| jhesketh | fungi: have you (or do you know of anybody who may have) scripted EOL'ing branches? | 00:44 |
|---|---|---|
| fungi | jhesketh: i've done nasty ad hoc one-liners for it in the past | 00:45 |
| fungi | basically i keep clones of everything we have in gerrit on my workstation, so i can just do a for loop over a list of repos, pushd into them, run arbitrary git/git-review commands, and then popd again | 00:46 |
| fungi | usually i'd whip up a text file containing the (short) tag message and pass that with an option to git so it was consistent and that way git wouldn't spawn an editor over and over | 00:47 |
| fungi | and then push an empty refname to the old branch name once done (and store a list of the prior shas of all the branches so i could undo it in case of a mistake) | 00:48 |
| jhesketh | fungi: righto.. also I'm guessing the general process is to add myself to the bootstrappers then it's just git push gerrit tag && git push gerrit --delete stable/kilo? | 00:48 |
| fungi | yep, that's basically it | 00:49 |
| fungi | but you know, with some logging and debug details so you can fix any accidents | 00:49 |
| jhesketh | okay I might script something up.. thanks :-) | 00:49 |
| fungi | thanks for looking into it! | 00:50 |
| fungi | jhesketh: also be mindful you'll need to check for (and abandon) any open reviews on the branches in question, or defer deletion until you can get those projects' core teams to do so | 00:51 |
| jhesketh | rgr | 00:51 |
| fungi | a quick gerrit query can double-check that for you | 00:51 |
| jhesketh | yeah, tonyb has already gotten that for us :-) | 00:52 |
| fungi | all' | 00:53 |
| fungi | s the better | 00:53 |
| tonyb | jhesketh, fungi: the only open ones remaining are the ones I don't have permission to close. | 00:53 |
| tonyb | tiny.cc/os-kilo-still-open is the dash of the open reviews but scripting would be quicker. | 00:54 |
| tonyb | If it's just those git commands I can add them to the same script if you like. | 00:57 |
| *** zul has joined #openstack-stable | 01:00 | |
| jhesketh | tonyb: I'm working on a tool to do it | 01:11 |
| tonyb | jhesketh: Cool, Thanks. | 01:11 |
| *** mrunge has quit IRC | 01:15 | |
| *** mrunge has joined #openstack-stable | 01:21 | |
| jhesketh | tonyb, fungi: my bash and git foo are a little weak, but here is my first pass: https://review.openstack.org/331910 | 01:52 |
| tonyb | jhesketh: Thanks | 01:56 |
| fungi | jhesketh: far from exhaustive, but a few suggestions left on that review | 02:01 |
| fungi | all in all looks good, mostly robustness items | 02:01 |
| jhesketh | Thanks :-) | 02:01 |
| fungi | spotted a syntax error too | 02:03 |
| * tonyb is too slow with his reviews | 02:08 | |
| jhesketh | tonyb: thanks, a quick comment inline about signing tags | 02:11 |
| tonyb | replied | 02:18 |
| jhesketh | ta | 02:18 |
| *** mriedem has quit IRC | 02:26 | |
| jhesketh | tonyb, fungi: new version up for review https://review.openstack.org/331910 | 02:53 |
| *** stevemar has quit IRC | 02:54 | |
| tonyb | jhesketh: Thanks +1'd sort of .... | 03:13 |
| jhesketh | tonyb: thanks.. tidied it up slightly | 03:39 |
| jhesketh | tonyb: should probably give this a go and retire an old branch | 03:39 |
| tonyb | jhesketh: go for it .... I can't do that thing | 03:39 |
| jhesketh | the gerrit documentation lies... it isn't matching a review to the commit sha and wants the changeid (which is much harder to grep for) | 03:49 |
| jhesketh | fine, I'll do it the nova way which requires jq to be installed | 03:52 |
| jhesketh | wait, that outputs sha's... | 03:54 |
| jhesketh | tonyb: look at https://review.openstack.org/Documentation/cmd-review.html and http://paste.openstack.org/show/520853/ and see if you can tell me what I'm doing wrong :-s | 03:56 |
| jhesketh | (please if yo uhave time) | 03:56 |
| tonyb | jhesketh: looking .... | 03:56 |
| tonyb | jhesketh: Well that looks right to me. | 04:04 |
| tonyb | jhesketh: and matches what nova's script does AFAICT | 04:05 |
| jhesketh | yep, hence the head+desk | 04:05 |
| jhesketh | tonyb: playing around, that second one works: http://paste.openstack.org/show/520855/ | 04:06 |
| jhesketh | ah okay, so it doesn't like the --branch param | 04:08 |
| tonyb | jhesketh: Oh so --project and --branch bork it unless they fix it? | 04:08 |
| jhesketh | otherwise it works | 04:08 |
| tonyb | Oh okay | 04:08 |
| jhesketh | unless they fix it? | 04:08 |
| tonyb | well the docs that that --project and --barch are used if the SHA/CHange# have duplicates | 04:08 |
| jhesketh | so if hte same commit is in multiple branches it may get confused, but that's not too bad as at least it'll still constrain to one project (rather than any commit from any project) | 04:08 |
| jhesketh | tonyb: oh okay... well it works both with and without --project | 04:09 |
| jhesketh | it just seems to be --branch it doesn't like | 04:09 |
| tonyb | jhesketh: okay. | 04:09 |
| jhesketh | and it'd be odd to only use --project or --branch if there were duplicates... so something weird is going on | 04:09 |
| jhesketh | either way, ommiting --branch should be safe | 04:09 |
| jhesketh | okay abanding part of the script works | 04:11 |
| tonyb | jhesketh: https://review.openstack.org/#/c/236642/2 look at the last few comments ... | 04:18 |
| jhesketh | tonyb: that was me tryign different abandon methods | 04:19 |
| tonyb | jhesketh: okay. | 04:19 |
| jhesketh | tonyb: so the --message flag appears to leave multiple messages on a change... it seems like a gerrit bug but I'm trying to confirm https://review.openstack.org/#/c/331937/ | 04:44 |
| tonyb | jhesketh: You know we have openstack-dev/sandboc for that right? | 04:45 |
| tonyb | *sandbox* | 04:45 |
| jhesketh | yes, I should have done it there | 04:45 |
| tonyb | I'm not seeing multiple messages | 04:46 |
| jhesketh | sorry, that change is a bad example... | 04:47 |
| jhesketh | 1st abandon was via the web ui.. the 1st "not needed" comment was abandoning with a message while it was already in abandoned state | 04:47 |
| jhesketh | so it added the message and kept the state | 04:48 |
| jhesketh | then it's restored and abanded via the cli.. this "Not needed" appears as a comment on the patch and then again on the Abandoned action | 04:48 |
| jhesketh | Then restoring a 2nd time, the last Abandon is via the CLI with no --message which abandons with just 1 comment (correctly) | 04:48 |
| tonyb | jhesketh: Hmm That's a bit of a pain :( | 04:51 |
| jhesketh | tonyb: right... so our options are double comment, fix upstream gerrit (err, no thanks), leave no message | 04:53 |
| jhesketh | I'm thinking double comment and moving on with our lives | 04:53 |
| tonyb | jhesketh: I basically agree, there is another option .. the REST interface doesn't seem to have that same issue with double comments | 04:55 |
| jhesketh | oh? did you run a test? | 04:55 |
| tonyb | jhesketh: I used the REST API to abandon 40somethign reviews that I had permissions to do. | 05:02 |
| tonyb | jhesketh: I s'pose I shoudl find one or two to verify that it didn't leave double messages. | 05:03 |
| jhesketh | tonyb: got a script? | 05:03 |
| tonyb | jhesketh: It's rough, ignore the 'gah_repos' stuff http://paste.openstack.org/show/520859/ | 05:05 |
| jhesketh | tonyb: maybe we should put that in the releases tooling and have the eol script call it? | 05:08 |
| tonyb | jhesketh: Yeah it'd need tweking to be used that way but adding it to release-tools is my plan. | 05:09 |
| tonyb | jhesketh: Also probably need a better way to get the http creds to it than via the cli | 05:11 |
| tonyb | jhesketh: Tweaked slightly: http://paste.openstack.org/show/520861/ | 05:15 |
| *** armax has joined #openstack-stable | 05:41 | |
| *** armax has quit IRC | 05:46 | |
| *** rcernin has joined #openstack-stable | 05:56 | |
| jhesketh | there aren't very many non-abandoned patches in your list... so I'm just going to do those manually and proceed with my script | 05:58 |
| jhesketh | tonyb: okay, shall we do this thing live? starting with devstack? | 06:03 |
| *** armax has joined #openstack-stable | 06:12 | |
| *** armax has quit IRC | 06:17 | |
| tonyb | jhesketh: So it looks like you've abandoned all the changes :) | 06:20 |
| tonyb | jhesketh: Thanks. | 06:20 |
| jhesketh | no worries | 06:20 |
| *** armax has joined #openstack-stable | 06:21 | |
| jhesketh | tonyb: so is openstack-dev/devstack ready to go? eg there are no jobs or anything depending on it etc | 06:22 |
| tonyb | jhesketh: It's my understanding that we're good to drop it. | 06:23 |
| tonyb | jhesketh: but I understand if you want to leave it alive for 24hours or so to verify that | 06:23 |
| jhesketh | what do we need to verity? | 06:23 |
| jhesketh | maybe we should start with an easier project | 06:24 |
| tonyb | jhesketh: Yeah maybe leave devstack and requirements 'til last | 06:31 |
| jhesketh | I was thinking the same for grenade and pbr | 06:31 |
| jhesketh | tonyb: so what about astara | 06:31 |
| jhesketh | do we need to do any due diligence there? | 06:32 |
| tonyb | jhesketh: Sounds good | 06:32 |
| jhesketh | tonyb: where did you get this list from btw... are they big-tent projects with kilo branhces? | 06:32 |
| tonyb | jhesketh: that's how it ended up, but it started as a much smaller set, then a few people opted in so I switched it around to opt out | 06:33 |
| tonyb | jhesketh: and only a few things opted out | 06:33 |
| tonyb | jhesketh: https://gist.github.com/tbreeds/7de812a5d363fab4bd425beae5084c87 at the end you can see the ones that opted out | 06:33 |
| jhesketh | tonyb: is that the most up to date list? ie should I use that instead of your mailing list one? | 06:34 |
| tonyb | So you're right we need to leave devstack/grenade/requirements until tomorrow and I'll verify that the projects staying active are going to work without those things | 06:34 |
| tonyb | jhesketh: the mailing list one was generated at the same time | 06:35 |
| tonyb | jhesketh: the list I gave you had all the 'tagged repos removed which I now know is wrong | 06:35 |
| jhesketh | tonyb: okay, lets do openstack/astara and see what happens | 06:36 |
| tonyb | jhesketh: because that will leave the kilo branch behind in those projects | 06:36 |
| tonyb | jhesketh: okay, then perhaps we shoudl wit until tomorrow | 06:36 |
| tonyb | Phoebe is very distratcing | 06:36 |
| jhesketh | tonyb: all good, we can wait until tomorrow for all of them | 06:36 |
| jhesketh | we have the tooling in place so we just need to make sure people are ready for branches to disappear and then we're good to go | 06:36 |
| tonyb | jhesketh: cool | 06:37 |
| tonyb | they *should* be fine they were warned about it several times etc | 06:37 |
| jhesketh | cool | 06:40 |
| *** agireud has quit IRC | 06:51 | |
| *** pcaruana has joined #openstack-stable | 06:52 | |
| *** agireud has joined #openstack-stable | 06:55 | |
| *** ihrachys has joined #openstack-stable | 07:25 | |
| *** apevec has joined #openstack-stable | 07:48 | |
| *** e0ne has joined #openstack-stable | 08:11 | |
| *** agireud has quit IRC | 08:19 | |
| *** agireud has joined #openstack-stable | 08:28 | |
| *** armax has quit IRC | 08:44 | |
| *** derekh has joined #openstack-stable | 08:44 | |
| *** armax has joined #openstack-stable | 09:34 | |
| *** armax has quit IRC | 09:58 | |
| *** ihrachys has quit IRC | 10:22 | |
| *** amrith has quit IRC | 10:28 | |
| *** amrith has joined #openstack-stable | 10:29 | |
| *** amrith has quit IRC | 10:32 | |
| *** amrith has joined #openstack-stable | 10:33 | |
| *** armax has joined #openstack-stable | 11:06 | |
| *** ihrachys has joined #openstack-stable | 11:24 | |
| *** ihrachys has quit IRC | 11:26 | |
| *** armax has quit IRC | 11:33 | |
| *** ihrachys has joined #openstack-stable | 11:34 | |
| *** armax has joined #openstack-stable | 11:43 | |
| *** dtantsur|afk is now known as dtantsur | 12:01 | |
| *** apevec has quit IRC | 12:17 | |
| *** armax has quit IRC | 12:20 | |
| *** armax has joined #openstack-stable | 12:27 | |
| *** takedakn has joined #openstack-stable | 12:29 | |
| *** armax has quit IRC | 12:35 | |
| *** zul has quit IRC | 12:35 | |
| *** armax has joined #openstack-stable | 12:35 | |
| *** zul has joined #openstack-stable | 12:35 | |
| *** fesp has joined #openstack-stable | 12:45 | |
| *** fesp has quit IRC | 12:46 | |
| *** apevec has joined #openstack-stable | 13:04 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:14 | |
| *** e0ne has quit IRC | 13:27 | |
| *** mriedem has joined #openstack-stable | 13:30 | |
| *** e0ne has joined #openstack-stable | 13:31 | |
| *** derekh has quit IRC | 13:36 | |
| *** eharney has joined #openstack-stable | 13:39 | |
| *** e0ne has quit IRC | 13:41 | |
| *** e0ne has joined #openstack-stable | 13:51 | |
| fungi | jhesketh: tonyb: fwiw, after playing around with the gerrit rest api a bunch, the ssh api is terribad by comparison. for a greenfield script i'd stick with using teh rest api for lookups and metadata/configuration. i only still use the ssh interface for git operations and the event stream | 14:08 |
| fungi | in theory git operations also work over https to gerrit now, i just haven't played around with it yet | 14:09 |
| jhesketh | fungi: agreed | 14:09 |
| fungi | and if mtreinish and i have our way, authenticated ssh access to the event stream will also be a thing of the past (via the "firehose") | 14:10 |
| jhesketh | fungi: so I think the script I hacked together today is ready to go fyi (well the abandon stuff is still cli but we've abandoned all the cahnges so it does nothing at the moment) | 14:10 |
| jhesketh | fungi: are you suggesting then we'd turn off other ssh access? | 14:11 |
| fungi | jhesketh: eventually maybe, not any time in the near future | 14:11 |
| jhesketh | sure | 14:12 |
| fungi | but the bridge to mqtt mtreinish wrote will make it possible for people to consume the gerrit event stream over an anonymous protocol | 14:12 |
| *** stevemar has joined #openstack-stable | 14:12 | |
| fungi | (and eventually i'd expect to see us add other data sources into the firehose too) | 14:13 |
| mtreinish | fungi: unfortunately hp really doesn't know how to keep the power on for their clouds. So my germqtt demo is no longer running | 14:13 |
| mtreinish | so we can't point to that anymore | 14:13 |
| fungi | meh. i could probably spin one up for us, but not this week | 14:13 |
| mtreinish | jhesketh: https://review.openstack.org/329115 | 14:14 |
| mtreinish | fungi: heh, as can I. It takes like 3 secs :) | 14:14 |
| jhesketh | might be something worth spinning up as an infra module | 14:14 |
| jhesketh | oh there's a spec | 14:14 |
| jhesketh | shiny | 14:14 |
| fungi | jhesketh: yeah, that's the plan in the spec | 14:14 |
| * jhesketh adds to review queue for tomorrow | 14:14 | |
| fungi | incomplete spec. i owe mtreinish some sections for it still | 14:15 |
| fungi | the "plan" i refer to may not _yet_ be in that spec. i think it's one of the things i still need to flesh out | 14:15 |
| mtreinish | fungi: I didn't put anything in there for the zuul to consume mqtt side | 14:16 |
| mtreinish | wasn't sure that was really part of the spec or not | 14:16 |
| fungi | it's not part of this spec. it's an option for the future | 14:20 |
| fungi | we can _mention_ in the spec that it would enable us to do things like that | 14:21 |
| fungi | and if it's easier for, say, people in china to get mqtt connections throug the great wall than it is ssh connections, i can imagine a new trigger type might show up as a patch for zuul pretty quickly | 14:22 |
| fungi | right now to run zuul behind the great wall, i gather you need a vpn tunnel (of questionable legality) to a rpoxy | 14:23 |
| *** armax has quit IRC | 14:23 | |
| mtreinish | yeah, especially if we enable it over a websocket on port 80 (which I think I put in that first draft) I can't imagine it would be worse for people in china | 14:23 |
| *** amrith has quit IRC | 14:28 | |
| *** amrith has joined #openstack-stable | 14:28 | |
| *** takedakn has quit IRC | 14:57 | |
| *** zul_ has joined #openstack-stable | 15:13 | |
| *** zul has quit IRC | 15:17 | |
| *** e0ne has quit IRC | 15:25 | |
| *** e0ne has joined #openstack-stable | 15:26 | |
| *** e0ne has quit IRC | 15:34 | |
| *** pcaruana has quit IRC | 15:35 | |
| *** ihrachys has quit IRC | 15:58 | |
| -openstackstatus- NOTICE: Gerrit is being restarted now to apply an emergency security-related configuration change | 16:06 | |
| *** stevemar has quit IRC | 16:29 | |
| *** e0ne has joined #openstack-stable | 16:32 | |
| *** armax has joined #openstack-stable | 16:51 | |
| *** catintheroof has joined #openstack-stable | 16:53 | |
| *** e0ne has quit IRC | 17:02 | |
| *** ihrachys has joined #openstack-stable | 17:03 | |
| apevec | dhellmann, fungi - firefox claims "releases.openstack.org uses an invalid security certificate" | 17:10 |
| apevec | it was fine few days ago iirc | 17:11 |
| apevec | or maybe I had exception? | 17:11 |
| fungi | apevec: try http. we don't have a cert for https://releases.openstack.org/ yet | 17:12 |
| *** ihrachys_ has joined #openstack-stable | 17:13 | |
| apevec | ah, yes I was using http before :) | 17:13 |
| apevec | sorry for the noise | 17:13 |
| apevec | fungi, while at it, has infra considered using letsencrypt? | 17:14 |
| apevec | we use it for rdoproject.org | 17:14 |
| fungi | apevec: i've seen far too few sites screw up certificate renewals for letsencrypt. maybe in a few years once it's matured | 17:14 |
| fungi | er, far too many | 17:14 |
| apevec | there are cron jobs :) | 17:15 |
| fungi | in the meantime i'm going to switch us to startssl, since we can get basically an unlimited number of certs and subjectaltnames for a reasonable flat rate | 17:15 |
| fungi | apevec: yep, cron jobs for that frighten me | 17:15 |
| apevec | old-school :) | 17:16 |
| fungi | i want to retrieve the new cert, manually check that it has the correct parameters still, load it myself and confirm the services are still working | 17:16 |
| fungi | and i don't want to do that every 90 days | 17:16 |
| *** ihrachys has quit IRC | 17:16 | |
| fungi | relying on a cron job for that means the site is down when the letsencrypt renewal process breaks for some reason, and stays broken until someone with root access is around, informed and addresses it | 17:17 |
| fungi | also, letsencrypt rubs me the wrong way. mofo gave cacert.org the runaround for years, then put together a coalition with google and some cabforum members to do basically the same thing, except it's still not community-maintained | 17:18 |
| fungi | also, mofo's trust store policies/cabal rub me the wrong way in general. backroom deals and quid pro quo | 17:20 |
| fungi | we still need to either get open process and community control over https or officially declare failure to actually secure http and look more seriously at alternatives | 17:21 |
| fungi | letsencrypt solves the easy problem and does nothing to address the hard issues | 17:21 |
| *** mriedem has quit IRC | 17:25 | |
| *** mriedem has joined #openstack-stable | 17:28 | |
| *** ihrachys_ has quit IRC | 17:32 | |
| *** ihrachys has joined #openstack-stable | 17:33 | |
| *** stevemar has joined #openstack-stable | 17:36 | |
| *** pcaruana has joined #openstack-stable | 17:40 | |
| *** rcernin has quit IRC | 17:43 | |
| *** ihrachys has quit IRC | 17:46 | |
| *** armax has quit IRC | 17:53 | |
| *** Rockyg has joined #openstack-stable | 18:06 | |
| *** e0ne has joined #openstack-stable | 18:19 | |
| *** pcaruana has quit IRC | 18:22 | |
| *** catintheroof has quit IRC | 18:51 | |
| *** catintheroof has joined #openstack-stable | 18:51 | |
| *** mriedem1 has joined #openstack-stable | 18:58 | |
| *** mriedem has quit IRC | 18:59 | |
| *** mriedem1 is now known as mriedem | 19:00 | |
| *** zul_ has quit IRC | 19:01 | |
| *** zul has joined #openstack-stable | 19:02 | |
| *** ihrachys has joined #openstack-stable | 19:25 | |
| *** hachi-ju has joined #openstack-stable | 19:28 | |
| *** e0ne_ has joined #openstack-stable | 19:28 | |
| *** e0ne_ has quit IRC | 19:28 | |
| *** arif-ali has quit IRC | 19:29 | |
| *** number80 has quit IRC | 19:29 | |
| *** hachi-ju has quit IRC | 19:29 | |
| *** e0ne has quit IRC | 19:30 | |
| *** number80 has joined #openstack-stable | 19:30 | |
| *** tonyb has quit IRC | 19:31 | |
| *** tonyb has joined #openstack-stable | 19:32 | |
| *** arif-ali has joined #openstack-stable | 19:35 | |
| *** dtantsur is now known as dtantsur|afk | 19:51 | |
| *** ihrachys has quit IRC | 20:04 | |
| *** rcernin has joined #openstack-stable | 20:12 | |
| *** eharney has quit IRC | 21:07 | |
| *** armax has joined #openstack-stable | 21:10 | |
| *** rcernin has quit IRC | 21:41 | |
| *** clayton has quit IRC | 21:46 | |
| *** stevemar has quit IRC | 21:47 | |
| *** clayton has joined #openstack-stable | 21:47 | |
| *** catintheroof has quit IRC | 21:51 | |
| *** Rockyg has quit IRC | 21:55 | |
| *** mriedem has quit IRC | 22:03 | |
| *** apevec has quit IRC | 22:34 | |
| *** armax has quit IRC | 22:50 | |
| *** stevemar has joined #openstack-stable | 23:07 | |
| *** stevemar has quit IRC | 23:12 | |
| *** stevemar has joined #openstack-stable | 23:18 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!