14:00:37 <jokke_> #startmeeting glance 14:00:40 <openstack> Meeting started Thu Dec 6 14:00:37 2018 UTC and is due to finish in 60 minutes. The chair is jokke_. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:44 <openstack> The meeting name has been set to 'glance' 14:00:46 <jokke_> #topic roll-call 14:00:50 <jokke_> o/ 14:01:27 <rosmaita> o/ 14:02:14 <abhishekk> o/ 14:02:28 <jokke_> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:03:43 <jokke_> #topic updates 14:03:56 <jokke_> I have only one thing on my list 14:04:23 <jokke_> our survey closes today, 9 responses and so far everyone are up for the default visibility change 14:04:39 <rosmaita> hooray! 14:04:44 <jokke_> so unless something dramatic happens today, we should be good to go with that 14:05:00 <abhishekk> great 14:05:06 <LiangFang> good 14:05:24 <rosmaita> i wonder if it will break tempest again :( 14:05:49 <jokke_> rosmaita: time will tell, most likely ;) 14:06:40 <jokke_> we'll fight that battle once we get the patch up to see how many thing it breaks. I'm assuming it breaking somewhere between lots and tons of our own tests as well 14:07:28 <jokke_> #topic release/periodic jobs updates 14:07:48 <abhishekk> that would be me 14:08:00 <abhishekk> We are in R-18 week, Stein 2 milestone will be on 11th January which will be exactly 1 month away from now 14:08:25 <abhishekk> so focus on our targets for milestone 2 and help with the reviews as well 14:08:25 <jokke_> abhishekk: you've got opportunity to follow if we have anything we need to take care of at this point? I'd assume we need at least stable cuts once the backports are in 14:08:41 <abhishekk> jokke_, right 14:08:57 <adam_zhang_> hello, everybody 14:09:01 <abhishekk> I have 2-3 patches up which needs to be backported as well 14:09:36 <abhishekk> Regarding periodic jobs, 2-3 jobs are failing today, yesterday it was all green 14:09:49 <jokke_> :( 14:09:58 <jokke_> any idea yet what broke? 14:10:00 <abhishekk> most of the tests are failing with String exception 14:10:35 <abhishekk> jokke_, not at the moment, but these are intermittent 14:10:38 <jokke_> that's not good, sounds like something really broke 14:10:56 <abhishekk> I will try to spend some time on this tomorrow 14:11:08 <jokke_> sounds great, thanks 14:11:30 <abhishekk> we can move to next topic 14:11:41 <jokke_> ok 14:11:53 <jokke_> #topic python-glanceclinet v1 support 14:12:00 <jokke_> rosmaita: you wanted to highlight this 14:12:12 <rosmaita> yeah, just in case anyone forgot 14:12:27 <rosmaita> i don't think i mentioned it last week because it's not really a config option 14:12:49 <rosmaita> but we have said that v1 client support ends in stein 14:13:13 <jokke_> that's a lots of cleanup to be happening \\o \o/ o// o/7 14:13:20 <rosmaita> i don't know who has bandwidth / how we want to prioritize it 14:13:39 <rosmaita> but we do still have the tests running against queens v1 glance 14:13:55 <rosmaita> so if we don't remove it, we can still be sort of confident that it still works 14:13:59 <jokke_> from master? 14:14:10 <rosmaita> i think so, gimme a sec to check 14:14:15 <abhishekk> rosmaita, probably I will take that 14:14:47 <jokke_> I was more thinking of dropping the tests and the option from the client to use it and start cleaning the code base block by block 14:14:52 <rosmaita> #link http://git.openstack.org/cgit/openstack/python-glanceclient/tree/.zuul.yaml 14:15:00 <jokke_> similar way as with the service 14:15:02 <rosmaita> looks like we chackout glanceclinet master 14:15:10 <jokke_> kk 14:15:26 <rosmaita> yeah, stable/queens devstack with glanceclient master 14:15:44 <jokke_> ok, so we can drop that test 14:16:07 <rosmaita> right, after we eliminate the code :) 14:16:43 <jokke_> just before ... can't drop the code if that change is gated agaist the test testing it is still there :P 14:16:44 <abhishekk> we also have some v1 related code in glance code base as well 14:16:54 <jokke_> abhishekk: yes we do 14:17:13 <jokke_> there is quite a few bits that still needs cleanup 14:17:46 <abhishekk> yes 14:18:51 <jokke_> ok, that's good reminder, thanks Brian 14:19:03 <rosmaita> happy to be of service 14:19:14 <jokke_> #topic Foreign key constrain patch 14:19:24 <jokke_> LiangFang: I think this was yours 14:19:40 <LiangFang> this is just a call to rosmaita:) 14:19:45 <rosmaita> i noticed! 14:19:51 <LiangFang> yes 14:20:03 <rosmaita> looks like this meeting will end early, i'll use the time to look at it 14:20:15 <rosmaita> it looks ok, but i wanted to look more carefully 14:20:22 <LiangFang> thanks, I have a question to rosmaita 14:20:23 <jokke_> #link https://review.openstack.org/#/c/617889/ 14:20:34 <rosmaita> sure 14:20:42 <LiangFang> postpone to open discuss section 14:20:46 <rosmaita> ok 14:20:54 <jokke_> well lets get in there 14:21:04 <jokke_> #topic Open Discussion 14:21:18 <abhishekk> I have PoC ready for cache-manage using v2 14:21:37 <jokke_> abhishekk: you wanna link those patches here you needed eyes on and needed backporting as well 14:21:38 <rosmaita> cool 14:21:39 <abhishekk> just pending some time to test it, may be couple of days 14:21:58 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2018-December/000665.html 14:22:04 <abhishekk> s/pending/spending 14:22:08 <abhishekk> jokke_, sure 14:22:22 <abhishekk> https://review.openstack.org/620781 14:22:23 <rosmaita> i don't know what to respond, i wasn't really paying a lot of attention to the cache management in the rocky cycle 14:22:29 <abhishekk> https://review.openstack.org/618468 14:22:36 <abhishekk> https://review.openstack.org/618093 14:22:56 <abhishekk> I will reply to that mail 14:23:02 <rosmaita> excellent 14:23:04 <rosmaita> thanks 14:24:16 <abhishekk> but I haven't received that mail :( 14:24:33 <jokke_> #link https://review.openstack.org/620781 14:24:42 <jokke_> #link https://review.openstack.org/618468 14:24:51 <jokke_> #link https://review.openstack.org/618093 14:25:02 <rosmaita> abhishekk: that's weird 14:25:24 <abhishekk> yeah 14:25:27 <rosmaita> but i have noticed that deliveries from openstack-discuss seem to lag from when stuff shows up on the website 14:25:44 <abhishekk> not in my inbox, searched all of it now 14:26:12 <abhishekk> that's why in the first place it didn't come to my attention I guess :( 14:26:21 <rosmaita> probably! 14:26:47 <LiangFang> rosmaita: i want to confirm with you, about refactoring location policy code, will you do this with two month? 14:27:01 <LiangFang> within two month? 14:27:05 <rosmaita> jokke_: how about you? 14:27:11 <abhishekk> jokke_, we need to discuss about location patch submitted by imacdonn 14:27:14 <jokke_> it's great when decently working things needs to be changed just for the sake of change, isn't it ;) 14:27:42 <rosmaita> LiangFang: i will not be doing any location code refactoring 14:27:48 <rosmaita> someone needs to, though 14:27:57 <LiangFang> seems have network delay 14:28:15 <jokke_> tbh I'm just incredibly scared to touch that 14:29:18 <abhishekk> I guess your concern about nova snapshot with location does not have to do with this patch, snapshot will create a new image it will not give a call to add_location (as per my understanding) 14:29:34 <jokke_> I know we have more and more need to do it and we still have the feature branch for it, so I think just need to take a hit and start poking it and see what all breaks 14:30:13 <jokke_> abhishekk: I'm not sure if nova calls add or replace 14:30:39 <abhishekk> Also we discussed 2 days before that we need to change nova and cinder to use the multiple backend support provided by glance, in that way when nova calls snapshot create they can specify which backend snapshot should be uploaded 14:30:42 <jokke_> but nova uses the locations API when it creates the ceph snapshot to register the location to the image 14:30:52 <abhishekk> same goes for cinder upload-volume-to-image 14:31:43 <jokke_> I do not know about cinder, do they upload the image or do they do it inband as well? 14:32:05 <abhishekk> jokke_, I will try to understand nova snapshot code 14:32:11 <jokke_> actually it must be add as that image goes to active 14:32:37 <abhishekk> I have tried to setup the environment but it is failing 14:33:58 <jokke_> abhishekk: thry to get in touch with some of our internal Nova folks. I'm very hopeful there is someone who understands how the ceph integration code works there 14:34:17 <abhishekk> jokke_, will do 14:34:51 <abhishekk> will update in/before next meeting 14:35:22 <abhishekk> kindly review my patches :D 14:35:53 <jokke_> iiuc the workflow is that Nova and Glance uses the same ceph backend for ephemeral storage and image store, Nova accesses the locations via the locations api and uses the direct path, when snapshotting it just takes snapshot on the ceph side and registers new image to glance and adds the location to the image 14:36:03 <jokke_> abhishekk: I will do 14:36:35 <abhishekk> jokke_, ack 14:37:15 <jokke_> In ceph that is just cow snapshot, which is also the reason why we have the delete dance on the ceph driver as ceph refuses to delete the parent as long as there is snapshots of the image 14:38:03 <abhishekk> jokke_, will go through it, thanks for the pointers 14:38:17 <jokke_> hopefully that helps 14:38:23 <jokke_> anything else? 14:38:25 <LiangFang> abhishekk: which failure when you setup environment? 14:38:44 <abhishekk> LiangFang, http://paste.openstack.org/show/736692/ 14:38:48 <LiangFang> ok 14:38:50 <abhishekk> jokke_, that's it 14:39:08 <jokke_> LiangFang: rosmaita: ? 14:39:29 <abhishekk> LiangFang, http://paste.openstack.org/show/736693/ 14:39:31 <LiangFang> I site beside a ceph team 14:39:36 <rosmaita> oops, not paying attention 14:39:38 <abhishekk> this is my local.conf 14:39:39 <LiangFang> sit 14:40:03 <LiangFang> so if you encounter some issue about ceph, I may can get help from them 14:40:06 <abhishekk> LiangFang, great. please drop me a mail if you find something 14:40:25 <jokke_> ok, 20min win! \\o \o/ o// o/7 14:40:29 <jokke_> Thanks all! 14:40:33 <LiangFang> thanks 14:40:40 <abhishekk> thank you all 14:40:48 <jokke_> #endmeeting