Wednesday, 2026-08-12

*** erlon9 is now known as erlon03:07
opendevreviewjayaanand borra proposed openstack/cinder-tempest-plugin master: Add image-cache cross-pool tempest tests  https://review.opendev.org/c/openstack/cinder-tempest-plugin/+/98998703:10
opendevreviewAdam Harwell proposed openstack/cinder master: image_utils: use concurrency-aware tpool wrapping  https://review.opendev.org/c/openstack/cinder/+/99500106:01
rm_workok, that cleaned up a lot with the changes that were incorporated into the parent, should be more minimal/focused now06:07
*** erlon5 is now known as erlon09:34
opendevreviewPrudhvi proposed openstack/cinder master: Implement per-pool image-volume cache promotion  https://review.opendev.org/c/openstack/cinder/+/98831509:38
*** auniyal2 is now known as auniyal_110:11
*** auniyal_1 is now known as auniyal__10:11
opendevreviewRajat Dhasmana proposed openstack/cinder-specs master: Propose replication v3 spec  https://review.opendev.org/c/openstack/cinder-specs/+/98595210:43
opendevreviewjayaanand borra proposed openstack/cinder master: Implement per-pool image-volume cache promotion  https://review.opendev.org/c/openstack/cinder/+/98831510:45
opendevreviewThomas Goirand proposed openstack/cinder master: Use sample_max to make cinder reproducible  https://review.opendev.org/c/openstack/cinder/+/99630811:33
opendevreviewThomas Goirand proposed openstack/cinder master: Use sample_max to make cinder reproducible  https://review.opendev.org/c/openstack/cinder/+/99630812:36
opendevreviewLeo Jenkins proposed openstack/cinder-specs master: Support for Hostname-based Ceph Mon References  https://review.opendev.org/c/openstack/cinder-specs/+/100072113:15
jbernard#startmeeting cinder14:00
opendevmeetMeeting started Wed Aug 12 14:00:36 2026 UTC and is due to finish in 60 minutes.  The chair is jbernard. Information about MeetBot at http://wiki.debian.org/MeetBot.14:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:00
opendevmeetThe meeting name has been set to 'cinder'14:00
yuvalis there a meeting?14:00
jbernard#link https://etherpad.opendev.org/p/cinder-hibiscus-meetings14:00
yuval0/14:00
jbernardjungleboyj 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 lutimura kaisers aloke_dev Anoop_Shukla erlon Jyotsna: courtesy ping (cinder meeting)14:00
jbernard#topic roll call14:01
jbernardo/14:01
simondodsleyo/14:01
agalicao/14:01
Jyotsnao/14:01
Luzio/14:01
Anoop_ShuklaO/14:01
jbernardhey everyone, im alive :) i managed to get quick sick a couple of weeks ago, feeling mostly better14:01
kaiserso/14:02
rosmaitao/14:02
fqio/14:03
mheno/14:03
jbernard#topic annoucements14:04
jbernardquickly, periodic tasks are all green14:04
rosmaita\o/14:04
jbernardwe need (probably) to have a second midcycle14:05
Leo[m]o/14:05
jbernardand im thinking we should shoot for soon if possible14:05
jbernardthe earliest would be next week14:05
jbernardfor reference, this is a single day session lasing 2-3 hours14:06
jbernardis anyone opposed to holding it next friday?14:07
jbernardis a different day or week preferred?14:07
rosmaitathat conflicts with the festival of reviews14:07
erlon\o14:07
simondodsleyi'm open to that - replace the festival with it14:08
jbernardahh, right.  we could overwrite the festival, or push to the following week14:08
rosmaitaanother possibility would be to hold it on wednesday, 1400-160014:08
rosmaitaeveryone already has 1400-1500 blocked off14:08
jbernardiow overwrite out weekly meeting next week14:08
jbernardthat works too, i would vote for that over friday personally14:09
simondodsleyeither14:09
erlonboth work for me , but Friday is better 14:09
yuvalwedenesday is better 14:09
yuval:)14:09
rosmaitai'd prefer wednesday, we are rapidly approaching M-314:10
yuvalwhats M-3 stand for?14:10
yuvalmilestone?14:10
jbernardyep14:11
yuvalgot it cause 28 is freeze right?14:11
jbernardok, lets plan for wednesday next week, in place of our regular meeting14:12
jbernardi will send mail14:12
agalicathe 28th of this month is freeze?14:12
jbernardthat will give me some time to try to catch up on some of my backlog14:12
jbernardagalica: https://releases.openstack.org/hibiscus/schedule.html feature freeze14:14
agalicagotcha, thx14:14
agalicaoof, none of my patches are really getting traction.  thought there was more time14:14
rosmaitathere is never enough time!14:15
jbernard#topic CBT support series (simon)14:17
simondodsleythanks - this is just a heads up to the vendors that we are creating CBT support 14:18
simondodsleythere is a series of patches listed in the reviews etherpad if people want to build their own support off those14:18
JyotsnaYes we do have CNT in our roadmap for new driver, perhaps the driver get merge on time.14:19
Jyotsna*CBT14:19
simondodsleytheree is also a spec which has been used to create the patch series: https://review.opendev.org/c/openstack/cinder-specs/+/98508914:19
yuvalare there driver that already implemented CBT?14:19
simondodsleythe spec hasn't been approved yet, but it shouldbe good14:20
simondodsleyonly Ceph supports CBT today14:20
Anoop_ShuklaNetApp plans to implement CBT in the next release.14:20
yuvalk14:20
simondodsleythis is for any other external system that wants to implement14:20
JyotsnaThanks simon for sharing 14:20
simondodsleythere is full usage documentation in the patch series as well14:21
simondodsleyincluding LVM support14:21
Anoop_ShuklaQuick question: CBT will be applicable for all protocols yes? Including NFS?14:21
simondodsleyit was designed so 3rd party backup vendors can use it as well14:21
jbernardnice, we can use some midcycle time to nail that down14:21
jbernardAnoop_Shukla: it should be, if you find an issue please raise it14:22
simondodsleyAnoop_Shulka: that will be down to the vendors implementation14:22
Anoop_ShuklaRight..we are currently analysing the support from our platform. That question came up during that analysis..so just making sure to clarify14:23
jbernard#topic NVMe-FC support series (simon)14:24
simondodsleySame again - vendor heads up - this is adding full NVMe-FC support to cinder and os-brick, including support for FCZM. There is LVM support as well, but Ubuntu doesn't ship nvme_fcloop so we can create tempest tests just yet14:24
Anoop_ShuklaAnother question was, are drivers allowed to have a binary dependency or a client dependency on the driver code? Or the binary client cannot be shipped via the openstack code?14:24
simondodsleythe centos-9-stream nodeset is flakey so we can't get that to work currently, which would be the only place we could do the tempest tests14:25
simondodsleyAnoop_Shukla: I don't beleive binaries can be shipped with openstack....14:26
jbernardAnoop_Shukla: dependencies can be defined from what setuptools allows (requirements.txt), but things external to pypi are not available14:26
jbernardAnoop_Shukla: basically, no ;)14:26
jbernardAnoop_Shukla: those things (could) get pulled in a packaging time14:27
Anoop_ShuklaOkay14:27
Anoop_ShuklaGot it14:27
yuvalappriciate for the heads up - what actually I need to validate/check?14:27
simondodsleyyou will need your vendor CI to run on that protocol and report back14:27
yuvalif my backend does not use the NVME-FC still?14:28
simondodsleythis add won't affect anyother nvme protocols, roce v2 or tcp14:29
yuvalok thanks14:29
jbernardthanks for working on those simon14:30
jbernard#topic drivers security stance (rosmaita)14:30
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/99672214:30
rosmaitai'd like to get some action on this patch14:31
Anoop_Shuklahttps://review.opendev.org/c/openstack/cinder/+/99931014:31
opendevreviewMerged openstack/cinder master: Imported Translations from Zanata  https://review.opendev.org/c/openstack/cinder/+/100023814:31
rosmaitait's a little vague about what exactly we will be requiring14:31
rosmaitabut i think that's ok, because the key point is to get the timeline set14:31
Anoop_ShuklaNetApp patch to improve Cinder driver security stance - https://review.opendev.org/c/openstack/cinder/+/99931014:31
rosmaitafor when we will start kicking out/marking unsupported drivers14:32
rosmaitaAnoop_Shukla: very nice!14:32
Anoop_ShuklaDefault to https and cert verify default to true for improving the security stance14:32
rosmaitasimon has a big patch up that we can use to discuss details14:32
Anoop_ShuklaNeed some eyes to the patch as we target to merge it in this release14:33
rosmaitahttps://review.opendev.org/c/openstack/cinder/+/99655414:33
rosmaitabut my patch just says that we will start enforcing this with the Indri release, that is, 2027.114:34
simondodsleyAnoop_Shukla: I'll look at it today14:35
jbernardrosmaita: does this count as a sufficient initial step related to what the foundation is hoping/asking for currently?14:35
rosmaitai think so ... the foundation is worried that bugs will be filed but not implemented14:36
simondodsleyAnoop_Shukla - you should recheck to pick up the fix that reolves the openstacksdk-functional-devstack gate failure14:36
rosmaitaso this gives us an enforcement mechanism14:37
Anoop_ShuklaThanks Simon..will pick up the change. Do you have the patch handy?14:37
simondodsleyit's merged - so a recheck will work now14:37
rosmaitaAnoop_Shukla: you may need to rebase on master14:38
jbernardrosmaita: excellent, thanks14:38
simondodsleytrue14:38
Anoop_Shukla@rosmaita: I think Manila is taking a stance of security in H release..14:38
rosmaitawell, since it would have to be done in the next week or so, i though H was too early14:38
rosmaitabut, we could say H14:39
Anoop_ShuklaHmm..sooner the better :)14:39
rosmaitaand then mark all the drivers (except maybe netapp, if your patch is merged) as unsupported14:39
simondodsleythe doc pathc i created will have to be manually changed whenever a driver changes its status - there isn't an automatic update14:39
Anoop_Shuklahaha14:39
rosmaitawell, we aren't saying that nothing needs to happen until Indri (from the perspective of driver development)14:40
simondodsleyand my driver-that is fine too rosmaita:14:40
simondodsleylol14:40
rosmaitaok, netapp and everpure14:40
rosmaitaanyway, please leave comments on the patch, and we should discuss simon's patch at the midcycle next week14:42
yuvalopen discussion?14:43
rosmaitai think mhen has something first?14:43
mhenyes, thanks14:43
mhen(oh and sorry if putting this under "topics" directly was bad etiquette - should I have put it under open discussion?)14:43
mhenso, this is in the context of the ongoing image encryption contribution14:44
rosmaitano, it's a topic, i think14:44
mhenwe were asked to address a resize function which is primarily used in remotefs/nfs based back14:44
mhenthen we found volume encryption (which we need) not working with nfs right now and filed a launchpad bug14:44
mhenthe potential fix from eharney linked in launchpad introduces behavior in Cinder that we see as problematic14:44
mhen#link: https://review.opendev.org/c/openstack/cinder/+/973870/comment/d923c431_09cca9b7/14:44
mhenit breaks usage in Nova and `image create --volume` as it introduces qcow2-based encryption to Cinder, which is a behavior other code parts don't expect from Cinder14:44
mhen(also not known to os-brick)14:44
mhenso, one question would be if Cinder really would want to introduce new encryption mechanisms just to fix nfs? (with all the impact on os-brick, Nova etc.)14:45
eharneyi haven't gone over the latest comment on 973870.  but nfs volume qcow2 encryption is already present in Cinder14:45
mhenand more broadly: could we move forward with image encryption without supporting nfs/remotefs-based backends for the time being while this is being addressed?14:45
mheneharney: it is? then I haven't looked close enough I guess14:46
mhenanyway, as written in the comment it breaks a few things14:46
mhenshouldn't that be known to os-brick then as well?14:46
mhenor am I misunderstanding things?14:46
eharneythere are some holes in it (especially around handling volume<->image etc) but you can create encrypted volumes with that method and attach them in nova currently14:48
Anoop_ShuklaSo the encrypted image now gets decrypted before raw conversion after this patch? Just trying to understand the flow.14:49
mhennormally, our image encryption implementation assumes Cinder is using raw LUKS in every place, so we make sure that the image is converted to that encryptiono14:50
mhenif the image is already raw LUKS encrypted, it is used as-is14:50
mhenotherwise it is re-encrypted basically14:50
Anoop_Shuklaokay14:50
mhenwe do require an encrypted target volume though14:50
Anoop_Shuklaright14:51
mhenhowever, we never expect Cinder to produce encrypted qcow2 currently14:51
mhenanyway, eharney if you have any pointers to existing encrypted qcow2 usage in Cinder feel free to comment on my comment and I'll look into it14:52
mhenwe don't need to go into detail here (looking at the time)14:52
mhenI just wanted to know if it would be feasible to skip those backends for now in the image encryption contribution and add support later14:53
rosmaitamhen: let's definitely follow up at the midcycle next week14:53
mhenok14:53
jbernard#topic open discussion14:55
yuvalopen discussion?14:56
JyotsnaHPE Alletra MP new driver patch  (Recieved +2 and workflow +1 and ready for merge) https://review.opendev.org/c/openstack/cinder/+/969373 Requirement Patch: https://review.opendev.org/c/openstack/requirements/+/974322  requirement patch - Addressed the comments from zigo and released flowkit version + update patch with latest version. Require appoval from reviewers. As we are approaching towards feature freeze very less time left for new driver14:56
yuvalI want to ask the core reviewers - what will help you for review? since I think there are a lot of patches going on14:57
jbernardLeo[m]: thanks for the hostname-based ceph patch, that's been on the list for a looong time14:57
jbernardyuval: ill be working a bit extra this week/weekend to try to catch up on what i can, ping me later this week, i should be able to make at least one pass14:58
agalicayuval: having solid reviews ahead of time on patches definitely helps.  if there's less issues to find it's easier to review14:59
agalicaspeaking from a "community should be reviewing as well" perspective14:59
Anoop_ShuklaMay be we can sign up for patches voluntarily to offload..15:00
simondodsleyMake sure all patches are in the review etherpad, or we don't know to look at them15:00
rosmaitaJyotsna: i will check on the requirements patch later, and then prod some people once zuul is +1 (i only have +1 in that repo)15:00
Jyotsnasure thanks <rosmaita>  i have ran recheck for zuul to pass so, if it passes we will need attention 15:02
simondodsleyand also make sure the patches go in the correct section of the etherpad15:02
jbernardok, last call15:04
Jyotsnacan we get attention from  Sean McGinnis and Jens for requirement patch as they have +2 rights 15:05
jbernardjust to recap, we review session this friday and we will have our midcycle next wednesday in this time slot (plus an hour or so extended)15:05
JyotsnaCan core reviewers help here in getting there attention after +1 15:05
agalicareview session is next friday, no?15:06
jbernardyes, sorry, next friday!15:06
jbernardlots of facetime next week15:06
agalicagoogletime, not facetime15:07
jbernardyup, in the general sense15:08
agalicasorry, that was a lame apple vs google joke15:08
agalicaI should probably be destroyed15:08
agalicaget that on the record15:09
jbernarddone!15:09
jbernardok, thanks everybody :)15:09
jbernard#endmeeting15:09
opendevmeetMeeting ended Wed Aug 12 15:09:20 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:09
opendevmeetMinutes:        https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-08-12-14.00.html15:09
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-08-12-14.00.txt15:09
opendevmeetLog:            https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-08-12-14.00.log.html15:09
agalicabed time for me I guess, night all15:09
agalicalol15:09
yuvalthanks!15:09
Jyotsnathanks15:10
zigoJyotsna: Added my +1, thanks for considering my remarks, much appreciated.15:13
Jyotsnathanks a lot zigo15:13
Jyotsna<rosmaita> are you around? we got zuul verified on patch and +1 from zigo can you please also provide +115:15
Jyotsnaalso can you request +2 reviewers for some attention 15:15
Jyotsna<zigo> can you please provide +1 on patch?15:16
Jyotsnasorry i am not able to see +1 from you 15:17
rosmaitaJyotsna: looking15:17
zigoJyotsna: When running tests, I still have hpe_storage_flowkit_py/tests/v3/test_session.py and hpe_storage_flowkit_py/tests/v3/test_task.py that are failing because they are trying to use the "src" folder.15:17
Jyotsnait is addressed in init_py as an alias is added 15:18
Jyotsnaso even if they are using src in import it will not fail 15:18
zigoJyotsna: That's what I'm getting when building the Debian package:15:21
zigohttps://paste.opendev.org/show/bKLMOTq0rW4MewtFq0tK/15:21
zigoSo, the issue seems to be mocking, not import.15:21
zigoI can write a debian specific patch for it though...15:21
Jyotsnaok i will check this 15:22
zigoJyotsna: I'm actually trying to fix it ! :)15:23
zigoHum... not sure how, actually.15:28
*** dviroel is now known as dviroel_lunch15:53
opendevreviewSimon Dodsley proposed openstack/cinder master: Add changed-block tracking contract and os-list_changed_blocks API  https://review.opendev.org/c/openstack/cinder/+/99092316:04
opendevreviewSimon Dodsley proposed openstack/cinder master: Consume Changed Block Tracking in cinder-backup  https://review.opendev.org/c/openstack/cinder/+/100074716:04
opendevreviewSimon Dodsley proposed openstack/cinder-tempest-plugin master: Add Changed Block Tracking (CBT) scenario tests  https://review.opendev.org/c/openstack/cinder-tempest-plugin/+/99574116:06
*** dviroel_lunch is now known as dviroel17:17
opendevreviewAnoop Kumar Shukla proposed openstack/cinder master: NetApp: enrich periodic Cinder EMS payloads  https://review.opendev.org/c/openstack/cinder/+/99620517:19

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!