14:00:08 #startmeeting cinder 14:00:08 Meeting started Wed Jan 10 14:00:08 2024 UTC and is due to finish in 60 minutes. The chair is whoami-rajat. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:08 The meeting name has been set to 'cinder' 14:00:13 #topic roll call 14:00:19 hi 14:00:31 o/ 14:00:31 hey everyone! 14:00:32 Hi 14:00:36 o/ 14:00:47 o/ 14:01:34 o/ 14:01:36 hi 14:01:50 #link https://etherpad.opendev.org/p/cinder-caracal-meetings 14:02:58 hello everyone 14:03:38 Happy new year to everyone who just came back from the break! 14:03:48 #topic announcements 14:03:57 Openstack "D" Release Voting 14:04:03 #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/N7OCEHENUZNMKKFOBBYRSJXEDJVJMGXI/ 14:04:17 the last day for voting is friday 14:04:21 we have 3 options 14:04:25 Dalmatian 14:04:25 Daikon 14:04:25 Dijon 14:04:27 shame they couldn't come up with some wild animals to follow the trend. Not the best choices to pick from. 14:04:41 o/ 14:04:43 last i checked, Daikon was winning 14:05:06 simondodsley: they're all foods (depending on where you live) 14:05:32 i see that - but it bucks the last 3 names trend - if there was one 14:06:27 i guess E will be Eggplant then... 14:06:41 you've got my vote! 14:07:14 Jeremy will be looking forward to the Fungi release 14:07:26 lol 14:07:43 :D 14:07:47 Deadline: January 12, 2024, at 20:00 UTC. 14:08:08 make sure to vote! 14:08:26 next, Team Signup for April 2024 PTG 14:08:32 #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/MIK6GXZXUFS5M3UOVILLDNDGURGH7UQV/ 14:09:02 the next PTG is planned to be conducted on the following dates: April 8-12, 2024 14:09:24 I've Signed up the Cinder team for PTG, which obviously we will be attending 14:10:05 \o/ 14:10:28 i can only do the first 3 days (if you care) 14:11:04 oh 14:11:08 so Cinder PTG (based on the past vPTG schedule) starts on the 2nd day (April 9 - April 12) 14:11:45 we can arrange your topics on the initial days, just mention the same against the topic 14:11:54 but that will be after i start planning for next PTG 14:12:37 next, Image Encryption 14:12:43 #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/ELMPPMIBGP4XW4YENHPZKQ57MJZVOAIE/ 14:12:56 this has been an ongoing effort for a long time 14:13:12 finally Josephine (and team) tested it with ceph backend 14:13:27 and wanted feedback on the patches and testing from cinder and glance teams 14:14:04 next, Interesting discussion on active/inactive projects 14:14:11 #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/YQILC7T246W2APPRAO24W3DWNUO4T6PU/ 14:14:56 this is unrelated to cinder but related to the discussion involved regarding an inactive project -- monasca here 14:15:42 if you are interested, you can add your opinions to the ML since i think Jay wanted feedback from the community, hence the mail 14:16:13 next, Caracal-2 Milestone 14:16:19 #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/6XXLHELC3NNW2AI3ILM2RIKTAAICDJP2/ 14:16:26 We don't observe any deadlines since we shifted the driver merge deadline to 26th Jan (because of year end break) 14:16:42 but i will come back to the driver discussion later as a topic 14:16:48 and that's all for announcements 14:16:55 anyone has anything else to announce? 14:17:10 reminder about the foundation election/bylaws changes 14:17:19 look for email with subject: "OpenInfra Foundation 2024 Individual Director Election and Bylaws Amendments" 14:17:26 contains an individualized link so you can vote 14:17:50 "Voting will close this Friday, January 12th, 2024 at 11:00am CST/1700 UTC. If you have any questions, please contact secretary@openinfra.dev" 14:18:21 (that's all) 14:20:11 thanks rosmaita , i think i forgot that since i was just looking at the ML archive for announcements 14:20:22 np 14:20:31 okay, let's move to topics 14:20:34 #topic Zuul Depends-On syntax 14:20:38 rosmaita, that's you 14:20:49 this is a quick one 14:21:34 as we all know, if you have a patch that depends on a change in a different code repo, you can add "Depends-on" in your commit message, and zuul will test your change with that other patch also applied 14:21:54 you see 2 ways of people doing this: 14:22:13 1 - Depends-on: Idf4d79527e1f03cec754ad708d069b2905b90d3f 14:22:25 2 - Depends-on: https://review.opendev.org/c/openstack/os-brick/+/895195 14:22:53 both work, but i was using #2 recently and was corrected by an infra team member that #1 is preferred 14:23:06 and this email explains why: 14:23:19 https://lists.openstack.org/pipermail/openstack-dev/2018-January/126535.html 14:23:42 just for awareness 14:23:45 (that's all) 14:24:10 i think you got it the wrong way around 14:24:10 except that i have that backwards! 14:24:16 :P 14:24:23 yes, the URL version is preferred 14:24:48 ok, so now that everyone is completely confused, i will shut up 14:25:22 I've been using the URLs for a long time, maybe someone corrected me long ago on this 14:25:30 but thanks for the awareness rosmaita 14:25:50 if that was all for this topic, we can move to the next one 14:25:53 #topic Regarding the dedicated backup status field discussed last week 14:25:56 that's all 14:26:01 crohmann, that's you 14:26:13 sorry i moved too fast 14:27:00 jep. Thanks for talking about this last week. I just wanted to know if you are interested in this change at all and if someone would implement this for the "D" release. 14:27:25 reading the description i think my comment from last week was misunderstood 14:27:43 what i meant was, no one reproposed it meaning no one is planning to work on it 14:28:11 So technically a reproposal means "I plan on working on this for the next cycle" ? 14:28:52 no, if the spec was reproposed for Caracal cycle, meaning they planned to work on it this cycle itself 14:29:02 But with a merged spec, is there any commitment to pick this up as a team, or does it always come down to someone indiviually starting the work? 14:29:07 if you are referring for the D cycle, we might see a resubmission next time 14:29:57 ok. So it's not about "me" having the intest in the feature to resubmit it, but rather the implementors doing so to communicate the intention to implement? 14:30:21 yes, correct 14:30:29 thanks. Understood. 14:30:29 at least that's my understanding 14:31:15 we have an "Untargeted" section on the specs homepage 14:31:30 https://specs.openstack.org/openstack/cinder-specs/#untargeted-specs 14:31:34 the implementation part depends on the team member's bandwidth which only an individual contributor can tell, if he/she can work on the implementation 14:32:15 i think crohmann's worry is that the spec will be lost and no one will know that it's available to work on 14:32:26 that rosmaita . 14:32:35 rosmaita, that might be a good place to put it 14:32:51 it might make sense to move the untargeted section to immediately after the current cycle's specs 14:32:54 rosmaita: i'll personally offer to host the fungi release party in late 2025 14:33:10 because right now, you have to scroll way down to see it 14:33:16 fungi: i'll be there! 14:34:03 rosmaita, makes sense to me, I can propose a patch for that unless you plan to work on it 14:34:07 I appreciate the spec not being "filed away" for good. But become part of some (stretched) roadmap 14:35:04 whoami-rajat: tell you what, i will propose the move-stuff-around patch, and you can look for other unimplemented specs to add to the list 14:35:20 since the previous cycle specs don't need work if implemented, the untargeted one seems more important 14:35:40 rosmaita, sure, sounds good 14:35:49 Thanks to you both for tackling this issue of unimplemented (but potentially interesting) specs. 14:36:00 whoami-rajat: i'll put up the patch, and then you can take it over to add the other specs 14:37:19 rosmaita, i can add the patch to move specs on top of your patch since i think 1) moving the untargeted section and 2) moving specs into it are 2 different tasks 14:37:30 ok, sounds good 14:38:11 thanks crohmann for the discussion, anything else on this topic? 14:38:17 not from e. 14:39:35 okay, moving on 14:39:44 #topic Regarding the Ceph caps 14:39:50 crohmann, again that's you 14:40:02 #link https://meetings.opendev.org/meetings/cinder/2024/cinder.2024-01-03-14.00.log.html#l-56 14:40:59 I kindly ask you to read my post on the ML at https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/5SVYDYOXWMG4TJKWEA6BFMPZZGC3Q5CS/ with my findings and suggestions to fix this. 14:41:46 Firstly I'd like to ask if you agree that there is some degree of disarray with the caps in the docs and code. 14:42:41 regarding a), there seems to be some issue with glance releasenote generation since the patch included an upgrade note mentioning this, it's strange it didn't show up in glance releasenotes 14:43:05 https://github.com/openstack/glance_store/commit/3d221ec529862d43ab303644e74ee9ad6ce8cd40#diff-e87ca98c41a1d5e6041634018a7b496e603acdd2489313156050f4d98920e159 14:44:54 Yes, but this does not mention the need to change the caps to grant glance read access to volumes. 14:45:35 oh sorry, it actually does 14:45:44 minimum ceph client version is greater than 'luminous' need to grant glance osd read access to the cinder and nova RBD pool. 14:47:11 glance team should look into this as to why it doesn't show up https://docs.openstack.org/releasenotes/glance_store/2023.2.html 14:48:22 anyways, we can go through the points after the meeting, thanks for bringing this up crohmann 14:49:17 Thanks. Shall we discuss on the ML then? I gladly write some patches if we could work out how things (caps) should look like today. 14:50:13 whoami-rajat: finally found it, it's in the yoga release notes for glance: https://docs.openstack.org/releasenotes/glance_store/yoga.html#relnotes-3-0-0-stable-yoga 14:50:24 but, obviously, not a lot of people have seen it! 14:50:53 And it's already obsolete 14:52:20 i don't know about the documentation point, does it need to be documented separately for each deployment tool? or in every project as to what are the right caps for it. Eg: cinder documentation can mention it's caps for glance and nova pools 14:52:23 That's my whole point. There are multiple set of caps floating around in code, deployment tooling and docs (and also potentially incomplete release notes). Least privilege and proper configuration to the requirements of Cinder and Glance as clients should be easily achievable. 14:53:24 No it does not need to be documented separately, thus my wording of "reference" caps (coming from the projects themselves). 14:53:29 crohmann: as part of your ML discussion, you can maybe ask where the best place to document this stuff is 14:53:46 rosmaita, oh it was merged in yoga 14:54:54 rosmaita: I did. See my last bullet points: 14:55:01 * fix the documentation and code devstack as "upstreams" first 14:55:08 distribute this as a reference to the deployment tools and also the 14:55:12 Ceph docs 14:55:38 * rosmaita is apparently not paying close attention 14:56:20 To me each project should be able to communicate the required access rights to e.g. Ceph, to function properly and document those. 14:56:34 Everything and everybody else (Ceph docs, deployment tooling, ...) just follows that. 14:57:25 I just made the observation that (as an operator) the current state of things is really confusing. 14:57:45 ok, so for cinder, i guess the appropriate place will be here: https://docs.openstack.org/cinder/latest/configuration/block-storage/drivers/ceph-rbd-volume-driver.html 14:57:50 makes sense, we can add to the cinder documentation regarding the caps cinder (volumes) pool need to nova and glance pools 14:58:14 +1 14:58:52 As soon as this happend, I gladly raise bugs with openstack-ansible, Ceph, ... about them being outdated :-) 14:59:10 sounds like you have a good plan! 14:59:38 we just have 2 minutes left so we can follow up on this discussion 14:59:46 there is one other important topic i wanted to cover 15:00:02 #topic Driver Merge Deadline (26th Jan) 15:00:10 #link https://etherpad.opendev.org/p/cinder-caracal-drivers 15:00:37 if you are planning to add your driver in the 2024.1 release, please mention the patch details on the etherpad 15:00:50 till now we have a blueprint from ZTE but we are unable to find any patches proposed against it 15:00:55 and we are out of time 15:01:02 will move the rest of topics to next week 15:01:10 rosmaita: whoami-rajat: Will you reply about the caps issue on the ML then? 15:01:24 crohmann, sure, i will take a look 15:01:35 awesome. Thanks a bunch! 15:01:40 np 15:01:50 we have a long list of review request to kindly look into it as well 15:01:53 thanks everyone for attending 15:01:55 #endmeeting