14:00:57 #startmeeting cinder 14:00:57 Meeting started Wed Mar 15 14:00:57 2023 UTC and is due to finish in 60 minutes. The chair is whoami-rajat. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:57 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:57 The meeting name has been set to 'cinder' 14:01:01 #topic roll call 14:01:49 o/ 14:01:52 o/ 14:01:58 o/ 14:02:11 o/ 14:02:11 o/ 14:02:31 hi 14:02:45 o/ 14:02:51 #link https://etherpad.opendev.org/p/cinder-antelope-meetings 14:02:54 o/ 14:03:07 o/ 14:03:17 \o 14:03:39 o/ 14:04:09 o/ 14:04:29 hello 14:04:35 good turnout, let's get started 14:04:40 #topic announcements 14:04:44 RC2 this week 14:04:54 17 March 14:05:30 o/ 14:05:35 we did release RC1 last week but we still have changes to get in into 2023.1 release 14:05:42 so we are planning to do RC2 14:05:51 #link https://etherpad.opendev.org/p/cinder-antelope-fixes-rc 14:06:19 I was reviewing the remaining patches and all are approved as of now except one 14:06:31 #link https://review.opendev.org/c/openstack/cinder/+/852654 14:06:57 I need to review it again after my comments were addressed, will do it after the meeting 14:07:10 in the meantime please review the above patch since that's the only one remaining 14:07:20 and zaitcev is fast in addressing comments 14:07:39 also thanks everyone for reviewing the other changes! 14:08:24 there is also my patch related to multiattach but it has conflicts on tempest side so nothing remaining from cinder side 14:08:38 that's it for RC2 14:08:53 next, 2023.2 (Bobcat) PTG 14:09:03 #link https://etherpad.opendev.org/p/bobcat-ptg-cinder-planning 14:09:18 we've Bobcat PTG in less than 2 weeks 14:09:31 o/ 14:09:35 so please add topics to the etherpad if you would like them to be discussed 14:10:23 i still don't see a lot of topics and it would take time to arrange them day wise (also there are cross project things) so please add them timely 14:10:44 next, OpenInfra live: Cinder update 14:10:51 #link https://lists.openstack.org/pipermail/openstack-discuss/2023-March/032638.html 14:11:18 we will be having project updates for 2023.1 release and i signed up for providing the cinder update 14:11:26 \o/ 14:11:32 so if you would like to see it, it should be live on youtube on Thursday, March 23 at 1500 UTC 14:11:45 \o/ 14:12:07 rosmaita, did a great update for Yoga and I missed it for Zed so this time I didn't forget about it 14:12:41 next, Cancel next weeks meeting? 14:12:58 we're close to PTG and next week is 2023.1 release 14:13:13 I don't think we will have a lot of topics, and if we do then better to discuss them at PTG 14:13:27 yay 14:13:33 but wanted to know what the team thinks, should we cancel meeting next week? 14:13:57 no objections from me 14:14:08 btw, i really like the enthusiasm in today's meeting! 14:14:43 That is good for me. 14:15:04 whoami-rajat: It is so much better when people respond, isn't it. :-) 14:16:07 jungleboyj, that's true, sometimes I feel I got disconnected when I'm the only one typing 14:16:22 :-) Been there. 14:16:33 :) 14:16:57 so I don't see any objections, I will send a mail to ML regarding canceling next week's meeting 14:17:04 sounds good! 14:17:20 that's all for the announcements 14:17:28 let's move to topics 14:17:33 #topic Outreachy update 14:17:35 enriquetaso, that's you 14:17:37 hi 14:18:00 I wanted to give an update regarding the outreachy applicants on the mailing list. I want to thank Jeremy Stanley and everyone who has taken the time to reply to their questions. 14:18:04 just wanted to mention, I see a lot of participation of outreachy members on the ML which is really good 14:18:18 true! 14:18:22 So far, this round of applicants looks good, and we have received a few contributions. As outreachy is often the first job in tech for many applicants, if you review these patches, please keep this in mind. Please highlight good practices or anything you think can help the applicants with their careers. 14:18:52 The deadline for this round is this month, and applicants need at least one patch merged in order to apply. So, if you could review these patches as soon as possible, it would be greatly appreciated. 14:18:57 #link https://review.opendev.org/c/openstack/cinder/+/877185 14:19:01 #link https://review.opendev.org/c/openstack/cinder-tempest-plugin/+/876905/ 14:19:05 #link https://review.opendev.org/c/openstack/cinder/+/877070 14:19:59 a lot of contribution which I wasn't aware about 14:20:14 i will commit to looking at those 14:20:17 today 14:20:24 enriquetaso, can you add these to the review request section so we've more visibility? 14:20:35 Yes.. most of them forgot to link the launchpad bug to the patches, maybe that's why, not sure 14:20:50 sure whoami-rajat 14:20:58 on that note, who has committed to looking at pete's restore-sparse patch? I know whoami-rajat said he will re-review, but who else? 14:20:59 I will take a look at the patches by this week (after I'm free from RC2 stuff) 14:21:20 thank whoami-rajat, and thanks everyone on the cinder team 14:21:22 I will do a review pass by today 14:21:28 rosmaita++ 14:21:43 enriquetaso++ for the great work on outreachy side! 14:22:02 sure, no worries :D 14:22:05 i agree, thanks enriquetaso 14:22:06 enriquetaso: i was just glad to be of help 14:22:39 I would really like rosmaita's question to be answered, anyone up for taking ownership of zaitcev's patch? 14:22:48 as the second core 14:23:01 I remember Jon reviewed it a couple of times 14:23:03 i can make time to review, but it might be better for someone who's already reviewed it 14:23:08 It's this: https://review.opendev.org/c/openstack/cinder/+/852654 14:23:29 there weren't a lot of folks around to answer question on the mailing list over the weekend, which seemed to be when some of the interns were trying to learn the workflow 14:23:31 and he's a core now, so he would be best 14:23:58 that's right, congratulations jbernard! 14:24:31 fungi: thanks for stepping up to answer, that was really helpful 14:25:02 rosmaita, ... I totally forgot to announce that 14:25:15 i saw it on the ML 14:25:20 at least i think so 14:25:58 sorry enriquetaso, we've totally mixed up multiple discussions with your topic 14:26:07 I will let that finish first 14:26:18 lol 14:26:22 thanks fungi for helping the outreachy interns out! 14:26:30 ++ 14:26:32 my topic is finish! 14:26:39 nothing else for my side 14:26:43 we went from no participation to too much participation! 14:26:49 fungi++ 14:27:18 yes, rosmaita, I'm shocked, i think leaving devstack aside for the first contribution maybe related 14:27:36 :D 14:27:54 that pytest/tox conflict is really troubling though 14:28:54 and further evidence we should be showing contributors how to use some sort of isolation like a venv rather than pip installing random tools into the system python context 14:29:05 yes.. it think devstack is having some problems with tox 4 as well 14:29:06 yeah, even I struggled with devstack recently and the issue faced by me was only faced by one other person dansmith who helped me resolve it 14:29:37 maybe it's the same issue i'm facing running tox whoami-rajat 14:29:41 clarkb has a change up to switch devstack to using venvs, i gether it's still mostly in good shape despite having sat unmerged for several years now 14:30:07 pep 668 is going to pretty much force our hand in implementing something like that anyway 14:32:30 so looks like outreachy interns are doing good progress and we need to review their changes actively 14:32:31 * enriquetaso fingercrossed 14:33:52 ok we've 2 more topics, anything else on this one enriquetaso ? 14:34:15 nop 14:34:24 great, thanks for the update! 14:34:40 #topic Discuss patch https://review.opendev.org/c/openstack/cinder/+/868485 that changes schema validation 14:34:44 #link https://review.opendev.org/c/openstack/cinder/+/868485 14:34:50 drencrom, that's you 14:34:52 hi o/ 14:35:09 I added this to the topics as I got a -1 telling that it should be discussed here 14:35:28 It already has a +2 14:36:32 i have a question about your schema definition ... i thought glance allowed arbitrary UTF-8 as property names? or is that only for values? 14:36:41 the commit message makes sense since that information is coming from glance 14:36:43 hi, sorry that was me, because I'm just raising attention because the patch is changing api/schemas 14:37:09 not sure why we're using extra_specs parameter type for validating glance image metadata, the change might modify extra_specs validation as well, I will add a comment 14:37:19 whoami-rajat: ++ 14:38:06 I did not change anything on the properties names actually, just made the values larger 14:38:45 I'm not sure about glance spacification for keys, I'd have to check 14:39:10 well, tbh, i guess cinder has never allowed UTF-8 for that, and no one has complained 14:41:45 I think extra_specs is also used in other places, that is why I added a new one 14:42:12 i suggest not calling it extra_specs at all 14:42:54 yeah, I'm not referring to your change but the convention we're using, extra_specs for volume, backup, image metadata doesn't look right 14:43:31 as rosmaita said, would be better to call it volume_image_metadata or something 14:45:08 rosmaita, or I can add a comment later but i think the discussion was about if the intent of the patch is correct 14:45:32 that is to allow metadata values greater than 255 characters 14:45:50 which IMO, if glance allows it, we should too 14:46:01 it's their image metadata after all 14:46:23 characters or bytes? if you get into multi-byte encodings that will make a difference 14:46:24 i was looking at the glance schema, some metadata is limited, but some are just JSON "string" type 14:46:49 Yes, that was how I found the bug. Dealing with large metadata from Glance. 14:47:31 a reviewer just needs to verify that we don't need a db change for this, and then it sounds ok to me 14:47:50 alternative would be to truncate glance metadata values at 255 chars 14:48:12 fungi: good question, i think mysql text is char count 14:48:34 i mean, we would truncate when we record them in cinder 14:48:40 for mysql it will depend on the field type and encoding specified in the schema 14:49:19 i can't remember if the encoding is real utf-8 or that weird mysql utf-8 14:49:27 there are at least a couple of utf8 types in mysql, for 3-byte and 4-byte 14:49:49 The DB is OK IIRC, there were cases where the metadata values where larger that 255 14:49:53 it gets confusing, especially for users who created the table with one and then the software tries to write in another 14:50:45 drencrom: yeah, but you are going to allow 65535 arbitrary unicode chars now 14:50:46 and yes, mysql field width limits are in bytes not characters, so the utf8 encodings basically shrink the allowed widths proportionally to their character sizes 14:51:16 so if it's 65535 bytes, we need to make the char count in the schema smaller 14:51:28 it can still be very large 14:52:16 don't we have the sam issue with the 255 limit? 14:52:43 probably, but it's limited ... so the schema rejects before the database barfs 14:53:27 like the keys are limited to basically a subset of ascii, which in utf-8 are all only one byte each 14:53:53 so char count == byte count for those 14:55:25 What about the other cases where extra_specs is used? Is the database size always larger that 255? 14:56:17 i don't know, tbh, but we have pretty tight control on what are used in extra_specs 14:56:51 we've 4 minutes left and one more topic, so let's continue this discussion on the patch or in #openstack-cinder channel 14:56:55 after the bug squad meeting 14:57:04 let's discuss the final topic 14:57:07 ok 14:57:17 #topic stable/rocky status 14:57:20 rosmaita, that's you 14:57:33 py27 unit tests have been failing for a month 14:57:49 let's talk about EOL at the vPTG 14:58:13 instead, i want to ask any cinder-coresec members to take a quick look at https://bugs.launchpad.net/cinder/+bug/2008705 14:58:28 rosmaita, sure, please add the topic to the planning etherpad https://etherpad.opendev.org/p/bobcat-ptg-cinder-planning 14:58:30 just need feedback on whether it's really a security issue, or we can work on it in the open 14:58:45 if not a security issue, we can discuss at the PTG also 14:58:52 whoami-rajat: will do 14:59:01 thanks 14:59:25 will take a look at the security issue though I'm not super familiar with it 14:59:41 thanks for the reminder. more generally, cinder-coresec feedback is encouraged on the other private security bugs for cinder and its deliverables 14:59:41 we've 3 review patches from otureachy members in the review request section 14:59:45 please take a look at them 15:00:37 we're out of time now, thanks everyone for attending and also making 2023.1 a successful cycle 15:00:40 see you all at the PTG! 15:00:43 #endmeeting