Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-04-25
10:55:37 sean-k-mooney that is just the polomer ui + a theme
10:55:40 aspiers nope
10:56:17 stephenfin aspiers: Oh, nice
10:56:19 aspiers this is just one of many specific improvements
10:56:28 aspiers I can't list them all for you
10:56:31 stephenfin Resolvable comments are indeed one of the things I'd like
10:56:38 stephenfin I like that in gdocs
10:58:06 aspiers also bear in mind those videos are 1.5 years old now
10:58:19 stephenfin yeah, so things will have moved on a lot since then
10:58:21 aspiers OpenStack really REALLY needs a Gerrit upgrade
10:58:45 stephenfin I imagine the infra folks are snowed under with the opendev migration but this sure would be a nice next step
10:58:47 sean-k-mooney aspiers: we used to have some downstream only changes
10:59:05 sean-k-mooney because we need to fix bug that had not merged upstream
10:59:07 aspiers stephenfin: yeah. it will happen
10:59:18 stephenfin big changes in 3.0 too, from the looks of things https://www.gerritcodereview.com/3.0.html
10:59:19 sean-k-mooney im not sure what is left to get upstream at this point
10:59:25 aspiers sean-k-mooney: if a patch to 2.13 hasn't landed by now, it never will
10:59:45 sean-k-mooney aspiers: yes which may mean we never update
10:59:59 aspiers sean-k-mooney: highly doubt that
11:00:20 sean-k-mooney there were one or two that we need to work at our scalse in the past the got push back which stop openstack updating
11:00:25 aspiers clarkb / corvus / fungi started working on it a long time ago
11:00:38 sean-k-mooney but i think most of that has gone upstream now
11:00:49 aspiers sean-k-mooney: yeah, we're not the biggest Gerrit user
11:01:03 aspiers there are several other huge scale installs
11:01:08 aspiers I met them all at the conference
11:01:19 sean-k-mooney i think we were 6 years ago when i started on openstack
11:01:21 aspiers and the others are all head in terms of HA too
11:01:30 aspiers ahead
11:01:38 aspiers Google is heavily investing in Gerrit
11:01:46 aspiers they have a team of 20 or so IIRC
11:02:13 aspiers stephenfin: 3.0 is more about ditching the dead stuff which everyone will have migrated off of by 2.16
11:02:25 stephenfin looks like it, yeah
11:02:37 stephenfin I'd imagine we'll be getting 2.16 first, in that case
11:03:12 aspiers for sure
11:12:37 openstackgerrit Adam Spiers proposed openstack/os-traits master: Update SEV trait docs to avoid misleading people https://review.opendev.org/655671
11:19:33 openstackgerrit Adam Spiers proposed openstack/os-traits master: Document policy of never removing traits https://review.opendev.org/655673
11:23:27 openstackgerrit Adam Spiers proposed openstack/os-traits master: Document policy of never removing traits https://review.opendev.org/655673
11:36:23 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/os-vif stable/stein: Prevent "qbr" Linux Bridge from replying to ARP messages https://review.opendev.org/655678
11:36:38 openstackgerrit Kashyap Chamarthy proposed openstack/nova-specs master: Add "CPU selection with hypervisor consideration" spec https://review.opendev.org/645814
11:36:50 kashyap efried: As promised ^^
11:38:37 jaypipes efried: morning. sorry, just saw your question from yesterday evening.
11:39:24 jaypipes efried: I'll be leaving around 1:30-2pm on the Friday. I'll just make do with side conversations etc at the summit about the various specs I'm concerned about (and continue to add reviews to said specs)
11:41:12 sean-k-mooney jaypipes: i think efried has tried to front load most of the topic that are in your usuall area. not sure they all fit
11:41:45 sean-k-mooney one thing i think will come up in the feeback session is the ptg and summit should over lap again
11:42:22 sean-k-mooney we will see but the room was pretty split at the last ptg on that topic
11:43:18 jaypipes ack
12:05:48 openstackgerrit Merged openstack/nova stable/queens: Fix regression in glance client call https://review.opendev.org/655186
12:10:29 openstackgerrit Merged openstack/nova stable/queens: libvirt: disconnect volume when encryption fails https://review.opendev.org/653033
12:23:37 fungi aspiers: stephenfin: sean-k-mooney: yes, current plan i think is going straight from 2.13 to 2.16 we just need to re-test the upgrade process on our staging server. the transition from mysql to notedb is going to be the biggest challenge as it will likely require a day-long service outage to complete
12:24:13 fungi we didn't want to conflate that work with the opendev switch for gerrit/git, but this ptg will be the time to make plans
12:24:47 sean-k-mooney is there an infra meetup at the ptg or just people will be about so it will come up
12:25:15 fungi infra is meeting thursday and friday all day (we share a room with qa)
12:25:25 sean-k-mooney ah ok
12:26:37 sean-k-mooney i kind of want to get a bit more involved with the infra side of things but on the other hand was not sure how to go about that so i was just going to say hi
12:54:05 kashyap fungi: What is "notedb"? /me looks up
12:54:42 kashyap Okay, it is this: "NoteDb is the next generation of Gerrit storage backend, which replaces the traditional SQL backend ..."
12:54:45 kashyap https://gerrit-review.googlesource.com/Documentation/note-db.html
12:55:40 fungi kashyap: yep, basically we'll be etl'ing gigabytes of data from an rdbms into git repositories
12:55:50 fungi not a speedy process
12:57:39 fungi at one point the upgrade process was necessary to be performed stepwise first moving some of the data into an h2 database and then moving it from there into git in a subsequent gerrit version, but we hear they've added the ability to migrate directly from sql to notedb now so it may go a lot more quickly, we just need to test
13:05:00 stephenfin fungi: Thanks for the context. That's all very good to hear :)
13:07:15 kashyap fungi: Yeah, noted
13:07:38 fungi also part of the upgrade process involved moving to a new server on a newer ubuntu release so we would have a new enough version of java to work with newer gerrit. we got that done last year, which was a nontrivial undertaking
13:08:20 sean-k-mooney fungi: out of interst shicne the data will now be sotred in git
13:08:37 sean-k-mooney could we use the notedb data in tools like gertty
13:08:54 sean-k-mooney or will that content not be downloadable by others
13:09:38 fungi well, gertty already has access to the relevant data (modulo access controls allowing your account to access it) through gerrit's rest api
13:09:42 sean-k-mooney gerrty effectivly buildis its own sqlite db replicting the gerrit one if that was not required it could be a nice improvemnt
13:10:19 fungi oh, i see what you're asking... i think the data in notedb is aggregated in ways which wouldn't make that useful for per-user data tracking
13:11:09 fungi what it does enable though is distribution of state, an there's a plugin recently contributed upstream to gerrit by the gerrithub folks which does multi-master replication between notedbs
13:11:36 fungi so we could finally have a decent cross-provider high availability solution for it
13:11:50 sean-k-mooney ya that would be cool
13:12:28 fungi i know the wmf folks have just upgraded to 2.16 and are looking at leveraging that plugin soon
13:12:49 sean-k-mooney wmf?
13:12:59 sean-k-mooney wiki media foundaiton?
13:15:13 sean-k-mooney actully looking at https://gitenterprise.me/tag/notedb/ it looks like it will allow you to work offline
13:15:26 stephenfin ffs https://review.opendev.org/651297
13:16:14 stephenfin there is definitely something buggy with one of our lower-constraint packages to say I keep seeing these StringException errors http://logs.openstack.org/97/651297/4/gate/openstack-tox-lower-constraints/758a74d/testr_results.html.gz
13:16:17 stephenfin But who knows what
13:19:25 efried stephenfin: That's bug 1825435
13:19:26 openstack bug 1825435 in OpenStack Compute (nova) "TestRPC unit tests intermittently fail with "'>' not supported between instances of 'NoneType' and 'datetime.datetime'" - maybe due to "Fatal Python error: Cannot recover from stack overflow."" [High,Confirmed] https://launchpad.net/bugs/1825435
13:20:00 sean-k-mooney did we find where that was intoduced yet
13:20:12 efried it seems almost exclusively isolated to py36, so it hits lower-constraints (which runs on py36) and the py36 UT.
13:20:40 efried sean-k-mooney: No. mriedem has been trying to debug it (https://review.opendev.org/#/c/654468/ and its dep) but we can't get it to reproduce with those patches
13:20:41 sean-k-mooney apparently it started to show up first in stephenfin cells v1 removal patches but that could also be a coincidnces
13:21:12 efried sean-k-mooney: If you can update the video models bp description as requested by kashyap, I'll approve it.
13:21:34 sean-k-mooney am sure ill go do that now.
13:21:35 efried jaypipes: I was specifically asking if those two topics were ones you wanted to be around for. If so, I can try to shuffle things around.
13:21:39 efried thanks sean-k-mooney
13:21:42 kashyap sean-k-mooney: Thank you!
13:22:25 efried aspiers: If you're around and want to fix that one typo in https://review.opendev.org/#/c/655673/ I'll fast approve.
13:22:36 sean-k-mooney i have it more or less in the relse note https://review.opendev.org/#/c/647733/3/releasenotes/notes/extend-libvirt-video-model-support-d630b99ef5039f51.yaml
13:22:39 efried If you're not around, I'll fix it.
13:22:52 efried sean-k-mooney: copy paste, baby
13:23:05 efried kashyap: You okay with the wording there --^ ?
13:23:35 sean-k-mooney i need to change it slightly but to make sense in a blueprint
13:23:46 sean-k-mooney mainly the tense but ill use it as teh basis
13:23:48 stephenfin efried: I don't think it is, actually
13:23:51 mriedem lyarwood: seems like an easy bug to fix, but you should look https://bugs.launchpad.net/nova/+bug/1825882
13:23:52 openstack Launchpad bug 1825882 in OpenStack Compute (nova) "Virsh disk attach errors silently ignored" [Undecided,New]

Earlier   Later