*** gtristan has joined #baserock | 06:34 | |
*** ctbruce has joined #baserock | 07:29 | |
*** paulw has joined #baserock | 07:40 | |
*** toscalix has joined #baserock | 07:45 | |
*** locallycompact has joined #baserock | 08:00 | |
*** locallycompact has quit IRC | 08:20 | |
*** locallycompact has joined #baserock | 08:20 | |
*** ChrisPolin has quit IRC | 08:36 | |
*** ChrisPolin has joined #baserock | 08:55 | |
*** tiagogomes has joined #baserock | 09:03 | |
*** jjardon_matrix has quit IRC | 09:21 | |
*** jjardon_matrix has joined #baserock | 09:36 | |
*** locallycompact has joined #baserock | 09:37 | |
*** CTtpollard has quit IRC | 09:46 | |
*** CTtpollard has joined #baserock | 09:50 | |
pedroalvarez | will gitlab push back the changes to git.baserock.org? | 10:02 |
---|---|---|
pedroalvarez | will working on git.baserock.org won't be allowed anymore? | 10:02 |
locallycompact | it still holds the bulk of the source | 10:05 |
locallycompact | we can lorry from gitlab | 10:05 |
locallycompact | if we want the premerge then MRs still have to be on gitlab | 10:06 |
locallycompact | or what do you want to happen | 10:06 |
pedroalvarez | right now is still posible to work in git.baserock.org, merging things there directly | 10:11 |
pedroalvarez | and gerrit replicates that | 10:11 |
pedroalvarez | and gerrit pushes back to g.b.o. after any merge in gerrit (not lorrying it, so that it's not in delta/) | 10:11 |
pedroalvarez | i just want to know how is this going to work | 10:12 |
mwilliams_ct | If gitlab.com is on EE then it would be able to mirror from git.baserock.org and that could lorry from gitlab in turn (though I wont speak for how robust the mirroring is) | 10:13 |
mwilliams_ct | EE = enterprise edition | 10:13 |
mwilliams_ct | sorry jargon | 10:13 |
pedroalvarez | I assume no EE here | 10:13 |
locallycompact | nop | 10:13 |
pedroalvarez | anyway, who is going to maintain all of this? | 10:14 |
pedroalvarez | (new infra) | 10:14 |
pedroalvarez | have the new runners been configured using some configuration management tools like ansible? | 10:15 |
locallycompact | Seems to kind of maintain itself | 10:15 |
locallycompact | We have our own runners yet? | 10:15 |
locallycompact | ybd is still on shared | 10:16 |
pedroalvarez | i understood that from the email | 10:16 |
locallycompact | definitions is on autoscale I guess that's our's | 10:17 |
locallycompact | where/what is that manager | 10:17 |
pedroalvarez | I don't follow | 10:20 |
locallycompact | definitions is connected to two runners called | 10:21 |
locallycompact | autoscale-runner-manager and baserock-mustang | 10:21 |
locallycompact | the first one is elastic 64bit but where is it | 10:22 |
locallycompact | jjardon, | 10:22 |
pedroalvarez | thanks for clarifying | 10:22 |
jjardon | locallycompact: is not elastic, Its the manager to generate the elastic nodes: Its an machine in DigitalOcean as well | 10:23 |
pedroalvarez | and everything maintains itself? | 10:24 |
pedroalvarez | security upgrades and all | 10:24 |
jjardon | pedroalvarez: securitu upgrades for what? | 10:25 |
*** ctbruce has quit IRC | 10:25 | |
mwilliams_ct | the operating system the machines run on? | 10:25 |
mwilliams_ct | and all the software on it? | 10:26 |
jjardon | what machines exactly? the runners? the only that have to be maintained manually is the ARM runner | 10:26 |
pedroalvarez | have they been configured using some configuration management tools like ansible? or just manually? | 10:26 |
mwilliams_ct | autoscale-runner-manager and baserock-mustang? | 10:26 |
jjardon | manually | 10:26 |
pedroalvarez | right, so who is going to maintain that, or who are we going to blame to if something is not working | 10:27 |
jjardon | but is the easiest thing in the world: install gitlab-runner; register the runner; done. For the autoscale manager a little more config is needed but its everything in a toml file | 10:28 |
jjardon | I can share it privately if you are curious | 10:28 |
mwilliams_ct | jjardon: I'd be interested in seeing the toml file as it's relevant to something I'm doing | 10:28 |
jjardon | mwilliams_ct: you have that file in our private tracker already :) | 10:29 |
mwilliams_ct | But that's not why pedroalvarez is asking. The Baserock ops team doesnt seem to have been involved in this at all, so how can they ensure it is secure? and available all the time? | 10:29 |
pedroalvarez | thanks for helping mwilliams_ct | 10:30 |
mwilliams_ct | I apologise if youre part of the ops team jjardon, last I knew it was gary_perkins, pedroalvarez and ssam2 | 10:30 |
*** ctbruce has joined #baserock | 10:30 | |
jjardon | mwilliams_ct: Im not part of any specific team | 10:30 |
jjardon | I setup this as a demo in my free time, to show the community and hoping it can be of interest; If that's the case we can formalize things | 10:31 |
pedroalvarez | so this is a demo | 10:32 |
pedroalvarez | right | 10:32 |
pedroalvarez | ok | 10:32 |
mwilliams_ct | jjardon: OK. I'm not really part of the baserock community, plus or minus a few patches but it strikes me that we should be phrasing this discussion as "can the ops team agree to take on the maintenance overhead of runners". I'll step back for now as I have other things to do, but I imagine it is quite a bit of work | 10:33 |
jjardon | but a demo that its much more powerful than anything we have had before; I wanted you to see its potential | 10:33 |
jjardon | no | 10:34 |
jjardon | runners are quite easy to maintain; I will do it myself if nobody has the time | 10:35 |
* pedroalvarez adds jjardon to the team :P | 10:35 | |
jjardon | Do I get a badge? | 10:39 |
pedroalvarez | haha :) | 10:40 |
pedroalvarez | (not sure if this is a genuine question, I've heard the badge term in some communities, but not in baserock, ever) | 10:41 |
jjardon | :) we need to implement that, then | 10:44 |
jjardon | other cool feature: rigth now the elastic runners are machines with 4cores and 4GB of RAM (same as shared runners in gitlab.com), but with a one line change in the toml file it can be as powerful as we want | 10:46 |
jjardon | for the curious, this is the entire toml configuration in the autoscale-runner-manager: | 10:47 |
pedroalvarez | I'm all for this change, honestly. | 10:47 |
jjardon | https://www.irccloud.com/pastebin/3QTjUC3F/ | 10:47 |
pedroalvarez | the configuration of the infra should go to infrastructure.git if possible | 10:48 |
jjardon | pedroalvarez: good to heard that :) | 10:48 |
pedroalvarez | I'm just not looking forward to move from the demo to well-done state | 10:48 |
pedroalvarez | s/to move/to move myself this new infra/ | 10:50 |
pedroalvarez | I'll definitely help though | 10:50 |
jjardon | pedroalvarez: remember that the autoscale runners is a nice-to-have; we can always use the free runners provided by gitlab.com is something goes wrong / Its not ready | 10:53 |
*** lc_ has joined #baserock | 11:03 | |
*** anahuelamo has quit IRC | 11:16 | |
tiagogomes | If you are moving definitions to gitlab, I think it makes sense to move lorry, lorry-controller, firehose, tbdiff, fhs-dirs, … | 11:18 |
pedroalvarez | (anything on gerrit) | 11:19 |
tiagogomes | What will happen to the gerrit instance, will it stay read-only ? | 11:22 |
pedroalvarez | Thats another question | 11:22 |
pedroalvarez | we will break links, and lose patches | 11:22 |
jjardon | tiagogomes: step by step :) Im mostly interested in definitions because the pre-merge builds, but we can do that, sure | 11:26 |
pedroalvarez | so, we are not killing gerrit then | 11:27 |
SotK | we should be doing pre-merge ci on everything | 11:27 |
SotK | and gating on the result | 11:27 |
SotK | at this point I don't really care with what tool | 11:27 |
tiagogomes | tbh I would prefer to move everything at the same time to avoid ending up in the mess where the canonical repos are spread in different servers. | 11:27 |
jjardon | if we agree on that, it can be easily done | 11:30 |
jjardon | setup those project as mirrors in gitlab -> stop mirroring when they are the canonical source | 11:31 |
tiagogomes | That's what I thought - it shouldn't be much hassle to move them at the same time | 11:31 |
*** anahuelamo has joined #baserock | 11:32 | |
*** ctbruce has quit IRC | 12:20 | |
*** franred has quit IRC | 12:20 | |
*** ctbruce has joined #baserock | 12:36 | |
*** franred has joined #baserock | 12:36 | |
*** gtristan has quit IRC | 12:47 | |
*** jjardon_matrix has quit IRC | 13:08 | |
*** jjardon[m] has joined #baserock | 13:17 | |
*** toscalix_ has joined #baserock | 14:29 | |
*** toscalix has quit IRC | 14:29 | |
*** franred has quit IRC | 15:08 | |
*** toscalix has joined #baserock | 15:34 | |
*** toscalix_ has quit IRC | 15:38 | |
jjardon | hi tiagogomes, is this branch something you plan to submit upstream or are you still working on it? Seems without it the usr-merge is not possible https://gitlab.com/tacgomes/ybd/commit/2a7963a3960c07c9e4caa1e7a4c0b69b800bd452 | 15:40 |
pedroalvarez | patches can be taken upstream by anyone, anyway | 15:41 |
tiagogomes | jjardon I am not planning further work on it. I'll create a PR | 15:47 |
tiagogomes | jjardon, https://gitlab.com/baserock/ybd/merge_requests/259 | 15:50 |
jjardon | tiagogomes: cool, thanks | 15:53 |
*** toscalix has quit IRC | 16:07 | |
*** ctbruce has quit IRC | 16:20 | |
*** franred has joined #baserock | 16:37 | |
*** jjardon[m] has quit IRC | 16:42 | |
*** CTtpollard has quit IRC | 16:45 | |
*** jjardon[m] has joined #baserock | 16:45 | |
jjardon | tiagogomes: not sure I using the correct branch, but I'm getting the same error using you branch: https://gitlab.com/baserock/definitions/builds/5647410 | 16:48 |
*** tiagogomes has quit IRC | 16:52 | |
*** locallycompact has quit IRC | 16:53 | |
*** franred has quit IRC | 16:54 | |
*** jjardon[m] has quit IRC | 17:56 | |
*** jjardon_matrix has joined #baserock | 18:06 | |
*** rdale has quit IRC | 18:15 | |
*** paulw has joined #baserock | 19:00 | |
*** paulw has joined #baserock | 19:23 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!