the other pool? 14:13:54 <abhishekk> that I am not sure, I was under impression that glance should be using same pool as nova 14:14:23 <dansmith> let me find the logs where we tell glance about the location and see if those are in our logs for that job, we can take it outside the meeting 14:14:36 <abhishekk> cool 14:15:03 <abhishekk> pdeore, could you please post your tempest link here or on the agenda page? 14:15:49 <pdeore> #link https://review.opendev.org/q/topic:New-Location-Apis - here you will get all the loc api realted patches :) 14:16:06 <pdeore> #link https://review.opendev.org/c/openstack/tempest/+/891026 14:16:12 <pdeore> tempest patch ^ 14:16:38 <abhishekk> this needs to depends on client patch as well 14:17:55 <pdeore> I will check and update after meeting 14:19:20 <pdeore> i don't think we need it, I have added the new add_loc method in te image_client, https://review.opendev.org/c/openstack/tempest/+/891026/1/tempest/lib/services/image/v2/images_client.py 14:19:43 <dansmith> tempest does not use the client 14:20:32 <abhishekk> ack 14:20:43 <abhishekk> then it is failing lot of jobs 14:21:00 <dansmith> yuuup 14:21:22 <abhishekk> and it should also check for min version required to test this job 14:21:30 <pdeore> yeah I've just quickly checked one job, but new tests are not failing there 14:22:00 <pdeore> abhishekk, yeah I was trying to find out for the existing version tests, but didn't get it 14:22:12 <pdeore> I will check it again and add it 14:22:26 <abhishekk> ack 14:22:34 <pdeore> shall we move ahead ? 14:23:24 <pdeore> ohk, moving ahead assuming it's yes :) 14:23:26 <pdeore> #topic Review Priority 14:23:38 <pdeore> #link https://review.opendev.org/q/topic:sqlalchemy-20+project:openstack/glance - SQLAlchemy 2.0 migration patches 14:24:12 <pdeore> please have a look, I think most of the patches got +W 14:24:25 <pdeore> abhishekk, croelandt Thanks ! 14:25:16 <pdeore> moving to next .. 14:25:18 <pdeore> #topic Importatnt Bugs ? 14:25:41 <pdeore> croelandt, anything you want to highlight here ? 14:26:55 <croelandt> not really :) 14:27:00 <pdeore> cool :) 14:27:12 <pdeore> let's move to open discussions ten 14:27:13 <pdeore> #topic Open Discussions 14:27:28 <abhishekk> when will be the last review party 14:27:58 <pdeore> 24th Aug 14:29:03 <pdeore> anyone has anything else to discuss ? 14:29:16 <abhishekk> we generally has it two weeks before, right 14:29:25 <mrjoshi> #link https://review.opendev.org/c/openstack/openstacksdk/+/881939 #link https://review.opendev.org/c/openstack/python-openstackclient/+/878631 14:29:25 <mrjoshi> #link https://review.opendev.org/c/openstack/python-openstackclient/+/883494 14:29:27 <abhishekk> that means it should be next week 14:29:35 <croelandt> we could move it to next week if you'd like 14:29:43 <mrjoshi> abhishekk, croelandt need you attention on these patches 14:29:56 <abhishekk> mrjoshi, ack, will have a look 14:29:57 <pdeore> we have it one week before but yea let's move it to next week 14:30:04 <croelandt> pdeore: ok :) 14:30:08 <croelandt> mrjoshi: I'll look at them 14:30:18 <croelandt> we can also include them in next week'sreview :) 14:30:18 <mrjoshi> thanks! 14:30:29 <abhishekk> :D 14:30:47 <pdeore> :) 14:31:27 <pdeore> anything else ? or we can wrap up early ? 14:31:35 <croelandt> let's wrap up! 14:31:45 <pdeore> ack :) 14:31:53 <pdeore> Thanks everyone for joining!! 14:32:06 <pdeore> #endmeeting