14:04:44 <jbernard> #startmeeting cinder 14:04:45 <opendevmeet> Meeting started Wed Oct 22 14:04:44 2025 UTC and is due to finish in 60 minutes. The chair is jbernard. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:04:45 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:04:45 <opendevmeet> The meeting name has been set to 'cinder' 14:04:52 <jbernard> courtesy ping: jungleboyj rosmaita smcginnis tosky whoami-rajat m5z e0ne geguileo eharney jbernard hemna fabiooliveira yuval tobias-urdin adiare happystacker dosaboy hillpd msaravan sp-bmilanov Luzi sfernand simondodsley zaubea nileshthathagar flelain wizardbit agalica 14:04:57 <jbernard> #topic roll call 14:05:01 <jbernard> #link https://etherpad.opendev.org/p/cinder-gazpacho-meetings 14:05:03 <agalica> o/ 14:05:04 <rosmaita> o/ 14:05:06 <hemna> mep 14:05:08 <hvlcchao1> o/ 14:05:10 <yuval> 0/ 14:05:12 <jbernard> o/ 14:05:16 <erlon> \o 14:06:19 <jungleboyj> o/ 14:06:29 <raineszm> o/ 14:06:31 <whoami-rajat> hi 14:06:40 <opendevreview> Eric Harney proposed openstack/cinder master: WIP: image_utils: Detect missing device before calling qemu-img convert https://review.opendev.org/c/openstack/cinder/+/964541 14:07:27 <raineszm> I hope I wasn't too presumptuous in adding an entry for today's meeting. 14:07:44 <jbernard> raineszm: not at all, that is what you're supposed to do 14:08:20 <raineszm> great. ty 14:08:25 <jbernard> welcome everyone 14:08:30 <jbernard> #topic annoucements 14:08:37 <jbernard> really quickly, next week is PTG 14:09:08 <jbernard> im working on finalizing the scheudle, but roughly we will have tuesday and wednesday for topics 14:09:34 <jbernard> thurs and fri can be used for collaborative reveiws, which will probably be more helpful than anything else 14:09:42 <jbernard> we have a few cross project sessions 14:10:06 <jbernard> there is an eventlet removal session on tuesday (across all projects) 14:10:19 <jbernard> wednesday we have a nova/cinder and a nova/cinder/ironic session 14:10:38 <jbernard> it looks like the nova/cinder/glance (encryption api) cross project will be on thursday 14:10:44 <jbernard> (waiting for final word from cyril) 14:11:33 <jbernard> most of the sessions will be concentrated between 1400 to 1700 UTC 14:12:02 <jbernard> im not sure if we want to try meetpad again (over google) but we can sort that out later 14:12:25 <jbernard> historicaly ive had trouble with the recordings, but that's most probably my fault 14:13:15 <jbernard> that's about all, i will send out a mail with more details and update the etherpad 14:13:20 <rosmaita> well, if the recordings are important, we should use whatever makes recording most reliable 14:13:36 <jbernard> that, i think, is why we've reverted to goolge in the past 14:13:55 <rosmaita> yep 14:13:57 <jbernard> it works, but it's different from what most others are doing 14:14:17 <jbernard> and can create some confusion, but i dont think it's signficant 14:14:18 <jayaanand> hi 14:14:19 <rosmaita> when was the last time we used meetpad? i don't remember 14:14:24 <erlon> If it would be nice if it was possible to get automated notes on those. Not sure if you guys have tried that before, but it's pretty handful the times I used 14:14:35 <jbernard> last ptg we tried initially 14:14:43 <jbernard> and failed and switch back to google 14:15:09 <jbernard> erlon: i think i can enable gemini to take notes from the live sessions 14:15:13 <erlon> Does meat pad supports automated notes? 14:15:16 <jbernard> erlon: is that what you mean? 14:15:35 <erlon> yeah, that's what I used on google meets 14:15:57 <jbernard> if we use google then i think it's straightforward, i need to look into meetpad if we try to go that route 14:17:18 <jbernard> that's all i have for annoucements, questions about next week? 14:17:28 <rosmaita> i must admit that i haven't been paying attention, is meetpad currently the preferred PTG meeting software? 14:17:37 <rosmaita> i notice that starlingx is using linux foundation zoom 14:17:43 <erlon> Are you organizing the schedule and planning for the ptg anywere (etherpad, docs)? 14:18:43 <erlon> Zoom also has meeting notes feature 14:18:44 <jbernard> we have a planning etherpad (https://etherpad.opendev.org/p/2026.1-ptg-cinder-planning) quite sparse at the moment, i will be transferring notes and shedule to the proper etherpad today 14:20:40 <jbernard> let me just say, if anything comes up just ask in the channel 14:20:59 <jbernard> most things are open to discussion and i welcome all input 14:21:33 <jbernard> #topic stable releases 14:21:43 <jbernard> rosmaita: just wanted to check in with you, how is that going? 14:21:49 <jbernard> do you need anythign from me? 14:22:37 <rosmaita> just need +1s on the patches :D 14:22:58 <jbernard> ok, will do 14:23:08 <rosmaita> so, at this point, no more merges into Caracal, that is, stable/2024.1 14:23:27 <jbernard> i think we landed everything we were targeting, no? 14:23:34 <rosmaita> yep 14:23:51 <rosmaita> there are no open non-WIP patches 14:24:16 <rosmaita> and i guess i should abandon all of them 14:25:08 <jbernard> yeah, else they just clutter the dashboard 14:25:08 <rosmaita> that's all from me 14:25:39 <jbernard> rosmaita: thanks for all of it! 14:25:43 <opendevreview> Merged openstack/cinder stable/2024.2: [Pure Storage] Manage Volume from GUI fails https://review.opendev.org/c/openstack/cinder/+/944301 14:25:44 <rosmaita> np 14:25:46 <opendevreview> Merged openstack/cinder stable/2024.2: [Pure Storage] Fix issue with LACP ports not being identified https://review.opendev.org/c/openstack/cinder/+/945785 14:26:35 <jayaanand> we have multiple back-porting patches from NetApp. Can someone take a look 14:26:40 <jayaanand> https://review.opendev.org/c/openstack/cinder/+/956755 14:26:40 <jayaanand> https://review.opendev.org/c/openstack/cinder/+/959293 14:27:04 <jayaanand> https://review.opendev.org/c/openstack/cinder/+/961431 14:27:04 <jayaanand> https://review.opendev.org/c/openstack/cinder/+/961432 14:27:18 <jbernard> #topic rbd driver blocksize 14:27:26 <jbernard> #link https://review.opendev.org/c/openstack/cinder/+/658283 14:27:40 <jbernard> raineszm: that's you i believe 14:27:50 <raineszm> So this is an /old/ patch that stalled out but is probably worth reviving. 14:28:03 <jbernard> jayaanand: if they're just review requests, put them in the etherpad 14:28:18 <raineszm> The basic question that seem to have come up that stalled the review was that there was a question about where to add the new options. 14:28:22 <jayaanand> sure than you! 14:28:40 <jbernard> raineszm: i vaguely remember - why can't we make it an option at the driver level instead of just rbd? 14:28:55 <raineszm> That seems like a good idea long term. 14:29:08 <raineszm> The only concern is how to handle the other drivers that already have an analogous option 14:29:22 <jbernard> hmm, do you have an example? 14:29:23 <raineszm> So there's combination of refactor and new feature there 14:29:48 <raineszm> yes 1 sec 14:29:49 <hemna> +A'd 14:31:32 <erlon> How common is this option across the drivers? Is this a libvirt feature that can be used by any backend? 14:31:48 <raineszm> Yes. So any driver can set the metadata 14:31:54 <raineszm> and libvirt picks it up 14:32:22 <raineszm> the solidfire driver has implemented it in the past I think 14:32:27 <raineszm> just trying to find a permalink 14:32:57 <erlon> Are the connectors (nfs, iscsi, fc) already passing that to libvirt on the nova side? 14:34:23 <raineszm> It seems so. So basically the logical_block_size and physical_block_size have default values that are advertised 14:34:27 <raineszm> in the connection data 14:34:46 <raineszm> the addition here is to allow setting those as an option 14:35:37 <raineszm> At least one other driver sets those by querying the disk geometry 14:36:08 <raineszm> Currently in the icsci driver it seems 14:36:40 <raineszm> https://opendev.org/openstack/cinder/src/commit/d02171164bdd702b12b59888b744d172f30d712d/cinder/volume/targets/iscsi.py#L127 14:38:09 <erlon> right, I see that in nova. I think my concern would be mostly related to the ability to backport this fix if we increase its scope 14:38:14 <raineszm> and here https://opendev.org/openstack/cinder/src/commit/d02171164bdd702b12b59888b744d172f30d712d/cinder/volume/driver.py#L2782 in the driver base 14:39:13 <erlon> What I like that idea to have this re-used across all drivers 14:40:35 <raineszm> I agree. I think that long term it should be moved to the driver level 14:40:57 <raineszm> The question is whether to do that combined with the fix for rbd or to do it in two steps 14:41:21 <raineszm> The motivation for the original patch being that it allows to fix performance issues with rbd and windows vms 14:42:43 <raineszm> jbernard: what do you think? 14:42:44 <erlon> yes, and I like the way the options are put in pairs like 512e -> "4096 512", it seems more meaningful too the user. 14:43:00 <jbernard> i lean towards doing it once, even if it takes a bit more (at the driver layer) 14:44:17 <erlon> How about backporting? It will be bigger 14:44:53 <jbernard> that is true, but it's not impossible, especially if care is taken in writing and review 14:45:07 <erlon> But apparently not impossible. 14:45:09 <erlon> :) 14:45:14 <erlon> yeah, agreed 14:45:19 <raineszm> Haha. Okay. Fair. 14:45:32 <raineszm> Yeah I'm game to hoist it up to the driver level. 14:45:43 <raineszm> If that's what we want to do 14:45:43 <jbernard> my sense is that faster fixes that require followup, more often than not, never get said followup 14:45:43 <erlon> If we can make it back to Caracal, I think I can make it happen 14:46:38 <raineszm> I'll start taking a look at that then. 14:46:50 <raineszm> thanks for chatting in through. 14:46:53 <raineszm> *it 14:47:04 <jbernard> np, thanks for working on it 14:47:28 <erlon> lets discuss more details on the option naming in the ptg 14:47:32 <jbernard> sure 14:48:01 <jbernard> #topic open discussion 14:48:16 <erlon> \o 14:48:34 <tobias-urdin> can i shamelessly plug my (now old) open patches :) https://review.opendev.org/q/project:openstack/cinder+status:open+owner:%22Tobias+Urdin%22 14:48:37 <agalica> jbernard: would it be possible to get cores assigned to our issues? We have 3 issues we need to backport - targeting for december. These are all important. One of them already has a core and another reviewer (and we're waiting on Rajat again on that one) 14:48:49 <erlon> just a nudge on whoami-rajat to get a look on the brick iscsi patch 14:49:11 <jbernard> tobias-urdin: you may :) 14:49:24 <agalica> tobias-urdin: do you need reviewers: My team can review if you do 14:49:26 <agalica> non-core 14:49:36 <agalica> My patches are these: Replication: https://review.opendev.org/c/openstack/cinder/+/869991 14:49:36 <agalica> Snapshot Expansion: https://review.opendev.org/c/openstack/cinder/+/850830 14:49:36 <agalica> B20: https://review.opendev.org/c/openstack/cinder/+/959901 14:50:19 <agalica> whoami-rajat: Polite nudge here as well that we have updated our B20 patch with the changes that hopefully satisfy your comments 14:50:53 <jbernard> agalica: cores assigned, let me think on that. it's not been our practice in the past but maybe we can do something to get closer and improve turnaround 14:51:32 <jbernard> it's always voluntary of course, as this is open source, but maybe i can help coordinate better 14:51:39 <agalica> jbernard: thanks. we're really struggling here, and no one seems to want to review. we're already behind in that we were meant to get merged into 2025.1 and missed that. Things are getting hot. 14:51:55 <tobias-urdin> thanks! they are quite old now but i think i rebased them recently, any feedback welcome :) 14:51:59 <agalica> we are happy to review many patches if tehre are 14:52:07 <agalica> ok tobias - any one in particular, or all of them? 14:52:18 <jbernard> agalica: ack 14:52:49 <erlon> @agalica the volume extend patch has a merge conflict. Can you rebase it? 14:53:00 <agalica> erlon: yes, we will do that today 14:53:06 <erlon> Volume Replication too 14:53:13 <agalica> ok, will do that today 14:53:44 <agalica> tobias-urden: do you have any specific patches you want reviewed in that list, or all of them? 14:54:55 <jbernard> agalica: i would go through them and review the ones that are passing ci without merge conflict 14:55:33 <agalica> ok, thanks - will do that 14:56:06 <wizardbit> Hey jbernard, were you able to get an update from Rajat on https://review.opendev.org/c/openstack/os-brick/+/955379 ? 14:56:46 <jbernard> whoami-rajat: ^ 14:59:01 <wizardbit> That will ping him? 14:59:54 <jbernard> maybe ;) 15:00:31 <jbernard> it's quite late now, he may see it in his scrollback buffer 15:00:59 <jbernard> we should be able to address all of these in next week's meetings thoug 15:01:18 <jbernard> we're at time, anything else? 15:01:20 <Anoop_Shukla> https://review.opendev.org/c/openstack/cinder/+/956221 15:01:44 <jbernard> Anoop_Shukla: ack 15:01:53 <jbernard> ok, thank you everyone, see you next week 15:01:55 <jbernard> #endmeeting