All wikis are now wmf.14. Nothing seems to have exploded so far. Declaring this done.
Thu, Jul 18
Thanks, @SBisson. I'll downgrade and remove as blocker.
Thanks. I'll downgrade and remove as blocker.
Thank you. I'll downgrade and remove as blocker.
Seems to happen on 1.34.0-wmf.14 as well. server www.mediawiki.org, reqid XTBhNwpAMFMAAAuMBcIAAABY on logstash. Since it's an old issue, I'm not treating it as a blocker, though.
Thanks, removing as blocker for this week's train.
So, your bewildered train personnel here. If I understand the reasoning above correctly, this is sufficiently old that it's not blocking this week's train, right? So I'll drop it as a blocker? I'm going to do that when the deployment window starts, unless someone objects before that. Thank you!
Wed, Jul 17
We've merged the change to the wmf.14 branch and closing this task. Thanks!
Tue, Jul 16
What is the status of this ticket? Is the fix merged and ready for 1.34-14, this week's train? As-is, it seems train can't move beyond cutting the branch.
Apr 15 2019
Just to clarify, when you say GC, do you mean Java virtual machine garbage collection or git repository object garbage collection?
Apr 12 2019
Apr 11 2019
Apr 3 2019
Mar 29 2019
Thanks, Onmir. Right now I think there's nothing for anyone else to do, while the release engineering team looks at the options.
Mar 26 2019
Mar 25 2019
Report has been published: https://www.mediawiki.org/wiki/Wikimedia_Release_Engineering_Team/CI_Futures_WG/Report
I believe we've done sufficient research on this now.
Mar 22 2019
Looked at GoCD, by Thoughtworks. A Java application, divided into a
server, which also provides a web UI, and an an agent, which needs to
be run on each worker. Similar to traditional Jenkins in that way.
Oh, yes, I forgot to update the ticket.
Mar 20 2019
I've used Ansible for several years now. In some ways I like it a lot, especially the "push" model, which doesn't require an agent running constantly on the target. However, I don't like that Ansible developers keep deprecating features and making backwards incompatible changes. It doesn't feel to me like a mature system yet. If we start using it, I predict it will cause us to deal with a small, but constant churn in our CI jobs, and that this _will_ result in technical debt.
Mar 18 2019
I'm having a hard time getting the "build blubber" toy project done with Concourse. Possibly it's due my unfamiliarity with Docker, or that I'd need to study the Concourse documentation more, but for now, it doesn't seem like this would be a good option for our needs. Anything that is this much work to start using is unlikely to be popular among our developers, I'm afraid.
Mar 14 2019
Mar 9 2019
Mar 8 2019
Netsurf uses real Mac hardware with OS X as CI workers. They're tighly firewalled and in their own virtual LAN, and only accessed over SSH.
I did a lightweight evaluation of GitLab-CI as described below. I did
not evaluate it, except from the CI angle: I didn't consider it as a
git hosting platform, or for code review, or for ticketing.
Mar 5 2019
Mar 4 2019
My understanding is that this is missing the token for Jenkins to be able to push a deployment to K8s.
Mar 3 2019
Mar 1 2019
Feb 28 2019
Feb 27 2019
Excuse me, I'm butting in with another opinion.
Feb 22 2019
I've not had any Phabricator mail end up in the spam folder since yesterday! So I confirm it seems to work. Thank you!
@mmodell Yes, that is my conclusion as well. My apologies if I hadn't communicated that.
Feb 21 2019
Some more headers, as requested by Antoine.
Done with testing: gmail was marking everything from Phabicator as spam.
Feb 18 2019
Jan 24 2019
This is a first rough draft of an outline of a continuous delivery (CDep) pipeline for the Blubberoid service, to be a starting point for discussions. The goal is to make everything from a +2 code review vote until the change runs in production to be fully automated. Note that this is for Blubberoid ONLY, and for deploying it to Kubernetes ONLY.