30 Jun 2020 |
_hc | Bubu: I sent you an email to get access to the jenkins.debian.net runner | 14:36:52 |
_hc | careful with it though, it is easy to break the job by messing with the local files there | 14:37:07 |
@freenode_izzy:matrix.org | _hc: any idea why that pipeline permanently runs into an endless series of "500 Internal Server Errors"? Waiting for it's completion almost 2h now. Canceled and restarted – it looks like it wants to process a bunch of changes completely unrelated to the corresponding MR. https://gitlab.com/fdroid/fdroiddata/-/jobs/618201426 | 19:21:41 |
@freenode_izzy:matrix.org | It e.g. wants to process a package "com.gocoronago.tracker" which wasn't even there at that commit. Neither can I find it at HEAD. Very strange. | 19:25:14 |
_hc | izzy: issuebot will retry on each 500 error from gitlab... | 19:27:27 |
_hc | so gitlab is/was broken | 19:27:36 |
@freenode_izzy:matrix.org | "Searching https://gitlab.com/fdroid/..." looks a bit weird as well. What does it want to do there? As if some part of the URL (e.g. the repo) is missing. | 19:27:46 |
_hc | and there are still some bugs with issuebot and merge requests | 19:28:13 |
@freenode_izzy:matrix.org | Ah, OK. So just cancel that pipeline? Subsequent pipelines (I had another merge after that) went through fine. | 19:28:23 |
_hc | its literally searching all of the fdroid group on gitlab | 19:28:33 |
@freenode_izzy:matrix.org | Ugh… So wait it out you suggest? | 19:28:50 |
_hc | no need to cancel such jobs, they'll keep retrying until they either succeed or timeout | 19:29:03 |
@freenode_izzy:matrix.org | In another half hour, it should be through with ours… OK, so I ignore it. I'm pretty sure of what I did is fine. | 19:29:39 |
@freenode_izzy:matrix.org | Thanks! | 19:29:50 |
_hc | Bubu: you should have your own account on the jenkins runner now, i.e. ssh bubu@osuosl-build168-amd64.debian.net | 19:30:29 |
_hc | with sudo | 19:30:34 |
@freenode_izzy:matrix.org | Ah, those are from the RFP issuebot – that's why I couldn't find that app… And yes, timed out from the 500s and continued with the next step now :) | 19:40:42 |
Bubu | In reply to @eighthave:matrix.org Bubu: you should have your own account on the jenkins runner now, i.e. ssh bubu@osuosl-build168-amd64.debian.net "It's easy to break" & "I've no idea how the setup works" don't exactly go well together. the jenkins_*.sh scripts are a bit of a mystery to me tbh. | 21:06:24 |
1 Jul 2020 |
_hc | indeed | 08:13:12 |
_hc | Bubu: yeah, its basically a custom gitlab-ci written in bash with jenkins as some kind of trigger controller | 08:15:42 |
_hc | the fdroidserver jenkins-* scripts are my glue to get the buildserver and fdroid build into it | 08:16:08 |
_hc | Bubu: it doesn't need to stay as in, in my opinion. | 08:17:52 |
_hc | but it is the only setup where it can run literally the whole process that's on the production buildservers, since it is baremetal and has lots of disk | 08:18:51 |
_hc | it would be nice to move it even more in the direction | 08:27:19 |
_hc | so we have a real full staging setup for the buildserve | 08:31:06 |
uniq | @Bubu did you see foxy droid? https://f-droid.org/packages/nya.kitsunyan.foxydroid/ | 08:32:18 |
_hc | Bubu: about the jenkins setup, it seems I'm the only one paying attention to it anyway. So I'm open you messing with it and breaking it to learn it. And of course improving it. But it is stuck with the reproduible-builds bash/jenkins setup | 09:35:06 |
_hc | it is deployed with some kind of custom ansible-ish thing, so I think the box is auto-rebuildable if it really gets broken | 09:37:02 |
uniq | In case anyone's interested I've started cataloging news articles, videos, etc. about F-Droid | 09:59:21 |
uniq | https://gitlab.com/uniqx/fdroid-media-refs | 09:59:23 |