| *** baoli has joined #openstack-sprint | 00:04 | |
| *** baoli_ has joined #openstack-sprint | 00:05 | |
| *** baoli has quit IRC | 00:09 | |
| *** rfolco has joined #openstack-sprint | 00:33 | |
| *** delattec has quit IRC | 00:56 | |
| *** baoli_ has quit IRC | 01:24 | |
| *** rfolco has quit IRC | 01:28 | |
| *** baoli has joined #openstack-sprint | 02:34 | |
| *** baoli has quit IRC | 02:40 | |
| *** crinkle_ has joined #openstack-sprint | 04:16 | |
| *** crinkle has quit IRC | 04:17 | |
| *** hashar has joined #openstack-sprint | 07:44 | |
| *** mestery has quit IRC | 10:46 | |
| *** baoli has joined #openstack-sprint | 11:37 | |
| *** baoli_ has joined #openstack-sprint | 11:46 | |
| *** rfolco has joined #openstack-sprint | 11:47 | |
| *** baoli has quit IRC | 11:49 | |
| *** larainema has quit IRC | 11:58 | |
| *** larainema has joined #openstack-sprint | 12:02 | |
| *** hashar has quit IRC | 12:25 | |
| *** hashar has joined #openstack-sprint | 12:26 | |
| *** mestery has joined #openstack-sprint | 12:52 | |
| *** rfolco has quit IRC | 14:03 | |
| *** rfolco has joined #openstack-sprint | 14:04 | |
| *** rfolco has quit IRC | 14:11 | |
| *** rfolco has joined #openstack-sprint | 14:12 | |
| *** hashar has quit IRC | 16:09 | |
| *** ianw has quit IRC | 16:13 | |
| *** ianw has joined #openstack-sprint | 16:43 | |
| zaro | online reindex testing is going on now https://etherpad.openstack.org/p/gerrit-online-index-testing | 17:01 |
|---|---|---|
| zaro | just started the online index and it should run for about 1.5 hrs. most projects are small and one really large project "kdc/C". please push changes to kdc/C since reindex will be running against that the longest. | 17:03 |
| Zara | okay, can probably give 5 mins or so | 17:05 |
| Zara | it's taking a long time to load, idk if that's expected or not | 17:08 |
| zaro | do you mean review-dev? | 17:09 |
| Zara | yeah | 17:09 |
| zaro | is it only slow on first load or is it continously slow? | 17:10 |
| zaro | UI is not slow for me. how about you pleia2 ? | 17:12 |
| pleia2 | not so far, but it's saying "Working ..." at the top always | 17:13 |
| pleia2 | (on a call, just fiddling until I have to talk) | 17:13 |
| Zara | yeah, same here re: 'working'. it was continuously slow for a bit, now it's not; not sure if it's my connection | 17:14 |
| zaro | CPU utilization is 250% | 17:16 |
| zaro | but no errors in logs or anything. | 17:17 |
| pleia2 | pushed a couple patchsets without trouble, edited my user info, added comments and votes | 17:19 |
| pleia2 | I'll queue up and push a few at once after this call | 17:19 |
| Zara | pushed a patchset and published edit to it on the ui without a problem | 17:20 |
| Zara | well, a patch, not a patchset | 17:20 |
| pleia2 | Zara: hahaha, I did patch about cats | 17:21 |
| Zara | :D | 17:22 |
| Zara | though hm, the only patches that show up in that project for me list zaro or me as owner | 17:23 |
| Zara | I wonder if that's an issue or if I'm missing something. would guess it's the latter. | 17:23 |
| pleia2 | on https://review-dev.openstack.org/#/q/status:open ? | 17:23 |
| pleia2 | oh, I picked a different project | 17:23 |
| pleia2 | gtest-org/gtest | 17:23 |
| Zara | oh, aha | 17:23 |
| Zara | I approve and have +1'd it. | 17:24 |
| zaro | you can use git review on kdc/C project | 17:26 |
| pleia2 | zaro: do you know what "Working..." in the UI means? Just an indication that things are a bit slow? | 17:26 |
| pleia2 | it's been coming and going as I wander through the UI | 17:26 |
| pleia2 | oh, I should pay attention to this call now, sec | 17:26 |
| zaro | yes, i believe it's the CPU is loaded and is trying to catch up. | 17:28 |
| zaro | i'm guessing http threads can use a little tweeking but luckly we've already done that on review.o.o | 17:28 |
| Zara | hah, related changes only just showed up for a patch I sent | 17:29 |
| Zara | so I think that's something that's affected. | 17:30 |
| zaro | good eye, i didn't notice that. | 17:30 |
| zaro | i don't see related changes at all. when do you get it? | 17:31 |
| Zara | it took a while, maybe a minute or more | 17:32 |
| Zara | it's gone again, maybe because I uploaded a new patch on that topic | 17:33 |
| zaro | i saw an error on query for topic:frogs. did that work? | 17:33 |
| Zara | the topic was frog | 17:33 |
| Zara | aha just appeared | 17:33 |
| zaro | ohh that's super slow | 17:33 |
| zaro | that makes sense i guess it's indexing while changes are getting added | 17:35 |
| Zara | hm, just tried rebasing something and pushing again and it failed, not yet sure whether it was an issue on my end | 17:38 |
| zaro | did you do that from UI? or command line? | 17:38 |
| Zara | command line | 17:39 |
| Zara | will paste error in a sec | 17:39 |
| Zara | http://paste.openstack.org/show/565258/ | 17:40 |
| Zara | that's make change, add, commit, git review, then make another change, add, commit, git rebase -i, squash so it's all one commit, git review | 17:40 |
| Zara | `git pull` also seems to take longer, though I don't have any useful measurements for that | 17:43 |
| zaro | yeah, it's slow. reindex is overtaking the CPU | 17:43 |
| zaro | but the rebase should work. let me try | 17:43 |
| Zara | tried again, same thing | 17:46 |
| zaro | it works for me | 17:46 |
| zaro | which change are you rebasing? | 17:47 |
| Zara | 107941 and 107937 | 17:47 |
| Zara | (not at the same time, haha) | 17:47 |
| Zara | it's the same workflow I use for storyboard things so I'm fairly confident I'm not doing anything too strange. | 17:48 |
| zaro | 107941 does not require a rebase. it's the same as master | 17:50 |
| zaro | i mean it's up to date with master | 17:50 |
| zaro | actually what i mean is that there's no conflict with master | 17:51 |
| Zara | yeah, I mean rebase to upload a new version of the patchset rather than to fix a merge conflict | 17:51 |
| Zara | http://paste.openstack.org/show/565263/ is in more detail, though it doesn't show the actual rebase step (which is just squash on the second change and then commenting out everything in it) | 17:52 |
| pleia2 | I am definitely noticing slowness here and there, but nothing untolerable, especially since an alert from the bot will exist during this time when we do it in production | 17:52 |
| zaro | Zara: so the rebase then push new patchset works for me with 107941 | 17:53 |
| zaro | Zara: let me try your scenario | 17:54 |
| zaro | Zara: hmm, just tried your scenario and it works for me | 17:58 |
| Zara | gah | 17:58 |
| zaro | git review should ask you to upload a series. did it do that? | 17:59 |
| Zara | no, but this should just be one commit | 17:59 |
| Zara | in this case I want it to put another file into the same commit | 18:00 |
| zaro | you do that with squash not a rebase. | 18:00 |
| Zara | yeah, I do an interactive rebase and then choose to squash it | 18:01 |
| zaro | ahh, i did not do that | 18:01 |
| zaro | you paste just said rebase -i master but hid what you actually did in the rebase | 18:01 |
| Zara | yeah, sorry, I mentioned it here but it wasn't too clear | 18:02 |
| Zara | this workflow definitely works for me in production but I don't know if this is reindex-specific or a more general difference with review-dev | 18:03 |
| Zara | will experiment a bit more | 18:04 |
| Zara | hm, okay. it seems it doesn't like it if you leave the original file untouched and squash in a new file. but if you edit the original file, and then squash that change, it's fine with it. | 18:07 |
| *** larainema has quit IRC | 18:08 | |
| Zara | I did that test after reading about the error here: http://stackoverflow.com/questions/16586642/git-unpack-error-on-push-to-gerrit/20329880#20329880 | 18:08 |
| zaro | use the scenario you described of squashing inside the rebase and it worked for me as well. | 18:09 |
| Zara | was it a squash where the original files weren't edited? | 18:10 |
| zaro | yes | 18:10 |
| zaro | https://review-dev.openstack.org/#/c/107948/ | 18:10 |
| zaro | but i used blank files though | 18:10 |
| zaro | maybe that's the difference | 18:10 |
| Zara | that only lists one patch set for me. | 18:11 |
| Zara | it worked for me here: https://review-dev.openstack.org/#/c/107946/ | 18:11 |
| Zara | but yeah, editing the same file | 18:11 |
| *** larainema has joined #openstack-sprint | 18:11 | |
| Zara | well this is weird but I'm doubting it's reindex specific | 18:12 |
| pleia2 | running out of things to test | 18:12 |
| zaro | Zara: yeah, doesn't seem to be due to reindexing | 18:13 |
| zaro | pleia2: have you done more complicated queries? | 18:13 |
| pleia2 | zaro: such as? | 18:14 |
| zaro | nothing in logs about indexing errors anyays | 18:14 |
| pleia2 | I did the things in the etherpad | 18:15 |
| zaro | pleia2: good question. let me take a look at gerrit-dash-creator for ideas :) | 18:15 |
| pleia2 | oh, queries like search queries | 18:15 |
| pleia2 | on it | 18:15 |
| zaro | ohh yeah tons of examples in https://github.com/openstack/gerrit-dash-creator/tree/master/dashboards | 18:16 |
| zaro | oops times up i guess. about 15 minutes less than i anticipated. | 18:17 |
| Zara | heh | 18:19 |
| zaro | i think that went well. | 18:19 |
| zaro | thanks for your help. now i'm looking at logs | 18:19 |
| zaro | some errors concerning draft changes, which we don't support. | 18:19 |
| Zara | (fwiw I just tried my failing rebase workflow on sandbox and it worked fine so it's defo a difference between dev and prod.) | 18:20 |
| zaro | for some reason got a db disconnect which seems kinda unexpected due to our activity. not sure what caused that. | 18:20 |
| zaro | but good to see gerrit continue to soldier on though. | 18:20 |
| zaro | Zara: try on review-dev again? reindex is no longer running now | 18:21 |
| Zara | okay | 18:21 |
| zaro | [2016-08-31 18:16:17,300] INFO com.google.gerrit.lucene.OnlineReindexer : Reindex to version 14 complete | 18:22 |
| zaro | [2016-08-31 18:16:17,300] INFO com.google.gerrit.lucene.OnlineReindexer : Using schema version 14 | 18:22 |
| pleia2 | :D | 18:22 |
| zaro | pleia2: ^ this is what to watch for | 18:22 |
| pleia2 | zaro: which log? | 18:22 |
| zaro | error_log | 18:22 |
| pleia2 | okie | 18:22 |
| Zara | okay, still not working, defo not reindex-specific :) | 18:23 |
| zaro | cool. | 18:23 |
| zaro | [2016-08-31 18:23:08,482] ERROR com.google.gerrit.sshd.BaseCommand : Internal server error (user Zara account 5421) during git-receive-pack '/kdc/C.git' | 18:26 |
| zaro | com.google.gerrit.sshd.BaseCommand$Failure: fatal: Unpack error, check server log | 18:26 |
| zaro | at com.google.gerrit.sshd.commands.Receive.runImpl(Receive.java:176) | 18:26 |
| zaro | at com.google.gerrit.sshd.AbstractGitCommand.service(AbstractGitCommand.java:101) | 18:26 |
| zaro | at com.google.gerrit.sshd.AbstractGitCommand.access$000(AbstractGitCommand.java:32) | 18:26 |
| zaro | at com.google.gerrit.sshd.AbstractGitCommand$1.run(AbstractGitCommand.java:70) | 18:26 |
| zaro | at com.google.gerrit.sshd.BaseCommand$TaskThunk.run(BaseCommand.java:445) | 18:26 |
| *** zaro has quit IRC | 18:26 | |
| pleia2 | bye zaro | 18:27 |
| *** zaro has joined #openstack-sprint | 18:27 | |
| pleia2 | zaro: paste bin? :) | 18:27 |
| zaro | opps. didn't mean to do that. | 18:27 |
| pleia2 | hehe | 18:27 |
| zaro | Zara: http://paste.openstack.org/show/565276/ | 18:27 |
| zaro | that's the error for your push | 18:28 |
| Zara | haha | 18:28 |
| zaro | it has nothing to do with online reindex | 18:28 |
| pleia2 | breaking other things, go Zara :D | 18:29 |
| Zara | this is why SotK hates me. | 18:29 |
| pleia2 | lol | 18:29 |
| zaro | could be that we updated gerrit.war on review.o.o without updating the plugins | 18:29 |
| pleia2 | zaro: so is there something we should have looked at to make sure the reindex did what we wanted? like a change of a project name, updating anything, or anything? | 18:29 |
| zaro | because i think that's the hooks plugin work | 18:29 |
| Zara | it seems like a strange edge-case that doesn't affect the production gerrit. | 18:30 |
| pleia2 | zaro: also, will you be around on Friday? | 18:30 |
| zaro | would need to investigate further but might be a bug in the hook. | 18:31 |
| zaro | ironically, i think the plugins on review-dev are definately more correct than the ones on review.o.o | 18:31 |
| zaro | pleia2: yes | 18:31 |
| zaro | pleia2: the reindex will reindex projects in i think alpha order | 18:32 |
| Zara | haha, I guess now we know it might bite us later. | 18:32 |
| zaro | so if the project we've changed the name to gets reindexed before the indexer finished you'll see the new project on gerrit even before the index finishes | 18:32 |
| pleia2 | zaro: ok, noted | 18:33 |
| zaro | just depends on when the project gets reindexed | 18:33 |
| zaro | nice work guys and thanks a bunch for volunteering! | 18:33 |
| Zara | np :) | 18:34 |
| * pleia2 considers loading up colornicks if she has to work with zaro and Zara a ot | 18:35 | |
| pleia2 | a lot | 18:35 |
| pleia2 | my poor brain | 18:35 |
| Zara | bwahaha | 18:38 |
| Zara | that's how I'll justify my capital letter | 18:38 |
| zaro | i thought everyone does colored nicks? i would go crazy if i didn't have it | 18:40 |
| Zara | hah, I don't use them either | 18:43 |
| Zara | though most of my lines go ':) :) :) hi !!!' so they're fairly easy to spot | 18:44 |
| *** rfolco has quit IRC | 19:22 | |
| *** rfolco has joined #openstack-sprint | 19:23 | |
| *** rfolco has quit IRC | 19:43 | |
| *** rfolco has joined #openstack-sprint | 19:44 | |
| *** rfolco has quit IRC | 20:04 | |
| *** rfolco has joined #openstack-sprint | 20:05 | |
| *** rfolco has quit IRC | 21:00 | |
| *** baoli_ has quit IRC | 21:02 | |
| *** baoli has joined #openstack-sprint | 21:03 | |
| *** hieulq has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!