| *** krtaylor has joined #openstack-sprint | 00:35 | |
| omrim | krtaylor: Hi | 10:01 |
|---|---|---|
| *** rfolco has joined #openstack-sprint | 11:09 | |
| fungi | i still need to get some breakfast, but will be ready to pitch in shortly | 13:00 |
| mmedvede | hi all! | 14:35 |
| fungi | howdy mmedvede! | 14:37 |
| fungi | i'm starting in on reviewing the https://review.openstack.org/#/q/topic:+module-split+status:open,n,z changes. if anyone has sprint-related changes they want reviewed make sure to use the "module-split" topic | 14:38 |
| mmedvede | fungi: This one could require special treatment https://review.openstack.org/#/c/150713/ | 14:38 |
| fungi | also reminder to those who are lurking, we're keeping notes at https://etherpad.openstack.org/p/puppet-module-split-sprint | 14:39 |
| mmedvede | basically, zuul project is already there, so that would prevent it from getting updated from new upstream | 14:39 |
| fungi | mmedvede: thanks for pointing that out. i'll double-check and yes, we'll need to think about whether it makes sense to push --force the new repo on top of that or propose the changes from it | 14:41 |
| fungi | on a positive note, it looks like the initial commit there is the one manage-projects creates, so i think it's no real loss if we overwrite | 14:42 |
| fungi | seems we added it back in june | 14:42 |
| mmedvede | fungi: yes, the repository is rather empty | 14:43 |
| fungi | and no changes in gerrit, open or closed | 14:43 |
| fungi | also 150713 isn't really necessary. that change won't do anything... the import work is going to end up being entirely manual for it | 14:44 |
| fungi | so no need to add that line to the projects.yaml file | 14:44 |
| fungi | i'll go ahead and take care of it once we have an additional person look over the github.com/Triniplex-Test/puppet-zuul contents to confirm | 14:45 |
| mmedvede | fungi: understood. I would mark it WIP, and abandon later. Probably would need to remove Depends-On from https://review.openstack.org/#/c/150712/ | 14:46 |
| mmedvede | fungi: Triniplex-Test repositories are not updated automatically. But I have push rights to it, so I have updated the ones needed for my patches | 14:47 |
| fungi | mmedvede: awesome. i'm going through it now just to make sure it's what i expect | 14:48 |
| mmedvede | just in case somebody thinks they are automatically updated | 14:48 |
| fungi | mmedvede: lgtm. let's get a second reviewer to make a pass through that repo and then i'll push it in over top the old one | 14:50 |
| mmedvede | ok | 14:52 |
| *** mestery has quit IRC | 15:01 | |
| *** dteselkin has joined #openstack-sprint | 15:04 | |
| krtaylor | hi everyone | 15:12 |
| *** SergeyLukjanov has joined #openstack-sprint | 15:13 | |
| SergeyLukjanov | hey folks | 15:13 |
| jesusaurus | hello | 15:13 |
| mmedvede | hi | 15:13 |
| clarkb | morning though I am not quite ready for the day to start | 15:14 |
| krtaylor | I'll start doing some reviews -> https://review.openstack.org/#/q/topic:module-split,n,z | 15:15 |
| *** jeblair has joined #openstack-sprint | 15:18 | |
| asselin_ | good morning | 15:29 |
| mmedvede | asselin_: morning | 15:30 |
| asselin_ | sweston should have the triniplex modules being updated. Either way, we should double check all repos | 15:32 |
| mmedvede | asselin_: I do not think Triniplex-Test repos are updated | 15:33 |
| mmedvede | asselin_: I have been pushing for my patches there though | 15:33 |
| mmedvede | asselin_, sweston: I have noticed some modules on Triniplex-Test are definitely not right, e.g. https://github.com/Triniplex-Test/puppet-lodgeit, I had force-pushed a couple | 15:35 |
| * asselin_ looking | 15:36 | |
| *** mrmartin has joined #openstack-sprint | 15:38 | |
| mrmartin | o/ | 15:38 |
| asselin_ | mmedvede, I don't see lodgeit here: https://github.com/openstack-infra/system-config/tree/master/modules | 15:38 |
| mmedvede | asselin_: it has been merged/split out already | 15:39 |
| asselin_ | mmedvede, today? | 15:39 |
| * asselin_ reviewing latest patches | 15:40 | |
| mmedvede | asselin_: no. Lodgeit did not use Triniplex-Test upstream | 15:40 |
| mmedvede | so it was not a problem | 15:41 |
| asselin_ | ok | 15:41 |
| mrmartin | asselin_: what's the next step with this review: https://review.openstack.org/#/c/149953/ ? | 15:44 |
| asselin_ | mrmartin, another core will +2 & +A | 15:45 |
| mrmartin | ok | 15:45 |
| asselin_ | mrmartin, then we need to remove it from system-config | 15:45 |
| asselin_ | mrmartin, example: https://review.openstack.org/#/c/150814/ | 15:46 |
| mrmartin | but after approval, I need to add the .gitreview and module dependency files? | 15:46 |
| asselin_ | mrmartin, yes that too | 15:46 |
| mrmartin | ok, then I'm waiting :) | 15:46 |
| asselin_ | mrmartin, you should have a patch like this to remove from system-config: https://review.openstack.org/#/c/150814/ | 15:47 |
| mrmartin | ok, then I'm writing those patches | 15:47 |
| asselin_ | mrmartin, thanks | 15:47 |
| fungi | okay, so did anyone have any specific thoughts on how (or whether) we need to plan for merge conflicts? i guess at this point they're merely changes to the sorted projects.yaml list but since the entries have a common prefix they're often going to be in context proximity anyway | 15:50 |
| fungi | i'm inclined to just upload rebases as needed while i'm reviewing | 15:50 |
| asselin_ | fungi, ++ | 15:51 |
| fungi | about to approve 149994 which will create openstack-infra/puppet-openstackid so wanted to make sure this is coordinated as much (or as little) as needed | 15:51 |
| asselin_ | looking | 15:51 |
| asselin_ | fungi, looks good to me | 15:52 |
| asselin_ | fungi, do you think merging non-alphabetical will help reduce merge conflicts? | 15:54 |
| asselin_ | also, should we try to keep https://review.openstack.org/#/c/150712/2/modules.env alphabetical? | 15:55 |
| mmedvede | asselin_: that is what I am trying to do | 15:56 |
| mmedvede | asselin_: should reduce probability of merge conflicts | 15:56 |
| asselin_ | the file is already non-alpha, perhaps we shoudl fix that first? | 15:56 |
| fungi | asselin_: not sure what you mean by merging non-alphabetical | 15:57 |
| fungi | what is non-alpha about that file? we have a job that checks it stays in alpha order | 15:57 |
| asselin_ | fungi, e.g. pick something in the first half of the alphabet, then the second half of the alphabet, then back to the first. | 15:58 |
| jesusaurus | doesnt order matter in modules.env because of possible dependencies? | 15:58 |
| fungi | asselin_: oh, sorry, i was talking about projects.yaml | 15:58 |
| fungi | now i see you were not | 15:58 |
| asselin_ | fungi, https://github.com/openstack-infra/system-config/blob/master/modules.env#L63 | 15:58 |
| fungi | yeah | 15:59 |
| asselin_ | jesusaurus, good question.... | 15:59 |
| jeblair | hi there, i'm here now | 15:59 |
| fungi | i'm not sure if it will help anyway. the number of rebases will likely be roughly the same regardless | 15:59 |
| jeblair | where's our etherpad? | 15:59 |
| asselin_ | jeblair, does order matter here: https://github.com/openstack-infra/system-config/blob/master/modules.env#L63 | 15:59 |
| asselin_ | jeblair, https://etherpad.openstack.org/p/puppet-module-split-sprint | 16:00 |
| fungi | we should just avoid rebasing the ones which start spontaneously merge-conflicting with one another until we're ready to approve them, so that we don't rebase them multiple times unnecessarily | 16:00 |
| asselin_ | fungi, ++ | 16:00 |
| jeblair | asselin_: i don't think order matters (it may have an effect, but shouldn't _matter_) | 16:00 |
| fungi | okay, fire in the hole. 149994 is approved | 16:01 |
| krtaylor | hehheh | 16:03 |
| krtaylor | I just pushed bup | 16:03 |
| fungi | did anybody else get a chance to look over github.com/Triniplex-Test/puppet-zuul ? short story there is that we've had a dead (initial manage-projects commit, no changes proposed ever) openstack-infra/puppet-zuul since june, so i'm planning to push --force that split content if it looks okay | 16:03 |
| * krtaylor looks | 16:03 | |
| asselin_ | fungi, I'll double check now | 16:03 |
| jeblair | okay, what do i do? pick a change ready to go and approve it? | 16:04 |
| mmedvede | fungi: I have rechecked, diff with current system-config shows no differences | 16:04 |
| asselin_ | fungi, matches: https://github.com/Triniplex-Test/puppet-zuul/commits/master https://github.com/openstack-infra/system-config/commits/master/modules/zuul | 16:04 |
| fungi | jeblair: that's what i've been doing. approve them all and let the gate sort them out ;) | 16:04 |
| krtaylor | re: puppet-zuul lgtm | 16:05 |
| clarkb | should we leave a note in -infra that system-config and project-config should be treated as frozen today while we do this? | 16:05 |
| fungi | clarkb: oh, good idea | 16:05 |
| krtaylor | +1 | 16:05 |
| mmedvede | clarkb: +1 | 16:05 |
| asselin_ | clarkb, +1 | 16:05 |
| fungi | s/frozen/ignored/ if you want the truth | 16:05 |
| asselin_ | fungi, :) | 16:06 |
| fungi | since we're all focused on reviewing sprint-related changes | 16:06 |
| *** reed has joined #openstack-sprint | 16:06 | |
| jeblair | are people writing removal changes after the create changes land? | 16:07 |
| fungi | mrmartin: thanks for 150831. i'll +2 it once i confirm creation of the new project worked | 16:07 |
| jeblair | oh, nm | 16:08 |
| jeblair | i had a bad query | 16:08 |
| asselin_ | jeblair, there are quite a few removal changes | 16:08 |
| jeblair | https://review.openstack.org/#/q/status:open+branch:master+topic:module-split,n,z | 16:08 |
| fungi | i think it should be safe to write them before the corresponding new project changes are approved | 16:08 |
| jeblair | is what i really want | 16:08 |
| fungi | ahh, yep | 16:08 |
| fungi | the query in the etherpad is a little overly-specific | 16:08 |
| jeblair | fixing | 16:08 |
| jeblair | oh, no, the ep is right, gerrit just was being helpful as i clicked around | 16:09 |
| fungi | ahh, heh | 16:09 |
| fungi | and yeah, i checked the etherpad after i said that. seems right | 16:09 |
| fungi | morningbrain | 16:09 |
| *** mestery has joined #openstack-sprint | 16:10 | |
| krtaylor | the description, why? is it needed? it seems worthless (just a restating of the project) | 16:11 |
| jeblair | krtaylor: it shows up in github (and possibly other places in the future) | 16:11 |
| krtaylor | jeblair, ah, ok | 16:11 |
| asselin_ | mrmartin, actually, can you update the commit messages a bit for yours? to include the removal patch? https://review.openstack.org/#/c/149953/ | 16:12 |
| mrmartin | asselin_: I can | 16:13 |
| asselin_ | mrmartin, like this one: https://review.openstack.org/#/c/150699/ | 16:13 |
| asselin_ | mrmartin, the two related patches should link to each other | 16:13 |
| asselin_ | mrmartin, thanks | 16:13 |
| mrmartin | Would you like to see just the Needed-By line? | 16:13 |
| asselin_ | mrmartin, I'm ok with that since today's the split | 16:14 |
| asselin_ | split sprint | 16:14 |
| mrmartin | but it will lost all of the +1's :( | 16:14 |
| jeblair | we could probably waive that today... | 16:15 |
| jeblair | if folks wanted to... since we're focused entirely on this | 16:15 |
| mrmartin | can I rewrite this on gerrit ui? | 16:15 |
| jeblair | mrmartin: yes | 16:15 |
| asselin_ | mrmartin, seems you can update the commit message in gerrit ui. didn't realize that | 16:17 |
| mrmartin | asselin_: yeap, it creates a new patches automagically | 16:17 |
| mrmartin | patches / patchset | 16:17 |
| fungi | i'm not at all concerned about needed-by, and am in fact double-checking regardless of depends-on headers since these follow an obvious pattern | 16:18 |
| jeblair | fungi, clarkb: erm, should we just kill recheckwatch? | 16:19 |
| clarkb | wfm | 16:19 |
| fungi | jeblair: i'd be fine with that. i never got any indication that it was every used by anyone | 16:19 |
| jeblair | including, at this point, us. | 16:19 |
| fungi | i've pointed out the url to the rss feeds periodically, but never seen anyone else mention it | 16:19 |
| jeblair | it has rss feeds? | 16:20 |
| fungi | oh, recheckwatch | 16:20 |
| fungi | sorry, i was thinking of openstackwatch | 16:20 |
| fungi | duh | 16:20 |
| fungi | yeah, totally safe to ditch recheckwatch | 16:20 |
| asselin_ | ok I see that's mine. :) | 16:22 |
| jeblair | asselin_: sorry i didn't spot that sooner :( | 16:23 |
| mrmartin | are the gate puppet apply FAILURES normal in that case? https://review.openstack.org/#/c/150831/ | 16:23 |
| asselin_ | jeblair, no problem | 16:23 |
| fungi | mrmartin: yes | 16:24 |
| mrmartin | ok. | 16:24 |
| fungi | mrmartin: they fail until the new project exists. i just rechecked it now that it does | 16:24 |
| jeblair | how about i stack some changes that have merge conflicts together and re-upload? | 16:25 |
| fungi | okay, i'm going to go ahead and push --force github.com/Triniplex-Test/puppet-zuul into gerrit now unless there are any objections or anyone else wants to look it over first | 16:26 |
| jeblair | fungi: ++ | 16:26 |
| fungi | jeblair: that also seems fine | 16:26 |
| asselin_ | jeblair, ok, updated the system config change to just delete it altogether: https://review.openstack.org/#/c/140431/ | 16:26 |
| fungi | asselin_: thanks! | 16:26 |
| jeblair | 16:28 < openstackgerrit> James E. Blair proposed openstack-infra/project-config: Add puppet-logstash as split out module https://review.openstack.org/150706 | 16:28 |
| jeblair | 16:28 < openstackgerrit> James E. Blair proposed openstack-infra/project-config: Add puppet-nodepool as split out module https://review.openstack.org/150714 | 16:28 |
| jeblair | 16:28 < openstackgerrit> James E. Blair proposed openstack-infra/project-config: Add puppet-meetbot as split out module https://review.openstack.org/150699 | 16:28 |
| jeblair | could folks re-review those real quick? | 16:28 |
| asselin_ | all I have a 30 minute meeting now. will be back soon | 16:30 |
| fungi | sure | 16:30 |
| jeblair | clarkb: please review my patches! ^ :) | 16:31 |
| jeblair | (well, they are not mine, but my rebases of them) | 16:31 |
| clarkb | are all of the triniplex repos up to date? | 16:32 |
| fungi | also, i think any non-urgent patches to those modules still sitting in system-config should just get -2'd when we're done and ask the change owners to resubmit to the appropriate new project | 16:32 |
| jeblair | fungi: ++ | 16:32 |
| fungi | sort of like we did with the project-config split | 16:32 |
| jeblair | or possibly even abandoned | 16:33 |
| mmedvede | clarkb: no | 16:33 |
| jeblair | oh neat! | 16:33 |
| mmedvede | clarkb: Triniplex-Test are not all up to date | 16:33 |
| jeblair | i thought this was all scripted and kept up to date :/ | 16:33 |
| mmedvede | jeblair: it used to. I can push there, so I did update some of the modules there to use for patches I submitted | 16:34 |
| mmedvede | We have to check them manually, unfortunately | 16:35 |
| jeblair | mmedvede: okay, so you're in charge of that for now. :) | 16:35 |
| jeblair | mmedvede: can you comment on those 3 patches and verify that their repos are up to date? | 16:35 |
| mmedvede | jeblair: on it | 16:36 |
| mmedvede | thanks for the rebase | 16:36 |
| clarkb | so I should hold off on +2'ing anything? | 16:39 |
| jeblair | clarkb: sounds like it.... i think we should establish some process here so we can get through these quickly. i'd like to stop doing this after today. :) | 16:40 |
| jeblair | so it seems like we need folks to (a) create temp split repos; (b) write split changes; (c) write removal changes; (d) write boilerplate puppet/gitreview changes; (e) verify the contents and up-to-date-ness of split repos; (f) review changes | 16:42 |
| jeblair | it seems like we're set up to have one person do a,b,c,d for each repo by grabbing a task in storyboard. shall we stick with that? | 16:43 |
| clarkb | I think so, it keeps units of work coupled so we don't miss a step for one project | 16:44 |
| jeblair | i sort of thought someone was doing (a) for all the repos, but i guess not | 16:44 |
| fungi | i thought the ones asselin_/krtaylor did were being maintained that way | 16:45 |
| fungi | using sweston's script in cron or something | 16:46 |
| jeblair | it would be a big help... but if that's the case, i don't understand why we have changes pointing to repos other than that, that may or may not be up to date :( | 16:46 |
| jeblair | clarkb: mmedvede have verified that those 3 repos are up to date, feel free to review those changes | 16:47 |
| sweston | sweston's ears perk up | 16:47 |
| clarkb | doing that now | 16:47 |
| jeblair | sweston: i thought there was a bunch of github repos with all of our puppet modules split out and being kept up to date. do you know if that's the case? | 16:48 |
| sweston | yes, so the last time I wrote a patch for asselin I stopped the script so that I could troubleshoot. I need to restart it, it has not run in a while | 16:49 |
| mmedvede | sweston: I have forced pushed a couple of repos there | 16:49 |
| jeblair | sweston: any chance you can get that going now, or should we just switch to having folks run things on their own? | 16:49 |
| sweston | jeblair: nope, I can do it in about 5 minutes | 16:50 |
| sweston | jeblair: are you doing the split today? | 16:50 |
| clarkb | I need to pop out to drop larissa off at work back in a bit (I did review that stack jeblair rebased) | 16:50 |
| jeblair | sweston: yes, today is our sprint, and we're going to get all of the modules split today. | 16:50 |
| sweston | jeblair: ok, I will make myself available today, and re-start the script now | 16:51 |
| sweston | jeblair: is it the 28th already? ;-) | 16:51 |
| mmedvede | sweston: check this https://github.com/Triniplex-Test/puppet-lodgeit | 16:51 |
| mmedvede | before restarting script | 16:52 |
| jeblair | sweston: great, thanks! where are the modules? | 16:52 |
| mmedvede | sweston: it is not correct, and there are several more repositories like that one | 16:52 |
| fungi | mrmartin: were you going to propose an initial change to openstack-infra/puppet-openstackid (adding .gitreview, rakefile and puppet metadata) now that it's created? | 16:52 |
| krtaylor | I used my github repo because I know (and verified) that is has not changed | 16:52 |
| sweston | jeblair: they are in the triniplex test repo mmedvede referenced | 16:53 |
| mrmartin | fungi: yeap I'm doing that | 16:53 |
| sweston | mmedvede: besides being out of date, is there anything else wrong with the repo? | 16:53 |
| mrmartin | only one question here, it must depend on puppet-redis module, how can I add this dependency to puppet metadata? | 16:53 |
| fungi | mrmartin: great. presumably we want that before we approve 150831 | 16:53 |
| jeblair | nibalizer, jesusaurus: ^ question from mrmartin | 16:54 |
| mmedvede | sweston: yes, it is not a split repo. Check the contents | 16:54 |
| mrmartin | so I guess, need to approve the puppet redis patch too before this | 16:54 |
| mmedvede | sweston: it is supposed to have only lodgeit, but it has all the modules | 16:54 |
| sweston | mmedvede: yup, you are correct, this may be a bug | 16:55 |
| sweston | mmedvede: give me a moment to restart the split script, and we will troubleshoot further | 16:55 |
| fungi | mrmartin: i'm looking at the redis one now | 16:55 |
| mmedvede | sweston: you are not force pushing to Triniplex-Test, correct? because if you do, please do not enable the script, it would rewrite good repositories that I have updated | 16:56 |
| sweston | mmedvede: I can push to Triniplex instead | 16:56 |
| jeblair | sweston, mmedvede: and we just approved 3 changes that clone from triniplex-test, so we should keep meetbot, logstash, and nodepool stable there for a while | 16:57 |
| mmedvede | jeblair: +1 | 16:58 |
| jeblair | fungi, mrmartin: i have +2d redis,askbot,drupal; did not aprv since fungi is looking | 16:58 |
| sweston | jeblair, mmedvede: ok, will do. I will run the script against Triniplex instead. Sound good? | 16:58 |
| mrmartin | jeblair: ok, we need the redis first, to finalize openstackid too | 16:58 |
| jeblair | fungi: what did you approve earlier? | 16:59 |
| fungi | jeblair: puppet-openstackid | 16:59 |
| mmedvede | sweston: wfm, thank you. | 16:59 |
| fungi | AJaeger had already +2'd | 16:59 |
| jeblair | i'm keeping status at the bottom of the etherpad | 16:59 |
| fungi | and i double-checked the repo content was current | 16:59 |
| fungi | oh, great idea | 16:59 |
| nibalizer | o/ good morning | 17:00 |
| pleia2 | good morning | 17:01 |
| fungi | for those who got kicked off the etherpad, looks like it fell over again when everyone from the nova meetup was coming online. seems more than coincidence | 17:01 |
| jeblair | nibalizer, pleia2: sweston is working on getting his script that auto-pushes split modules to github going again | 17:02 |
| nibalizer | cool | 17:02 |
| pleia2 | great | 17:02 |
| jeblair | nibalizer, pleia2: probably the most helpful thing would be to claim some modules in https://storyboard.openstack.org/#!/story/302 and write changes for them, assuming that sweston will have up to date modules at https://github.com/Triniplex | 17:03 |
| jeblair | fungi: are you going to aprv askbot and drupal too? | 17:04 |
| pleia2 | sure thing, thanks jeblair | 17:05 |
| fungi | jeblair: yep, already looking at them | 17:05 |
| fungi | just getting the bookeeping straight | 17:05 |
| fungi | er bookkeeping (no keeping boo!) | 17:05 |
| asselin_ | Hi, I'm back | 17:06 |
| nibalizer | cool | 17:07 |
| fungi | it's worth noting that mrmartin's split repo for puppet-askbot already has the boilerplate added | 17:10 |
| fungi | so it won't in theory need a separate change for that | 17:10 |
| nibalizer | mrmartin: the dependencies section of this can be used as an example of how to set one module dependent on another https://github.com/puppet-community/puppet-module-puppetboard/blob/master/metadata.json | 17:11 |
| mrmartin | fungi: I can remove that if required | 17:11 |
| fungi | mrmartin: no need. technically that one's not a module "split" since it's a new project anyway | 17:11 |
| asselin_ | fungi, mrmartin is that really a split or a new one? | 17:11 |
| asselin_ | fungi, ++ | 17:11 |
| mrmartin | asselin_ :) good point | 17:11 |
| fungi | but might as well roll it into this | 17:11 |
| mrmartin | you need to roll in that anyway, so this is the best time :) | 17:12 |
| fungi | yep, not complaining just making sure everyone's aware it's a little special | 17:12 |
| mrmartin | nibalizer: how will you add a dependency, for example our puppet-redis module? | 17:13 |
| nibalizer | in that location put openstackci-redis | 17:14 |
| nibalizer | for versions idunno | 17:14 |
| * jeblair creates two new gertty dashboards | 17:15 | |
| mrmartin | ok, so it means, those modules will be uploaded to puppetforge under openstackci ? | 17:16 |
| jeblair | i pasted my gertty dashboard config into the etherpad if anyone is interested | 17:17 |
| fungi | saw that--thanks! | 17:17 |
| nibalizer | mrmartin: eventually, there is some stuff in flight for geting automated puppetforge deploys, i don't think thats in scope today though | 17:18 |
| jeblair | logstash/meetbot/nodepool are cloned now | 17:18 |
| mrmartin | ok, because if we add this dependency in metadata.json it won't harm us, because our modules are coming from git repo directly defined in modules.env, right? | 17:19 |
| sweston | jeblair, mmedvede: ok, script is re-started ... the first split takes about an hour, I will update you with progress as I have it | 17:19 |
| jeblair | sweston, mmedvede: we have imported the 3 repos in triniplex-test now, so i believe there are no approved changes depending on any triniplex repo now. you're all clear to do whatever you need | 17:20 |
| nibalizer | mrmartin: correct | 17:20 |
| sweston | jeblair: ok | 17:21 |
| fungi | http://git.openstack.org/cgit/openstack-infra/puppet-zuul is up to date now too, so adding it to the created list | 17:21 |
| mmedvede | sweston: awesome | 17:21 |
| jeblair | asselin_: can you write the puppet boilerplate changes for meetbot, logstash, and nodepool? | 17:22 |
| asselin_ | jeblair, sure | 17:22 |
| mmedvede | asselin_: I can add the boilerplate if you did not start on it yet | 17:26 |
| asselin_ | mmedvede, didn't start yet. just about to | 17:27 |
| mmedvede | asselin_: I have it all setup, should be quick for me | 17:27 |
| asselin_ | mmedvede, ok, you do it. I'll pick another module | 17:27 |
| jeblair | yay more modules! | 17:27 |
| pleia2 | nibalizer: can you set a topic of module-split on your changes? | 17:35 |
| nibalizer | sure | 17:36 |
| nibalizer | is the only way to do that cli to set the branch tite? | 17:36 |
| nibalizer | er branch title | 17:36 |
| jeblair | pleia2, nibalizer: alse recall gerrit lets you set the topic through the ui | 17:36 |
| jeblair | (gertty too) | 17:36 |
| jeblair | nibalizer: and in the future, you can do "git review -t module-split" so it is set when you upload regardless of branch name | 17:36 |
| pleia2 | have to be the owner of the change to set it though | 17:37 |
| pleia2 | (re: gerrit ui) | 17:37 |
| jeblair | pleia2: oh! | 17:37 |
| nibalizer | cool | 17:38 |
| jeblair | pleia2: i guess that must be another permission that has leaked into administrators... :( | 17:38 |
| pleia2 | jeblair: yeah, seems so, I can only fiddle with commit message and change-id on other people's patches | 17:38 |
| jeblair | failed: [git02.openstack.org] => {"changed": false, "disabled": false, "error": true, "rc": 1} | 17:39 |
| jeblair | stdout: Notice: Run of Puppet configuration client already in progress; skipping (/var/lib/puppet/state/agent_catalog_run.lock exists) | 17:39 |
| jeblair | fungi: ^ do you grok that? | 17:39 |
| nibalizer | pleia2: topic attached :) | 17:40 |
| jeblair | fungi: switching to #openstack-infra | 17:41 |
| asselin_ | jesusaurus, these don't quite match?? https://github.com/jesusaurus/puppet-cgit/commits/master https://github.com/openstack-infra/system-config/commits/master/modules/cgit | 17:42 |
| asselin_ | maybe it's ok? need more eys | 17:42 |
| pleia2 | nibalizer: fyi, looks like we're also using Depends-on: and Needed-by: in the commit messages to match up -system and -project patches that are related, I'm cribbing from asselin_'s patches, ie: https://review.openstack.org/#/c/150859/ & https://review.openstack.org/#/c/150860/ | 17:42 |
| nibalizer | ah okay | 17:43 |
| asselin_ | fungi, jeblair are we still doing the depends-on and needed-by? not sure if that's needed anymore? | 17:44 |
| jeblair | pleia2, asselin_: i think we can waive that for simplicity during the sprint | 17:48 |
| pleia2 | ah ok, good to know :) | 17:48 |
| asselin_ | nibalizer, ^^ | 17:49 |
| nibalizer | cool! | 17:49 |
| jeblair | is mostly for coordination under normal conditions, but now we've got etherpads with lists of things and we're really focused on it, so not strictly necessary | 17:49 |
| pleia2 | wfm | 17:49 |
| clarkb | I am making tea. will be back for reviews shortly | 17:50 |
| nibalizer | so, not to rock the boat, but the subversion module is 4 lines, and is only used once, so we could remove it entirely and replace the usage of it in wiki.pp with package { 'subversion': } | 17:50 |
| nibalizer | thoughts? | 17:50 |
| jeblair | nibalizer: ++ | 17:51 |
| clarkb | +1 | 17:51 |
| nibalizer | aight ill spin up that change real fast | 17:51 |
| nibalizer | PSA to others, be on the watch for trivial modules like subversion that can be just refactored into /dev/null | 17:53 |
| nibalizer | (git review -t)++ | 17:57 |
| jeblair | nibalizer: ++ we also nixed recheckwatch -- even though it is not trivial, it's not actually used (we just need to remove it) | 17:57 |
| nibalizer | heh | 17:57 |
| nibalizer | nicely done | 17:57 |
| jeblair | mmedvede, asselin_: what's the best way to verify that a repo is up to date? | 17:58 |
| asselin_ | jeblair, I use github: | 17:58 |
| mmedvede | jeblair: I was doing recursive diff between the repo in question and the latest checkout of system-config/modules/<module> | 17:58 |
| asselin_ | 1. open the new project in github & click on "commits" | 17:59 |
| mmedvede | jeblair: and also what asselin_ is saying, just making a quick look through commits | 17:59 |
| asselin_ | 2. https://github.com/openstack-infra/system-config/tree/master/modules, pick your module | 17:59 |
| asselin_ | then there's an 'circle like icon' on the top right | 18:00 |
| asselin_ | "browse commits for this branch" | 18:00 |
| asselin_ | 3. quick visual inspection of the 2 | 18:00 |
| jeblair | asselin_: makes sense. also "git diff ." in the local module dir should match the commit list for the new repo. | 18:01 |
| jeblair | mmedvede: i like the clone + recursive diff -- 2 commands and provides a conclusive negative quickly | 18:02 |
| jeblair | so maybe it's good to do the recursive diff first and double check with the commit list | 18:02 |
| jeblair | anyway, it looks like triniplex-test/puppet-gerrit is out of date by one commit | 18:02 |
| jeblair | so i -1d that change | 18:02 |
| mmedvede | jeblair: this is command ' diff -r -u -x.git <repo_in_question> system-config/modules/$module' | 18:02 |
| mmedvede | jeblair: checking puppet-gerrit | 18:03 |
| asselin_ | another 30 minute meeting. back soon | 18:03 |
| nibalizer | jeblair: im glad that you are checking the upstreams rigorously | 18:04 |
| jeblair | mmedvede: oh i may have been wrong about puppet-gerrit | 18:05 |
| jeblair | nibalizer: well, trying at any rate :) | 18:05 |
| mmedvede | jeblair: rechecked puppet-gerrit, seems fine | 18:05 |
| mmedvede | still up to date | 18:05 |
| jeblair | mmedvede: yep, i agree. sorry about that. so puppet-gerrit is ready | 18:06 |
| jeblair | fungi, clarkb: want to aprv? ^ | 18:06 |
| clarkb | jeblair: looking | 18:06 |
| mrmartin | this seems to be empty: https://git.openstack.org/cgit/openstack-infra/puppet-redis/ but this patchset seems to be passed: https://review.openstack.org/#/c/149953/ | 18:07 |
| jeblair | and yeah, anyone who wants to do the recursive-diff and commit list check to verify that repos are up to date, please do so, and leave a comment to that effect on changes | 18:07 |
| nibalizer | i'm setting my project-config changes to depend on each other | 18:07 |
| mrmartin | and it is not synced to github | 18:07 |
| nibalizer | that way its not rebase-city | 18:07 |
| jeblair | nibalizer: sounds good | 18:07 |
| clarkb | jeblair: done | 18:07 |
| jeblair | mrmartin: yeah, i believe that's because of the ansible bug we found (i moved the discussion of that over to infra) | 18:07 |
| nibalizer | itll still maybe be rebase city because we're all messing about int projects.yaml, but at least this way we have a chance | 18:08 |
| jeblair | mrmartin: i was hoping the next pass would fix it | 18:08 |
| mrmartin | am I need to do something? | 18:08 |
| clarkb | so I assume once we have a bunch of new projects we then go through the system-config changes and review/approve those to start using the new modules | 18:08 |
| jeblair | clarkb: yeah, keep status updated at the bottom of the etherpad: https://etherpad.openstack.org/p/puppet-module-split-sprint | 18:08 |
| clarkb | mrmartin: let me look | 18:08 |
| jeblair | clarkb: should basically just move projects down through that list | 18:08 |
| clarkb | jeblair: I see the list now thanks | 18:09 |
| fungi | mrmartin: i don't think so, we just need to double-check that it fixed itself (or manually intervene) | 18:09 |
| mrmartin | ok. | 18:09 |
| jeblair | fungi, clarkb, mrmartin: i think puppet is running there now | 18:09 |
| jeblair | let's see the result of this run | 18:09 |
| nibalizer | wow storyboard as a tool to quickly add people to subtasks is AMAZING | 18:11 |
| * mrmartin going to eat something, I'll look back later | 18:14 | |
| pleia2 | hm, are we using Triniplex-Test or Triniplex? | 18:16 |
| jeblair | pleia2: sweston is getting the script going on triniplex | 18:17 |
| jeblair | so aiui, we should write changes assuming that will be up to date (soonish) | 18:17 |
| pleia2 | ok | 18:17 |
| sweston | jeblair, pleia2, mmedvede: looks like first sync is done, and updates are running. I will look for inconsistencies in about half an hour, as soon as I am free, but feel free to help yourself in the meantime. | 18:18 |
| pleia2 | thanks sweston | 18:19 |
| sweston | pleia2: yw | 18:19 |
| jeblair | fungi, mrmartin: redis,drupal,askbot repos are created | 18:20 |
| mmedvede | sweston: thanks | 18:21 |
| sweston | mmedvede: yw | 18:21 |
| jeblair | ha! i don't know whether "puppet-elastic_recheck" comes before or after "puppet-elasticsearch" in our alphabetization check :) | 18:23 |
| * fungi hopes it's using posix (ascii byte value) sort order | 18:25 | |
| nibalizer | ooops, i've been doing triniplex-test | 18:28 |
| * nibalizer will update | 18:28 | |
| nibalizer | fungi: i wonder if puppet can handle a module called elastic\arecheck :D | 18:29 |
| jeblair | i tried one, we'll see. :) | 18:29 |
| nibalizer | erp | 18:29 |
| jeblair | and forgot to set topic. fixing | 18:29 |
| asselin_ | i'm back | 18:30 |
| jeblair | clarkb, fungi: 150698 150705 150711 are ready for aprv | 18:32 |
| clarkb | jeblair: rgr looking now | 18:32 |
| asselin_ | yeah. I hit the alphabetic issues back here: https://review.openstack.org/#/c/140523/ so in doubt, follow those other files | 18:33 |
| jeblair | i've rechecked the other newly created changes, should have results on those soon | 18:34 |
| fungi | on it | 18:34 |
| fungi | sorry, was stuffing my face real quick | 18:35 |
| clarkb | jeblair: and when you say those are ready you checked that the repos exist now? /me assumes so and reviews | 18:35 |
| jeblair | i was planning on only eating puppet modules | 18:36 |
| clarkb | oh tests wouldn't pass until that happens right? | 18:36 |
| jeblair | clarkb: yes and yes. i move them from "split merged" to "repos created" after i verify that they exist on git.o.o. then i 'recheck' | 18:36 |
| clarkb | fungi is beating me to them | 18:36 |
| fungi | jeblair: clarkb: 150831 too | 18:36 |
| fungi | right? | 18:36 |
| jeblair | yep, missed that one | 18:37 |
| fungi | clarkb: sorry, since you're reviewing them too i'll leave off the approval | 18:37 |
| asselin_ | jeblair, recheckwatch is used here: remove it? https://github.com/openstack-infra/system-config/blob/master/modules/openstack_project/manifests/zuul_prod.pp#L104 | 18:38 |
| clarkb | fungi: no problem I sa don't wait for me | 18:38 |
| clarkb | fungi: its mostly cookiecutter anyways and jeblair did the hard bit | 18:38 |
| jeblair | asselin_: yep. can all go away. | 18:38 |
| fungi | yup | 18:38 |
| asselin_ | will do | 18:38 |
| jesusaurus | asselin_: oh was i not on latest master? | 18:41 |
| * jesusaurus looks at the diff | 18:41 | |
| asselin_ | jesusaurus, not sure.... | 18:41 |
| jeblair | approved asterisk creation | 18:42 |
| jeblair | and iptables | 18:44 |
| clarkb | ok puppet-gerrit is on git.o.o I am going to recheck the system-config change for that module | 18:44 |
| jeblair | approved reviewday add | 18:46 |
| jeblair | yeah, the alphabetization is the other way | 18:47 |
| mrmartin | fungi: the openstackid fails on puppetlint, here: https://github.com/openstack-infra/puppet-openstackid/blob/master/manifests/init.pp#L52 do you remember for this boolean value? | 18:47 |
| fungi | fixing merge conflicts on 150705 and 150711 now | 18:47 |
| mrmartin | what you suggest, make a real boolean from this value, or some change required in lint rules? | 18:48 |
| clarkb | mrmartin: making it a real boolean should be fine just make sure you double check ewhere it is used and fix an comparison that might expect a string | 18:48 |
| clarkb | mrmartin: but ruby/erb should coerce true the boolean to 'true' the string when writing it out if that is necessary they just won't be equivalent when compared to each other | 18:49 |
| jesusaurus | asselin_: they look the same to me, what differences are you seeing? | 18:52 |
| asselin_ | jesusaurus, hmm...maybe it's just different in the github ui that I was using? | 18:53 |
| asselin_ | jesusaurus, 4th commit here is different: https://github.com/jesusaurus/puppet-cgit/commits/master | 18:54 |
| asselin_ | than this one: https://github.com/openstack-infra/system-config/commits/master/modules/cgit | 18:54 |
| asselin_ | maybe it's just the 2 "reverts" | 18:55 |
| clarkb | fungi: I reapproved the two you rebased so hopefully those get in soon | 18:56 |
| jeblair | approved planet add | 18:57 |
| jesusaurus | asselin_: oh, interesting, git-subtree seems to have dropped both the revert commit and the reverted commit | 18:57 |
| fungi | clarkb: thanks | 18:57 |
| jeblair | that is interesting. i did not know it would do that. | 18:58 |
| mmedvede | jesusaurus: I've noticed that too, subtree split tends to get rid of null-effect commits | 18:58 |
| jesusaurus | today i learned... | 18:59 |
| fungi | 150828 and 150834 are also good to go but going to merge-conflict so i'll rebase them now | 19:00 |
| jeblair | so i'd prefer the full history with reverts, but i'm also not so tied to that that i think we need to rework the current process. so i'm inclined to say that we should accept the output of subtree (where it removes null-effect commits) | 19:04 |
| jeblair | anyone object? | 19:04 |
| asselin_ | let me check triniplex | 19:04 |
| clarkb | jeblair: I too prefer full history but am not worried enough about it to change the process | 19:04 |
| asselin_ | but i'm not tied to it either | 19:04 |
| jeblair | asselin_: oh, do you think triniplex has a different process? | 19:05 |
| asselin_ | just want to compare | 19:05 |
| mmedvede | jeblair: I also prefer history with reverts | 19:05 |
| asselin_ | seems to have the reverts: https://github.com/Triniplex/puppet-cgit/commits/master | 19:05 |
| mmedvede | I would say if there is a flag for subtree to allow null-effect commits, we should use it. Otherwise we have to change the workflow | 19:06 |
| jesusaurus | oh, has triniplex already made repos for all of the modules? | 19:07 |
| asselin_ | jesusaurus, yes, sweston is running his script to create all the repos | 19:08 |
| jeblair | jesusaurus: yes, sweston just ran the script to update them (so we need to carefully double check them) | 19:08 |
| mmedvede | hehe | 19:08 |
| jeblair | that's interesting that has has the commits. seems like we should use that then. | 19:08 |
| jesusaurus | i was going off of storyboard to see which ones weren't in progress yet | 19:08 |
| asselin_ | jesusaurus, no, storyboard is accurate. | 19:09 |
| asselin_ | jesusaurus, sweston's script just creates the repo. but you can use it for the actual moduel you pick in storyboard | 19:09 |
| nibalizer | jeblair: ++ to using subtree and ya it sucks that its not perfect but hey lets get this done | 19:09 |
| jesusaurus | asselin_: oh, ok, so all of the repos are created, and storyboard 'progress' just means that we are creating the changes to propose in gerrit? | 19:12 |
| jeblair | sweston's script does a bit more than just subtree | 19:12 |
| asselin_ | jesusaurus, yes | 19:12 |
| pleia2 | https://github.com/Triniplex/puppet-mediawiki looks a bit funny with the templates/apache bit there, thoughts? (templates/apache as a directory structure does exist in system-config) | 19:18 |
| clarkb | pleia2: if you hover over it github tells you why they are being special | 19:19 |
| mrmartin | pleia2: what's wrong with that? | 19:19 |
| clarkb | "this path skips through empty dirs" | 19:19 |
| pleia2 | clarkb: ah, thanks | 19:19 |
| jesusaurus | asselin_: ive updated the patch to point at triniplex-test | 19:20 |
| mrmartin | we have a very old apache module, and it is a typical way to customise the vhost config of apache | 19:20 |
| asselin_ | jesusaurus, ok | 19:20 |
| clarkb | pleia2: they use a similar font color for submodules too | 19:20 |
| pleia2 | clarkb: I guess I just don't look at github much | 19:22 |
| * nibalizer gonna go hunt down a burrito, be back in a bit | 19:25 | |
| jeblair | nibalizer: you have comments on 150888 and 150891 | 19:25 |
| clarkb | finally I think https://review.openstack.org/#/c/150814/ is ready | 19:30 |
| pleia2 | I need to run out for some pre-travel things, probably lunch too | 19:30 |
| jesusaurus | whats the difference between triniplex and triniplex-test? is it important to not use triniplex-test? | 19:31 |
| fungi | for puppet metadata.json are we okay ignoring dependencies for now and worrying about that later? for example, mmedvede's comment at https://review.openstack.org/#/c/150885/1/metadata.json | 19:31 |
| asselin_ | jesusaurus, use triniplex. triniplex-test seems to be outdated | 19:31 |
| clarkb | fungi: for us it should continue to work because of how we use install modules. but eventually we should clean it up and properly list deps | 19:32 |
| asselin_ | otherwise it's the same...we didn't want to update triniplex-test b/c some in-flight changes | 19:32 |
| jesusaurus | ahh, ok | 19:32 |
| fungi | clarkb: yeah, i figured for the sake of expediency that can be left as a future lhf exercise | 19:32 |
| fungi | just want to make sure we're all on the same page | 19:32 |
| jeblair | pleia2: thanks! | 19:33 |
| jeblair | i just approved statusbot, sudoers, tmpreaper, accessbot, releasestatus, graphite, mediawiki, unbound, user, and ssh | 19:34 |
| fungi | awesome | 19:35 |
| clarkb | 150814 too | 19:36 |
| clarkb | pretty sure its ready. jenkins says +1 | 19:36 |
| fungi | nibalizer: the only complaint with 150888 seems to be whitespace, so i'll fix that for you | 19:36 |
| *** krtaylor has quit IRC | 19:41 | |
| jeblair | i've updated the etherpad with info for commiters and reviewers based on our current processes | 19:42 |
| sweston | mmedvede: if it's really crucial, I believe that we can do a filter on the split for null-effect commits | 19:44 |
| sweston | if anybody has any questions I am back, and can answer them now ;-) | 19:44 |
| jeblair | sweston: my understanding is that whatever you are doing for triniplex is correct -- i think asselin_ said that your repo does have the revert commits | 19:45 |
| mmedvede | sweston: I think we came to conclusion we want to keep null-effect commits | 19:45 |
| asselin_ | sweston, yes it did | 19:45 |
| * mmedvede in meeting for the next 30 minutes | 19:46 | |
| sweston | jeblair, mmedvede, asselin_ : ok. yes, all commits are kept by the current logic | 19:46 |
| jeblair | i am going to make and eat food now. hopefully some changes will have merged by the time i get back. | 19:46 |
| fungi | nibalizer: i'm updating 150891 for you based on reviewer feedback too | 19:46 |
| jeblair | (looks like there are a few more project-config changes that could be reviewed) | 19:47 |
| clarkb | I am going to make a review pass now before lunch | 19:47 |
| clarkb | pleia2: https://review.openstack.org/#/c/150903/ needs alphabetizing | 19:48 |
| clarkb | jeblair: any reason you didn't approve 150926? | 19:49 |
| *** krtaylor has joined #openstack-sprint | 19:53 | |
| fungi | 150885 is ready too | 19:54 |
| nibalizer | fungi: thanks for doing that, sorry to make more work for you | 19:55 |
| fungi | nibalizer: not at all. that was trivial | 19:55 |
| clarkb | asselin_: for https://review.openstack.org/#/c/150854/3 was there something specific that concerned you about the two repos being different? | 20:00 |
| clarkb | asselin_: diff for me says they are the same but I figured I would double chekc since your comment implies maybe it needed more scrutiny | 20:00 |
| asselin_ | clarkb, we got that resolved | 20:01 |
| clarkb | asselin_: cool so I am good to approve? | 20:01 |
| asselin_ | yes | 20:01 |
| clarkb | done | 20:01 |
| jeblair | clarkb: i did; mmedvede uploaded a new patch | 20:06 |
| jeblair | looks like a rebase | 20:06 |
| jeblair | aprvd | 20:06 |
| clarkb | jeblair: oh missed that sorry | 20:06 |
| jeblair | or rather, asselin_ uploaded a new patch | 20:07 |
| jesusaurus | clarkb: the difference was in git-subtree dropping a commit and the revert immediately following it | 20:10 |
| clarkb | oh that was in that repo | 20:10 |
| asselin_ | these are a bit different: https://github.com/Triniplex/puppet-exim/commits/master | 20:16 |
| asselin_ | https://github.com/openstack-infra/system-config/commits/master/modules/exim | 20:16 |
| asselin_ | nibalizer, you comment that they're good right? maybe visual inspection isn't sufficient? ^^ | 20:18 |
| asselin_ | https://review.openstack.org/#/c/150946/1 | 20:18 |
| * asselin_ going for lunch | 20:20 | |
| nibalizer | asselin_: so what i've been doing is subtree splitting in my own local checkout | 20:24 |
| nibalizer | then comparing that to the commit history on the triniplex github | 20:25 |
| nibalizer | and im largely ignoring incidents of commit/revert showing up or not showing up, since we've established we don't care too much about that | 20:26 |
| fungi | 150894 and 150892 are merge-conflicting with master now, so i'll rebase them | 20:26 |
| jeblair | back | 20:28 |
| *** rfolco has quit IRC | 20:30 | |
| fungi | 150924 and 150859 look like they're probably shortly getting kicked out of the gate on merge conflicts with other approved changes | 20:32 |
| fungi | yep, 150924 just came back. rebasing now | 20:33 |
| jeblair | i approved the deletes that were reday | 20:33 |
| fungi | thanks | 20:34 |
| nibalizer | asselin_: minor issue https://review.openstack.org/#/c/140431/ | 20:34 |
| clarkb | pleia2: is there a project-config change for graphtie split? I see the system config change but not the project-config change | 20:38 |
| asselin_ | nibalizer, thanks fixed | 20:38 |
| jeblair | asselin_: i think that in 150946 the difference is that a merge commit is missing. the actual commits are there. i believe that's probably to be expected in a history rewrite like this and is okay. | 20:38 |
| pleia2 | clarkb: looking, I thought I had | 20:38 |
| pleia2 | clarkb: doh, half done, finishing now | 20:38 |
| pleia2 | and needs a rebase now anyway, so will redo from current master | 20:39 |
| jeblair | asselin_: would you agree with that? | 20:39 |
| * asselin_ looking | 20:39 | |
| pleia2 | clarkb: wait, it was already merged: https://review.openstack.org/#/c/150916/ | 20:40 |
| clarkb | pleia2: oh no wonder I didn't see it. just waiting for the repo to end up on the mirrors then? | 20:40 |
| clarkb | it merged recently enough that I probably just need to be more patient | 20:40 |
| pleia2 | clarkb: yeah, it was recent | 20:40 |
| pleia2 | I'm so happy we're doing this today, fosdem talk on Saturday is about puppet and we're in much better shape than the first couple times I've given similar talks :) | 20:42 |
| pleia2 | look how great we are with all our split out modules | 20:42 |
| asselin_ | nibalizer, jeblair ok I think it's fine | 20:42 |
| krtaylor | rebased both bup patches, they should be ready for review 150830 and 150895 | 20:48 |
| jeblair | pleia2: ++! | 20:48 |
| krtaylor | hehheh | 20:48 |
| fungi | pleia2: heh, the tarnished silver soup tureen calling the kettle black, are we? | 20:51 |
| jesusaurus | is the remove_nginx class still used/needed or can we just delete it? | 20:51 |
| jeblair | jesusaurus: just delete | 20:51 |
| fungi | jesusaurus: i believe it can just disappear | 20:51 |
| fungi | if it _is_ used anywhere, delete the use of it while you're at it | 20:52 |
| jesusaurus | will do | 20:52 |
| fungi | but i think i already found/removed a lot of them if not all | 20:52 |
| jeblair | puppet-askbot has a boilerplate change that can be aprvd | 20:52 |
| pleia2 | fungi: turns out it's actually valuable in the talk that we did it "wrong" previously, we have a whole spec that talks about to do it the right way | 20:53 |
| fungi | pleia2: absolutely | 20:53 |
| pleia2 | or that's what I tell myself | 20:53 |
| jeblair | puppet-gerrit as well | 20:54 |
| jeblair | mmedvede: i think puppet-elasticsearch is ready for a boilerplate patch | 20:58 |
| jeblair | mmedvede: oh, sorry, my repo was up to date. it has already landed. nm. | 20:59 |
| jeblair | will update etherpad | 20:59 |
| fungi | the redis and openstackid boilerplate changes are also ready for another +2/approval | 20:59 |
| pleia2 | https://github.com/Triniplex/puppet-lodgeit is empty (and was called "lodgetit" in storyboard, fixed after giggles) | 21:00 |
| jeblair | fungi: can you re-examine 135376 (there are comment replies to you there) | 21:00 |
| jeblair | sweston, mmedvede: can you look into puppet-lodgeit? | 21:02 |
| pleia2 | oh, there is already a puppet-lodgeit module | 21:02 |
| fungi | jeblair: nibalizer: thanks, approved 135376 | 21:02 |
| mmedvede | jeblair: looking | 21:02 |
| pleia2 | http://git.openstack.org/cgit/openstack-infra/puppet-lodgeit/ | 21:02 |
| pleia2 | so I don't think we need to do that one afterall | 21:02 |
| jeblair | pleia2: oh, oops, it's already done | 21:02 |
| fungi | pleia2: maybe we missed the follow-on delete change for it? | 21:02 |
| mmedvede | oh yes, I think it was merged / split a while back | 21:02 |
| jeblair | mmedvede, sweston: sorry, nevermind | 21:03 |
| fungi | pleia2: nope, it's definitely absent from system-config so we're fine | 21:03 |
| pleia2 | fungi: it's actually been there for a while, come to think of it | 21:03 |
| jeblair | it's marked as merged, but has no assignee in storyboard | 21:03 |
| fungi | yeah, i seem to remember it going in early as a proving ground for the process | 21:03 |
| jeblair | added jesusaurus to it for completeness | 21:04 |
| asselin_ | yes I marked it merged but didn't know who | 21:04 |
| asselin_ | also 'github' is merged, but I put myself there accidently. not sure who did that one | 21:04 |
| asselin_ | (and I can't remove myself) | 21:04 |
| clarkb | hrm still no puppet-graphite | 21:04 |
| jeblair | clarkb: most recent git04 encountered the timeout bug | 21:05 |
| jesusaurus | jeblair: oh whoops, thanks | 21:05 |
| fungi | jeblair: we're also hitting it in centos jobs, like this just now http://logs.openstack.org/85/150885/1/gate/gate-infra-puppet-apply-centos6/a4f455f/console.html#_2015-01-28_20_29_42_741 | 21:06 |
| nibalizer | fungi: neato! | 21:06 |
| sweston | jeblair: no worries | 21:07 |
| jeblair | fungi: it seems like the git servers are under a lot of load right now? | 21:07 |
| *** ianw has joined #openstack-sprint | 21:07 | |
| fungi | jeblair: i'm looking in cacti for any sign of the issue | 21:07 |
| clarkb | I recently looked at cacti and bw use is low | 21:07 |
| clarkb | which I would expect to correspond to load | 21:07 |
| clarkb | but I was looking at front ends not backends | 21:08 |
| jeblair | they have been seeing load spikes recently | 21:09 |
| mmedvede | jhesketh: are you working on ehterpad_lite module? It was unassigned in storyboard | 21:09 |
| clarkb | jeblair: appears to correspond to spikes in disk activity | 21:10 |
| clarkb | but not bw | 21:10 |
| asselin_ | 2 modules left unassigned in storyboard: mysql_backup & packagekit | 21:11 |
| fungi | yeah, load spikes up to ~100 | 21:11 |
| ianw | is being unable to login to storyboard a known thing? | 21:11 |
| jeblair | ianw: no, best to move back to #infra and ask about that | 21:12 |
| fungi | those load spikes on the git backends are hitting different servers at different times, so almost certainly something pounding git.o.o from a single source and getting persisted to one backend at a time | 21:13 |
| jeblair | boilerplate for puppet-zuul is ready for aprv | 21:15 |
| clarkb | jeblair: reviewing | 21:16 |
| jeblair | rebasing 150888 | 21:17 |
| clarkb | puppet zuul boilerplate is approved | 21:18 |
| clarkb | stepping out for food back in a bit | 21:18 |
| pleia2 | asselin_: all assigned \o/ | 21:19 |
| fungi | we've got a nice pipeline of changes flowing through that etherpad now | 21:20 |
| jeblair | rebasing 150932 | 21:20 |
| ianw | fixed storyboard link in 136959 (puppet-httpd) | 21:20 |
| fungi | rebasing 150936 | 21:21 |
| fungi | oh, and 150926 since i had them stacked previously | 21:22 |
| asselin_ | pleia2, yup :) | 21:22 |
| nibalizer | asselin_: merge conflict on https://review.openstack.org/#/c/150926/ | 21:23 |
| * asselin_ looking | 21:24 | |
| nibalizer | er it looks like fungis on it | 21:24 |
| asselin_ | ok | 21:24 |
| fungi | yeah, mentioned above i was rebasing them | 21:25 |
| fungi | rebasing 150859 | 21:28 |
| *** mrmartin has quit IRC | 21:30 | |
| jeblair | rebasing 150882 | 21:31 |
| jeblair | and 150881 | 21:31 |
| fungi | rebasing 150891 | 21:34 |
| asselin_ | FYI: I "assigned" in storyboard the previously merged/deleted modules to the commiter who removed them | 21:36 |
| jeblair | ++ | 21:36 |
| jesusaurus | rebasing 150920 | 21:36 |
| fungi | puhpow | 21:37 |
| jeblair | 150830 can use a +3 | 21:38 |
| jeblair | and 150974, 150978, and 150984 | 21:40 |
| jeblair | asselin_: 150968 has a -1 comment | 21:44 |
| asselin_ | lookin | 21:44 |
| asselin_ | jeblair, fixed & rebased | 21:46 |
| jeblair | logstash is already in boilerplate merged (last step) | 21:49 |
| asselin_ | pleia2, needs rebase https://review.openstack.org/#/c/150914/ | 21:51 |
| jeblair | rebasing 150873 | 21:52 |
| asselin_ | pleia2, I'll rebase 150914 | 21:52 |
| pleia2 | k, thanks | 21:53 |
| fungi | yeah, at this point i think if a sprint change needs a rebase or cosmetic fixes, it's faster to just do it than -1 or ping someone | 21:53 |
| jhesketh | Morning | 21:54 |
| jhesketh | Feel free to ping me for reviews etc | 21:54 |
| jeblair | jhesketh: awesome; check out the info on the etherpad: https://etherpad.openstack.org/p/puppet-module-split-sprint | 21:55 |
| jeblair | jhesketh: all of the modules have been assigned, so we're mostly reviewing/approving/rebasing changes now | 21:55 |
| jeblair | jhesketh: at the bottom of the etherpad, i'm moving names of modules down those 4 sections to help keep track of the workflow for them | 21:56 |
| jhesketh | Okay cool, thanks | 21:57 |
| jeblair | rebasing 150894 | 21:57 |
| fungi | we have a pile of new needed rebases. i'll take 150950, 150903 and 150830 | 21:59 |
| clarkb | fungi I really want a picture of the no rebase sign now | 22:00 |
| fungi | clarkb: http://static.lolyard.com/lol/no-rebase.jpg | 22:01 |
| fungi | (proving my point of why bother to bring a camera when someone else has already taken all the pictures of the places i'm going and posted them to the internet) | 22:02 |
| jeblair | rebaisg 150910 | 22:03 |
| jeblair | rebasing 150933 | 22:05 |
| fungi | rebasing 150905 | 22:06 |
| nibalizer | fungi: whoa | 22:06 |
| nibalizer | thats kinda mind blowing | 22:06 |
| nibalizer | plus they probably got the lighting right and had a better camera anyways... | 22:06 |
| pleia2 | fungi: your camera logic is no fun :) | 22:07 |
| fungi | pleia2: but characteristically lazy | 22:08 |
| fungi | rebasing 150857 | 22:08 |
| jeblair | btw, we should be keeping the modules.env list in alpha order | 22:11 |
| jeblair | i know there isn't a check for that, but it helps (a little) with rebases | 22:11 |
| jeblair | i will update 150852 to reorder | 22:11 |
| nibalizer | jeblair: it didn't start that way but yea i agree its a good idea | 22:11 |
| fungi | 150960 has an alpha order failure in projects.yaml. i'll upload a fixed version | 22:13 |
| asselin_ | At this point, I think we should fix as part of rebasing changes | 22:14 |
| jeblair | asselin_: makes sense | 22:14 |
| mmedvede | Are all of the modules need to be split? Because some of them are not on storyboard, e.g. kerberos | 22:14 |
| jeblair | ah, that's a late addition | 22:15 |
| asselin_ | mmedvede, I synced storybaord with the spec. | 22:15 |
| jeblair | i think it was added after spec approval. we can/should split it i think. | 22:15 |
| asselin_ | mmedvede, also the premerge of project-config is only what was in the spec | 22:15 |
| fungi | asselin_: jeblair: yeah, as i rebase changes to modules.env i've been trying to tease them closer into alpha order | 22:15 |
| mmedvede | asselin_: ok, thanks for that. Forgot about the pre-merge... | 22:16 |
| nibalizer | what should i be doing to help? | 22:16 |
| jeblair | nibalizer: i think we're mostly watching for conflicts and rebasing at this point | 22:18 |
| jeblair | mmedvede: might have some more modules to add to the list? | 22:18 |
| mmedvede | jeblair: checking now | 22:18 |
| mmedvede | I have pushed kerberos | 22:19 |
| mmedvede | But forgot about the pre merge, need to fix that | 22:19 |
| jeblair | mmedvede: i'm guessing there will similarly be an openafs module | 22:19 |
| jeblair | fixing 150961 | 22:20 |
| nibalizer | jeblair: ya, it seems like at this point it becomes de-parallelized and cores have to line things up/rebase/+a | 22:20 |
| jeblair | well, anyone can rebase/recheck; i'm doing very little +Aing | 22:20 |
| jeblair | i am half-tempted to try to stack all of the changes, but what's already in zuul, etc, makes that a bit complex | 22:21 |
| jesusaurus | fungi: oh, you fixed 150960. i was trying to figure out why it didnt think i was making a change | 22:21 |
| fungi | jesusaurus: yep, mentioned in scrollback | 22:22 |
| nibalizer | i feel like if I try to rebase anything i'll just step fungi's toes | 22:22 |
| * jesusaurus goes back to adding boilerplate | 22:22 | |
| fungi | nibalizer: not at all! before i rebase anything i mention it in here and keep an eye on what anyone else says they're rebasing/fixing | 22:22 |
| fungi | so as to avoid any more duplication of effort than we're already stuck with | 22:23 |
| jeblair | rebasing 150962 | 22:23 |
| fungi | taking a break for a few to heat up dinner | 22:24 |
| nibalizer | okay | 22:25 |
| nibalizer | im gonna take a stab then! | 22:25 |
| mmedvede | Is it possible to unassign task in storyboard? Can not add a new task without assigning it to myself... | 22:25 |
| jeblair | mmedvede: i would ask krotscheck in #openstack-infra | 22:26 |
| asselin_ | mmedvede, not that I know of. but I've been able to create new tasks unassigned | 22:26 |
| mmedvede | jeblair, asselin_: thx | 22:26 |
| * asselin_ going for a 30 minute meeting | 22:30 | |
| * nibalizer gonna rebase 150861 | 22:31 | |
| clarkb | https://review.openstack.org/#/c/150960/ looks like it got caught in rebasing? should be ready for approval though | 22:31 |
| clarkb | jeblair: fungi ^ | 22:31 |
| fungi | rebasing 150972 | 22:32 |
| fungi | rebasing 150917 | 22:35 |
| clarkb | whenever people say "rebasing" I think of the "reloading" yells in L4D | 22:35 |
| nibalizer | gonna rebase 150882 even though jim did it an hour ago | 22:36 |
| fungi | there will be a lot of that unless we stack them all | 22:37 |
| mmedvede | added 5 more modules into storyboard. They all have no changes pre-merged yet | 22:37 |
| nibalizer | yea... should we try to stack? | 22:37 |
| nibalizer | im trying to mentally figure out if we eventually converge or if it just cycles | 22:37 |
| nibalizer | i think it conversges | 22:37 |
| jeblair | if we do so, we need to stack _everything_ and we can't let anything merge that isn't stacked because of the octomerge issue | 22:37 |
| krtaylor | octomerge? sounds like a Bond character | 22:39 |
| jeblair | re-alphabetizing 150888 | 22:41 |
| jeblair | jhesketh: are you around to start helping out on approvals? | 22:44 |
| jeblair | rebasing 150932 | 22:45 |
| jeblair | oh that one just needed to be un-based (its parent was messed up) | 22:46 |
| jeblair | (stacking fail) | 22:46 |
| jhesketh | jeblair: sorry in about 20min? | 22:50 |
| fungi | jhesketh: there's still plenty to chew on, i think | 22:52 |
| nibalizer | jeblair: so this one is wildly unalphabetized https://review.openstack.org/#/c/150857/2/modules.env does that mean i should rebase it? | 22:52 |
| jeblair | nibalizer: only if it needs a rebase | 22:54 |
| jeblair | nibalizer: (zuul queue position is too precious now to waste :) | 22:55 |
| nibalizer | mk | 22:56 |
| nibalizer | so i can detect that by pulling it down and attepmting to rebase onto master | 22:56 |
| nibalizer | or i can press that button | 22:56 |
| nibalizer | but im worried about that button, never been super sure what it does | 22:56 |
| jeblair | nibalizer: the button will (if needed) create a new patchset, so don't do that right now | 22:57 |
| jeblair | nibalizer: zuul will return a patch conflict immediately on enquing, so once it's in check, zuul will tell us if it conflicts with master; however, once it moves to gate, it still might conflict with something ahead | 22:58 |
| jeblair | nibalizer: so sometimes we get +1 and the -2 on these | 22:58 |
| jeblair | nibalizer: the only thing you could do locally to work ahead of zuul would be to try to apply it to master + whatever else is in gate | 22:59 |
| jeblair | nibalizer: if you do that and it fails, it's worth cherry-picking it onto the tail of the gate queue | 22:59 |
| * nibalizer nods | 23:00 | |
| * asselin_ is back | 23:04 | |
| jeblair | rebasing 150927 | 23:09 |
| jhesketh | jeblair: okay, I'm going to reviews now.. any in particular you wanted me to look at? | 23:19 |
| asselin_ | is metadata.json or Metadata.json correct? or either is fine? | 23:20 |
| clarkb | jhesketh: I have just been going through the main list of open reviews for that otpic and finding changes without +2's or approvals | 23:20 |
| jeblair | nibalizer: ^? | 23:20 |
| jhesketh | clarkb: will do | 23:20 |
| fungi | jhesketh: https://review.openstack.org/#/q/topic:module-split+status:open,n,z | 23:20 |
| fungi | churn through all the things | 23:20 |
| jeblair | jhesketh: and also watching out for failures that may need rebasing / reapproval | 23:20 |
| asselin_ | nibalizer, is metadata.json or Metadata.json correct? or either is fine? | 23:21 |
| jeblair | gerritbot repo was just created, so i'm rechecking 150983 now | 23:21 |
| jhesketh | jeblair: do we have automatic puppet updates turned off during this sprint? | 23:22 |
| jeblair | jhesketh: nope, everything is operating normally | 23:22 |
| nibalizer | asselin_: lower case please | 23:23 |
| fungi | rebasing 150935 | 23:23 |
| asselin_ | ok | 23:23 |
| jeblair | jhesketh: all the changes should be noops | 23:23 |
| jeblair | jhesketh: and we've already completely finished gerrit (and many others) without breakage | 23:24 |
| jhesketh | cool | 23:24 |
| jeblair | 151000 is ready for aprv | 23:26 |
| clarkb | looking | 23:26 |
| jhesketh | asselin_: 151039 any reason why you didn't put the rake file or the metadata in that change? | 23:28 |
| asselin_ | jhesketh, just added them | 23:28 |
| asselin_ | for to include them in the first patch | 23:29 |
| asselin_ | forgot | 23:29 |
| jeblair | the check queue graph is trending downward; we might speed up soon (cross fingers) | 23:29 |
| fungi | yeah, the massive gate breakage is fixed, so now we just have the massive backlog it caused | 23:29 |
| mmedvede | Added bulk changes needed for remaining 5 modules https://review.openstack.org/#/c/151041/ | 23:30 |
| jeblair | mmedvede: cool, thanks | 23:32 |
| jeblair | mmedvede: so i left comments on that that i thought we didn't want to split project_config, but then i realized i was probably wrong about that and it's fine... | 23:32 |
| jeblair | does that sound right to everyone? | 23:32 |
| clarkb | wait what is puppet-project_config? | 23:33 |
| clarkb | oh the thing tha installs project-config | 23:33 |
| jeblair | clarkb: yep | 23:33 |
| mmedvede | yes | 23:34 |
| jeblair | clarkb: so it should take "your project config repo git url" as a parameter and check it out for you | 23:34 |
| clarkb | and define some parameters | 23:34 |
| jeblair | and it also provides puppet vars that other modules can use | 23:34 |
| clarkb | kk | 23:34 |
| clarkb | wait did puppet-subversion merge? | 23:35 |
| clarkb | I see it in the projects list | 23:35 |
| clarkb | nibalizer: ^ | 23:35 |
| jeblair | clarkb: is it just in the base for that patch? | 23:35 |
| jeblair | clarkb: if so, you're probably looking at the pre-merge stuff that was added a long time ago | 23:36 |
| jeblair | clarkb: we'll want to clean that up, but it's not urgent | 23:36 |
| clarkb | ya I think so | 23:36 |
| jeblair | as long as it's not in projects.yaml | 23:36 |
| nibalizer | clarkb: puppet-subversion just got refactored out | 23:38 |
| nibalizer | it was a whole class for package { 'subversion': } | 23:38 |
| nibalizer | i dont care if project_config splits or not | 23:39 |
| nibalizer | could go either way | 23:39 |
| fungi | clarkb: right, the removal patch for it merged | 23:41 |
| fungi | there was no split patch, just one that replaced it by the package resource in the one module using it | 23:41 |
| fungi | i'm starting to wind down over here, but will check in a bit later on to see if things are getting stuck and need me to jump back in | 23:57 |
| fungi | amazing progress on the sprint so far | 23:57 |
| nibalizer | indeed! | 23:59 |
| mmedvede | fungi: I have 5 more modules coming in soon | 23:59 |
| fungi | mmedvede: awesome! | 23:59 |
| jeblair | i have to run out for a bit but will be back soon | 23:59 |
| jeblair | ping clarkb and jhesketh if approvals are needed :) | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
