15:00:02 <mnasiadka> #startmeeting kolla 15:00:02 <opendevmeet> Meeting started Wed Nov 3 15:00:02 2021 UTC and is due to finish in 60 minutes. The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:02 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:02 <opendevmeet> The meeting name has been set to 'kolla' 15:00:09 <mnasiadka> #topic rollcall 15:00:14 <mnasiadka> o/ 15:02:40 <mgoddard> \o 15:02:51 <adrian-a> o/ 15:03:10 <parallax> \o 15:04:43 <frickler> o/ 15:05:00 <yoctozepto> o/ 15:05:13 <yoctozepto> mnasiadka: remember about the extra agenda in etherpad :-) 15:05:22 <mnasiadka> #topic agenda 15:06:10 <mnasiadka> * Announcements 15:06:10 <mnasiadka> * Review action items from the last meeting 15:06:10 <mnasiadka> * CI status 15:06:10 <mnasiadka> * Release tasks 15:06:10 <mnasiadka> * Yoga cycle planning 15:06:12 <mnasiadka> * Security bugs to squash 15:06:12 <mnasiadka> * Switch docs to recommend installing from git repo; re: https://review.opendev.org/c/openstack/kolla-ansible/+/815043 15:06:14 <mnasiadka> * New core-reviewer 15:06:14 <mnasiadka> * Open discussion 15:06:33 <mnasiadka> #topic Announcements 15:06:46 <mnasiadka> I'm off next week, any volunteer to run the meeting next Wed? 15:07:11 <yoctozepto> let me check 15:07:25 <yoctozepto> ok, I'm available 15:07:32 <mnasiadka> Ok then 15:07:45 <mnasiadka> #action yoctozepto to run the meeting next week 15:07:48 <mnasiadka> Thanks! 15:07:53 <yoctozepto> yw :-) 15:07:57 <mnasiadka> #topic Review action items from the last meeting 15:08:19 <mnasiadka> Seems none last week 15:08:33 <mnasiadka> #topic CI status 15:09:02 <mnasiadka> seems green 15:09:08 <mnasiadka> Kayobe is amber on Wallaby (disk issues) 15:09:32 <mnasiadka> change was merged, should we update back to green? 15:10:26 <mnasiadka> I'll check history later and see if it's green again (and check with priteau) 15:10:35 <mnasiadka> #topic Release tasks 15:10:46 <mnasiadka> mgoddard, yoctozepto: is it time to cut RC2? 15:10:56 <mgoddard> sure 15:11:10 <mnasiadka> https://review.opendev.org/c/openstack/kolla-ansible/+/814942 this was merged (mentioned on last weeks meeting) 15:11:18 <mnasiadka> and I think we reverted the problematic patch 15:11:20 <mnasiadka> ok then 15:11:39 <yoctozepto> what about mariadb patch? 15:11:45 <mnasiadka> which one? 15:11:54 <yoctozepto> the one I mentioned yesterday 15:11:55 <yoctozepto> https://review.opendev.org/c/openstack/kolla-ansible/+/814276 15:12:31 <mnasiadka> Ah, I did not look into that, since mgoddard reviewed that earlier - so I waited for him to act upon it ;-) 15:13:06 <mgoddard> I'll try to have a look later 15:13:16 <yoctozepto> so wait for this for rc2 15:13:27 <yoctozepto> this should get mariadb to its regular glory 15:13:38 <mnasiadka> added rc2-blocker hashtag and will keep an eye for that 15:13:56 <mnasiadka> and once it merges (and is backported) will post rc2 releases patches 15:14:14 <mnasiadka> ok, let's move on 15:14:26 <mnasiadka> #topic Yoga cycle planning 15:14:41 <mnasiadka> I haven't been able to populate Priorities on the whiteboard yet 15:15:36 <mnasiadka> But will do at latest tomorrow 15:15:50 <mnasiadka> Any other things that we need to consider at this point? 15:17:26 <yoctozepto> nothing specific from me 15:18:17 <mnasiadka> Other day I was thinking if we shouldn't follow what some other projects do - post ,,bugs'' with [RFE] prefix and target them to Yoga milestone in Launchpad - but I think priorities on the whiteboard worked quite OK 15:19:36 <mnasiadka> yoctozepto, mgoddard: opinions? (sometimes I feel the whiteboard is a bit ,,overcrowded'') 15:20:48 <mgoddard> mnasiadka: you're basically reintroducing the blueprints that we dropped the other week? 15:21:04 <mnasiadka> not really reintroducing, but fair point - let's stick to whiteboard for now :) 15:21:18 <mnasiadka> #topic Security bugs to squash 15:21:28 <mnasiadka> yoctozepto: do we have any? 15:22:05 <adrian-a> Whiteboard is crashing Chrome on Android :) and some prefixes or tags (in a similar idea like in Gitlab https://ibb.co/RpRjNng ) could be useful 15:22:57 <yoctozepto> mnasiadka: we do 15:23:51 <yoctozepto> mnasiadka: this topic to make you look at them 15:24:17 <mnasiadka> well, where do I find them in launchpad? 15:24:43 <yoctozepto> gimme a sec 15:25:05 <yoctozepto> kolla https://bugs.launchpad.net/kolla/+bugs?field.searchtext=&orderby=-importance&search=Search&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.informat 15:25:05 <yoctozepto> ion_type%3Alist=PRIVATESECURITY&field.information_type%3Alist=USERDATA&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=&field.tags_combinator=ANY&field.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_patch.used=&field.has_ 15:25:05 <yoctozepto> branches.used=&field.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.has_no_blueprints.used=&field.has_no_blueprints=on 15:25:19 <yoctozepto> kolla-ansible https://bugs.launchpad.net/kolla-ansible/+bugs?field.searchtext=&orderby=-importance&search=Search&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONS 15:25:19 <yoctozepto> E&field.information_type%3Alist=PRIVATESECURITY&field.information_type%3Alist=USERDATA&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=&field.tags_combinator=ANY&field.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_patch. 15:25:19 <yoctozepto> used=&field.has_branches.used=&field.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.has_no_blueprints.used=&field.has_no_blueprints=on 15:26:08 <yoctozepto> I'm open to explaining/discussing mine in there if you need more details 15:26:20 <yoctozepto> as for the others, I guess we should decide whether not to open them 15:26:26 <yoctozepto> or close entirely (-: 15:26:42 <mnasiadka> oh boy, nice links - I'll make some shortened urls and look through them ;-) 15:27:03 <yoctozepto> that's launchpad for ye 15:27:44 <mnasiadka> ok, the first two (haproxy and horizon dir listings) seem self explanatory 15:27:56 <mnasiadka> do we have any volunteers to work on those? 15:28:04 <yoctozepto> yeah, need to check if they happen still 15:28:35 <yoctozepto> I hoped you could spend some resources :-) I'm overloaded these days 15:28:46 <mnasiadka> and the last time seems not so trivial, because we would need to skim the logs in search of those passwords 15:28:52 <adrian-a> yoctozepto, can you please shorten that URL? I can't concat it to something meaningful 15:29:15 <mnasiadka> Ok, let me at least triage those - and get back with some updates to those bugs. 15:29:25 <mnasiadka> adrian-a: it's a private list, I don't think you have access. 15:29:34 <adrian-a> o, mkay 15:29:53 <yoctozepto> yeah, he does not 15:29:55 <mnasiadka> #action mnasiadka to triage security bugs and update them with resolution plan (if needed) 15:30:06 <yoctozepto> great, mnasiadka :-) 15:30:13 <mnasiadka> #topic Switch docs to recommend installing from git repo; re: https://review.opendev.org/c/openstack/kolla-ansible/+/815043 15:30:18 <mnasiadka> yoctozepto: that's yours? 15:30:26 <yoctozepto> mnasiadka: always mine 15:30:56 <yoctozepto> there is some discussion as to how to handle the recommendation on the source of kolla-ansible code 15:31:08 <yoctozepto> I argue we are better off recommending git 15:31:33 <yoctozepto> as this is what we test and I guess also run in production as we don't release often enough 15:31:44 <yoctozepto> moreover, the versioning is confusing 15:32:09 <yoctozepto> each component has a different version so hard to tell what has been installed from the version only 15:32:12 <yoctozepto> unless reading renos 15:32:17 <yoctozepto> which folks don't do (-: 15:32:29 <yoctozepto> or at least not often enough to make me glad they do 15:33:03 <mnasiadka> well, from one perspective I'm ok with that, from other - maybe we should do releases more often 15:34:38 <mnasiadka> but it's fine for the docs to point to git instead of pypi 15:34:42 <mnasiadka> any other voices of reason? 15:34:57 <adrian-a> sounds good 15:35:23 <mnasiadka> mgoddard, frickler? 15:35:54 <mgoddard> fine by me 15:36:06 <frickler> git is good 15:36:31 <mnasiadka> #agreed to recommend installing from git repo in the docs 15:36:37 <yoctozepto> frickler: I'll print that quote and frame it 15:36:50 <yoctozepto> "git is good" 15:36:59 <mnasiadka> yoctozepto: are you going to follow up? 15:37:16 <yoctozepto> yeah, action me on that to keep this flowing on 15:37:23 <mnasiadka> well, I think the author is adrian-a, right? 15:37:38 <adrian-a> I'll add a comment to the review and link to IRC log and I'll complete the commit with git 15:37:47 <mnasiadka> great, case solved 15:38:08 <mnasiadka> let's move on 15:38:22 <mnasiadka> #topic New core-reviewer 15:39:22 <mnasiadka> I think it's time to add a new core reviewer - especially if that person is outside of StackHPC - out of the list of contributors - I think kevko is a good candidate (with proper review stats and good knowledge of kolla/kolla-ansible code) 15:40:01 <yoctozepto> adrian-a: many thanks! 15:40:17 <adrian-a> yoctozepto, yw :) 15:40:23 <mnasiadka> If there are no objections - I'll propose him through the mailing list. 15:40:41 <mnasiadka> And the question is - first kolla or kolla-ansible or both? 15:40:49 <mgoddard> both 15:40:52 <yoctozepto> mnasiadka: both 15:41:16 <frickler> +2 ... oh wait, I can only +1 ... ;) 15:41:46 <mnasiadka> ok then, both 15:41:47 <yoctozepto> frickler needs to work a bit more to gain the core title :-) 15:42:11 <frickler> yes, I plan on doing that, but that'll take some time I agree 15:42:29 <yoctozepto> no rush, quality over quantity :-) 15:42:37 <mnasiadka> yup 15:43:03 <mnasiadka> #topic Open discussion 15:43:14 <mnasiadka> Phew, we made to open discussion this time ;-) 15:43:41 <yoctozepto> oh noez 15:43:49 <yoctozepto> let's check if we have not missed some point in the agenda (-: 15:44:05 <yoctozepto> c'est impossible ! 15:44:16 <yoctozepto> we have gone through all of 'em 15:44:21 <yoctozepto> congrats mnasiadka 15:46:15 <mnasiadka> ok, no open discussion points from anybody? ;-) 15:48:49 <yoctozepto> c'est impossible aussi ! 15:49:01 <headphoneJames> well, I have a FYI that I'm trying to make a minimal change for switching over service configs in keystone_authtoken from project scoped tokens to system scope. 15:49:18 <headphoneJames> should be ready soon 15:49:22 <adrian-a> I see 3 options on this (last comment), what do you think? https://review.opendev.org/c/openstack/kolla-ansible/+/816076 15:50:17 <yoctozepto> headphoneJames: sounds wonderful 15:50:58 <mgoddard> adrian-a: A2 or C 15:51:02 <mnasiadka> yoctozepto: haven't we tried to get rid of init-runonce from tools/ at some point? 15:51:33 <mgoddard> where C is pip install python-openstackclient -c <upper constraints URL> 15:51:35 <yoctozepto> mnasiadka: we did 15:51:48 <yoctozepto> mgoddard: ++, in docs only though 15:51:51 <yoctozepto> adrian-a: ^ 15:52:03 <yoctozepto> mnasiadka: I guess we should then hurry 15:52:27 <yoctozepto> and amend the docs not to recommend that as something necessary 15:52:32 <mnasiadka> Well, hurry or not - I think we should discuss if init-runonce is the toolset we want to maintain and if users should be really running it ;-) 15:53:06 <mgoddard> well, it kind of works, and we recommend against running it in production 15:53:17 <mgoddard> but we need something like it for testing 15:53:17 <yoctozepto> mnasiadka: let's discuss then - we should not recommend it :D who objects? 15:55:00 <yoctozepto> no objections 15:55:02 <mnasiadka> I think we already recommend against it, I'm fine in adding a message to post-deploy to install python-openstackclient (and some other clients - preferably in a venv ) or in docs 15:55:17 <mnasiadka> but not really to automate that part 15:55:24 <yoctozepto> action me to hide it properly 15:55:28 <yoctozepto> and go-go-go 15:56:00 <mnasiadka> #action yoctozepto hide properly init-runonce 15:56:03 <mnasiadka> (whatever that means) 15:56:13 <adrian-a> So should I leave pip install python-openstackclient + maybe add a note this installs the latest stable client and some example pip commands with git URL and tags for other releases? 15:56:14 <yoctozepto> lol 15:56:20 <adrian-a> *in docs (leave) 15:56:27 <yoctozepto> adrian-a: see mgoddard's message about -c 15:56:39 <EugenMayer> mnasiadka since i really automated a lot of this in the last 3 weeks. I find installing the client is entirely mandatory since openstack is useless with GUI only. init-runonce i ack that it is just not the right thing and instead of running it (what i did in the start) i did something like 15:56:39 <EugenMayer> https://github.com/EugenMayer/openstack-lab/blob/stable/ovn/README.setup.md 15:56:54 <mnasiadka> leave in docs, add upper constraints 15:56:56 <adrian-a> yoctozepto: I haven't found a '-c' flag in pip, not sure what that does 15:57:13 <yoctozepto> adrian-a: constrains versions 15:57:20 <yoctozepto> it will do the right thing (TM) 15:57:40 <adrian-a> and it's not documented? haven't found it, e.g. https://manpages.debian.org/stretch/python-pip/pip.1 15:57:42 <mgoddard> yoctozepto: kayobe uses init-runonce, so at a minimum please add a symlink to avoid breaking us 15:57:42 <mnasiadka> EugenMayer: sure, but it depends on users environment - we can fix the example in docs how to install the openstack client - but I don't think we should be automating that in post-deploy 15:58:10 <yoctozepto> mgoddard: oh my, you mean in CI, right? 15:58:15 <yoctozepto> RIGHT?! :D 15:58:22 <EugenMayer> mnasiadka: i would say, that installing the openstack client,when using kolla, on the deployer is an key concept. It cannot be really optional IMHO 15:58:23 <mgoddard> yoctozepto: well, testing, yes 15:58:30 <yoctozepto> mgoddard: ok 15:58:57 <mnasiadka> EugenMayer: but we can't cover all cases, and sometimes you don't use the deployment host as the client host, I still prefer docs :) 15:59:10 <mnasiadka> ok then, I think it's enough for today 15:59:12 <yoctozepto> EugenMayer: it can, the deployment machine might not ever be used as the client 15:59:15 <mnasiadka> Thanks for attending! 15:59:16 <yoctozepto> ;-) 15:59:19 <mnasiadka> #endmeeting