14:03:36 <jbernard> #startmeeting cinder 14:03:36 <opendevmeet> Meeting started Wed Aug 21 14:03:36 2024 UTC and is due to finish in 60 minutes. The chair is jbernard. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:03:36 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:03:36 <opendevmeet> The meeting name has been set to 'cinder' 14:03:40 <jbernard> #topic roll call 14:03:43 <whoami-rajat> hi 14:03:45 <simondodsley> o/ 14:03:45 <Luzi> o/ 14:03:46 <jbernard> o/ hello everyeone 14:03:47 <rosmaita> o/ 14:03:50 <akawai> o/ 14:03:51 <eharney> hi 14:03:53 <msaravan> Hi 14:04:04 <mhen> o/ 14:04:04 <jbernard> #link https://etherpad.opendev.org/p/cinder-dalmatian-meetings 14:04:24 <jbernard> #link https://releases.openstack.org/dalmatian/schedule.html 14:04:32 <yuval> 0/ 14:05:58 <yuval> whats the meaning: "roll call" 14:06:29 <Sai> o/ 14:06:30 <whoami-rajat> i think it means attendance 14:08:33 <jbernard> yuval: yes, just who's here 14:08:35 <rosmaita> yes, traditionally the names of members of an organization are kept on a "roll", and the "roll call" is just asking each member to identify as present 14:09:05 <simondodsley> ah - the good old days 14:09:30 <jbernard> alright, welcome everyone 14:09:41 <jbernard> thanks for filling out the adgenda 14:09:50 <jbernard> before we get started, 14:10:18 <jbernard> the cycle is winding down, and reviews are one of my highest priority right now 14:10:18 <yuval> the roll is virtual 14:10:25 <yuval> is it powered by nova? 14:10:40 <jbernard> im aware of the dell powerstore and lightbits review series 14:10:42 <rosmaita> probably!!! :D 14:11:06 <jbernard> there are a few others, i just wanted to say that I/We are aware and working on it :) 14:11:31 <yuval> Thank you jbernard more mentioning - yes we do need your review - I am available for anyone that need my review 14:12:24 <jbernard> simondodsley: your summary for the dell powerstore series (specifically the NVMe work) will be helpful, I'll ask about that in a bit 14:12:31 <jbernard> ok 14:12:33 <whoami-rajat> I haven't been active in reviews recently but can help if we have a list of priority items somewhere 14:12:58 <yuval> I will also review nvme related 14:13:12 <jbernard> whoami-rajat: ill make one, the requests in today's adgena are a starting point 14:13:24 <jbernard> whoami-rajat: thanks 14:13:30 <whoami-rajat> okay, will take a look, thanks 14:13:36 <jbernard> #topic dalmatian os-brick release 14:13:57 <jbernard> im not sure who added this, rosmaita maybe? 14:14:07 <rosmaita> yeah, forgot to put my nick 14:14:25 <jbernard> tl;dr we need a few patches merged asap for the upcoming brick release 14:15:19 <rosmaita> basically, release is supposed to happen tomorrow 14:15:36 <rosmaita> though we can sometimes stretch it by a day 14:15:37 <jbernard> yes, libraries is tomorrow, clients are next week I believe 14:16:38 <rosmaita> basically, what we need to do is have people commit to reviews right now if we are going to get them into the release 14:17:59 <jbernard> rosmaita: does updating requirements fix anything, or is it just something we do before a release? 14:18:12 <jbernard> rosmaita: wondering if/how to prioritize 14:18:57 <rosmaita> hard to say ... what happens is that we are testing in the gate with whatever upper constraints are set to, which may be quite different from what's in requirements.txt 14:19:30 <eharney> i think we only have one open brick requirements change? 14:19:33 <rosmaita> it's probably not a big deal unless we are relying on a feature of a req that's not present in the minimum 14:22:07 <rosmaita> so jbernard, basically up to you 14:23:20 <jbernard> ok, if we can get the patches that need landing reviewed, then I'm happy to submit a patch for requirements 14:23:46 <jbernard> we can circle back post-lunch and see where we're at 14:24:13 <jbernard> how does that sound? 14:24:51 <rosmaita> so how do people feel about the image encryption password handling patch: https://review.opendev.org/c/openstack/os-brick/+/926293 14:25:04 <rosmaita> i am willing to look at that right after this meeting 14:26:10 <jbernard> i can too 14:26:16 <rosmaita> ok, cool 14:26:55 <mhen> nice :) 14:26:58 <rosmaita> the other one i identified is a small change in a privsep function, but will require a careful look: 14:27:05 <rosmaita> https://review.opendev.org/c/openstack/os-brick/+/922298 14:27:16 <rosmaita> but we could always hold off and backport it as a bugfix 14:27:47 <rosmaita> i didn't see anything else that looks key to the release 14:27:58 <rosmaita> so if anyone has a favorite os-brick patch, speak up now! 14:28:39 <whoami-rajat> I'm taking a look at that change 14:29:08 <Nilesh_Thathagar> Hi can we get review this os-brick patch https://review.opendev.org/c/openstack/os-brick/+/916062 ? 14:30:10 <rosmaita> it's isolated to a single connector, so kind of low risk ... i am willing to look at it for the release 14:30:26 <jbernard> me too 14:30:33 <jbernard> i added it to the list 14:30:49 <yuval> me too 14:30:54 <Nilesh_Thathagar> thanks 14:32:30 <rosmaita> ok, so eharney please respond to whoami-rajat's comment on https://review.opendev.org/c/openstack/os-brick/+/916737 and we can get that one moving along 14:32:36 <jbernard> simondodsley: re os-brick patches, 14:32:44 <simondodsley> yes 14:33:02 <jbernard> simondodsley: dell have a couple, i belive you've looked at 14:33:09 <jbernard> let me get the urls, one sec 14:33:18 <jbernard> https://review.opendev.org/c/openstack/cinder/+/907133 14:33:27 <jbernard> and https://review.opendev.org/c/openstack/os-brick/+/923345 14:33:31 <jbernard> i believe 14:33:39 <simondodsley> yes - we discussed this last week on the call. Neither Gorka or I are happy with the lack of FCZM support 14:34:24 <jbernard> simondodsley: ok, i just wanted to make sure that's still the case, I believe the cinder patches rely on the brick patch 14:34:26 <simondodsley> I actually have an NVMe-FC test system up right now and am going to look at the os-brick side of it 14:35:14 <jbernard> i might be wrong about that though, Yian was asking earlier and I want to make sure I make a accurate statement 14:35:31 <simondodsley> yes - its a pre-req 14:35:36 <jbernard> simondodsley: that's good news \o/ 14:35:45 <jbernard> simondodsley: ok, thank you! 14:36:44 <jbernard> whoami-rajat: do we have a url naming standard for review priority etherpads? 14:37:45 <jbernard> whoami-rajat: i think getting the list out of the meeting minutes will help folks to focus, certainly myself 14:38:42 <jbernard> #topic review requests 14:38:56 <jbernard> as i mentioned, this is the current priority 14:39:06 <jbernard> ill make an isolated etherpad 14:39:07 <whoami-rajat> jbernard, sorry got distracted with reviews, which milestone are we referring to? I think i created one for os-brick, another for M1/M2 etc 14:39:26 <jbernard> whoami-rajat: im just wondering if we have a convention, before I just make one up :) 14:39:50 <whoami-rajat> oh i don't think we have one, they are temporary and just useful until the milestone 14:40:01 <whoami-rajat> so anything should be good 14:40:42 <jbernard> ok 14:40:45 <jbernard> #link https://etherpad.opendev.org/p/cinder-dalmatian-reviews 14:40:51 <jbernard> ^ that will be it 14:40:59 <jbernard> ill move our adgenda requests there 14:41:11 <jbernard> plus a few others on my dashboard 14:41:35 <yuval> can we add the etherpad to the cinder header? 14:41:45 <yuval> The Block Storage Project | https://wiki.openstack.org/wiki/Cinder | https://tiny.cc/CinderPriorities 14:42:13 <jbernard> yuval: that's a good question, i think a channel operator must do this 14:42:29 <jbernard> yuval: and im am not one of those, to the best of my knowledge 14:42:33 <jbernard> does anyone know? 14:42:57 <rosmaita> i think we will have to ask the openstack-infra team 14:43:22 <yuval> the tiny.cc goes to: 14:43:24 <yuval> foreach=%28project%3Aopenstack%2Fcinder+OR%0Aproject%3Aopenstack%2Fpython-cinderclient+OR%0Aproject%3Aopenstack%2Fpython-brick-cinderclient-ext+OR%0Aproject%3Aopenstack%2Fos-brick+OR%0Aproject%3Aopenstack%2Fcinderlib+OR%0Aproject%3Aopenstack%2Frbd-iscsi-client+OR%0Aopenstack%2Fcinder-tempest-plugin%29+status%3Aopen&title=Cinder+Priorities+Dashboard&High+Priority+Changes=label%3AReview-Priority%3D2&Priority+Changes=labe 14:43:24 <yuval> l%3AReview-Priority%3D1&Blocked+Reviews=label%3AReview-Priority%3D-1 14:43:40 <yuval> https://review.opendev.org/dashboard/? 14:44:01 <simondodsley> If you add the meeting URL to the header it will have to be changed every cycle 14:44:25 <yuval> yes - well if its a big deal leave it 14:45:18 <jbernard> #topic open discussion 14:46:01 <jbernard> reviews aside, any issues or things that need to be raised, please do so 14:48:18 <yuval> I have something to say but not related to current release 14:48:24 <jbernard> shoot 14:48:50 <yuval> in os-brick, file: os-brick/os_brick/privileged/__init__.py 14:48:55 <yuval> have these 2 lines: 14:48:55 <yuval> os-brick 14:48:56 <yuval> os_brick 14:48:56 <yuval> privileged 14:48:56 <yuval> __init__.py 14:49:05 <yuval> if os.environ.get('VIRTUAL_ENV'): 14:49:06 <yuval> capabilities.append(c.CAP_DAC_READ_SEARCH) 14:49:23 <yuval> I am working with kolla and its making my life hell 14:49:52 <simondodsley> lol - kolla can do that 14:50:52 <yuval> I have to change the privsep section in the kolla-ansible files for cinder and nova to add "sudo -E..." and add the "VIRTUAL_ENV" variable 14:50:59 <yuval> to all dockers 14:51:46 <eharney> it seems weird to allow DAC_READ_SEARCH in that context but not others, i wonder why we can't just generalize it 14:52:11 <yuval> adding this capability regardless to the fact that I am using a virtual env or not - is a big risk? 14:52:13 <eharney> we already give SYS_ADMIN so there's probably a good argument for just saying that we need DAC_READ_SEARCH for things to work 14:52:59 <rosmaita> check the blame on that ... i seem to remember it being added specifically for virtual envs, maybe the commit message explains 14:53:24 <yuval> there is acomment above: 14:53:24 <yuval> # On virtual environments libraries are not owned by the Daemon user (root), so 14:53:25 <yuval> # the Daemon needs the capability to bypass file read permission checks in 14:53:25 <yuval> # order to dynamically load the code to run. 14:53:55 <eharney> it was added to let privsep work in a venv 14:54:18 <eharney> the implication there is that production services wouldn't be running in a venv equivalent 14:54:27 <eharney> but if you are running in kolla and it runs the service that way... 14:55:44 <yuval> kolla - running in a docker. but sometimes it will run inside the docker a venv (on dev mode) adding the variable to kolla using kolla-ansible is really not trivial 14:56:38 <yuval> anyway - next release I would be happy to dig in the issue and understand what can be done (maybe the change need to be in the kolla side) 14:57:55 <eharney> rosmaita: whoami-rajat: i left a thought for you to chew on on https://review.opendev.org/c/openstack/os-brick/+/916737 -- probably doesn't taste too good though 14:58:07 <yuval> gorka is still around? 14:59:32 <rosmaita> don't see him around today 14:59:45 <jbernard> yuval: he will repond to email as well, i dont' think he's online right now 14:59:52 <jbernard> that's our time 14:59:59 <jbernard> thank you everyone, happy reviewing! :) 15:00:04 <jbernard> #endmeeting