14:01:03 #startmeeting cinder 14:01:03 Meeting started Wed Apr 19 14:01:03 2023 UTC and is due to finish in 60 minutes. The chair is whoami-rajat. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:03 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:03 The meeting name has been set to 'cinder' 14:01:05 #topic roll call 14:01:14 #topic roll call 14:01:26 o/ 14:01:29 o/ 14:01:29 hi 14:01:34 o/ 14:01:36 o/ 14:01:37 o/ 14:01:41 o/ 14:01:45 o/ 14:01:50 o/ 14:01:52 o/ 14:03:36 Hi all. Is there an agenda for today's meeting? 14:03:51 #link https://etherpad.opendev.org/p/cinder-bobcat-meetings 14:03:55 drencrom, ^ 14:05:32 hi 14:05:46 o/ 14:05:55 hello everyone, let's get started 14:06:02 #topic announcements 14:06:07 hi 14:06:08 first, Xena EM 14:06:15 #link https://review.opendev.org/c/openstack/releases/+/878882 14:06:44 I reviewed all the open patches in xena and saw many of them not merged into more recent branches 14:06:50 * jungleboyj sneaks in late 14:06:52 so we've excluded them from the final xena release 14:07:30 looks like Jon isn't around but i will ping him later to propose a final release patch 14:07:48 and then move xena to EM (as done in the release patch above) 14:09:19 we can get the "pin tox <4" patches into xena later, but we should get them in (and out of my life) 14:09:51 ah yes, maybe i missed reviewing them or already done it, can't remember ... 14:10:15 i think you reviewed them 14:11:11 doesn't matter for release, just thought I'd mention it :) 14:11:59 yep i see i haven't, will do that after the meeting (since we're sticking with tox) 14:12:33 next is, Update on EOL Stein and Rocky 14:12:44 #link https://lists.openstack.org/pipermail/openstack-discuss/2023-April/033386.html 14:13:27 so Elod replied to rosmaita's email saying that they will put up a rocky-eol patch 14:13:32 for all projects 14:13:49 it's still not clear for stein but hopefully the same will be done for that 14:14:11 from cinder perspective, we won't be actively reviewing patches in those branches 14:14:22 i think we (cinder project) can just go ahead and EOL stein ourselves 14:14:30 nobody objected on the ML 14:15:42 yeah that's also possible 14:16:22 I was thinking if the whole openstack does it, it would be even better 14:16:43 iirc, nova and neutron have done it already 14:18:21 ok, we can do it for cinder then, Jon or I can put up a release patch for it 14:18:35 i think Jon is planning to do it 14:18:56 yeah, just checked, nova and neutron both have rocky-eol and stein-eol tags 14:20:45 ok great, let's do it for cinder as well 14:22:48 final announcement, Festival of XS reviews this week 14:23:06 #link https://etherpad.opendev.org/p/cinder-festival-of-reviews 14:23:17 we will be having festival of xs reviews on 21 April 14:23:22 which is this friday 14:23:23 wow, it's the 3rd week of april already 14:23:34 i will send out a mail later today 14:23:39 :-) April has been very fast. 14:24:02 I'm amazed it's April already :D 2023 is also fast 14:25:30 anyway, that's all the announcements 14:25:36 i see jbernard joined in 14:25:46 o/ 14:25:49 we were discussing about rocky and stein EOL 14:26:02 for rocky, the release team will be putting up patches as written in their email 14:26:28 and for stein, rosmaita suggested that we put up our own patch 14:26:43 im workign on it now, no problem 14:27:10 great thanks! 14:27:20 jbernard: probably do both rocky and stein so there's no confusion 14:27:32 that was my plan 14:27:48 i can abandon anythign that conflicts with release team proposals 14:27:54 if the release team hasn't done it already then sounds good ^ 14:27:55 sounds good 14:28:55 i think that's all for announcements, let's move to topics 14:29:14 #topic Discuss patch that changes schema validation 14:29:17 #link https://review.opendev.org/c/openstack/cinder/+/868485 14:29:21 drencrom, that's you 14:29:56 Yes, we discussed this a few weeks ago. The point is that the limit control counts characters but mysql limit is in bytes 14:30:27 I've checked what glance does and it seems it just fails when mysql fails to insert the data 14:30:54 And does not control the size 14:32:27 so basically, this changes the request schema but doesn't have to mess with the database because that's already big enough 14:32:43 the database has a 64k cap? or is uncapped? 14:33:08 Mysql text fields have a 64k limit 14:33:40 but text can have utf8 chars so that the char limit is not really useful 14:34:38 drencrom: do you mean the mysql field is measured in bytes, not chars? 14:34:52 yes 14:35:26 ok, since the stuff is coming in via json, it's definitely utf8 14:35:49 iirc, worst case for utf8 is 4 bytes? 14:36:02 (i mean to represent a character) 14:36:08 yes 14:36:51 so should we limit the field to 64K / 4 ? 14:37:00 i'm a bit unsure why our image metadata table uses a Text field when other tables use a String field for such things 14:37:38 glance uses a text field also 14:37:53 i guess just because we wanted it to hold a larger value 14:40:12 can we have the schema validator just check it as bytes instead of string length? 14:40:28 In theory it can still fail with a 16k chars limit if glance has a text shorter than 64k but larger than 16k chars 14:40:46 i think eharney is onto something here 14:41:16 we can have the custom validator determine the byte size and reject if necessary 14:41:49 key think is that would give you a 400 before even contacting the DB 14:41:52 *thing 14:42:49 that looks like it shouldn't be too hard to do 14:43:37 I don't know how to do that but can help if you give me some pointers 14:43:52 (add a new jsonschema validator method in validators.py etc and change the type to a new type instead of "string") 14:44:37 drencrom: ping me if you don't see a good example in validators.py and we can figure something out 14:46:12 ok, will look into it, thanks 14:46:36 great 14:49:14 cool, is that all for your topic drencrom ? 14:50:43 yes 14:51:18 to summarize, sounds like we are OK with increasing the value allowed because no database change is needed, but we would like a validator so that we can reject a string that will be too big in byte size 14:55:00 thanks for the summary rosmaita 14:55:05 that's all the topics we had 14:55:08 let's move to open discussion 14:55:11 #topic open discussion 14:55:56 Hey guys can we get a review on https://review.opendev.org/c/openstack/cinder/+/874813/4?tab=comments https://review.opendev.org/c/openstack/os-brick/+/876284 14:57:01 Also JeanPierre would like to discuss how to proceed on https://bugs.launchpad.net/cinder/+bug/2002535 14:57:01 Tony_Saad: Looking. :-) 14:57:06 Thanks! 14:59:28 Tony_Saad: Why is it needed to document a generic feature like image volume cache in a driver documentation? 15:00:01 Or actually, let's document my objection with a -1, not here. 15:00:12 I believe it is documented in Dells other producs 15:01:11 we're out of time 15:01:15 thanks everyone for attending 15:01:16 #endmeeting