Jenkins is having some trouble


#1

Yes, again. Gah. And bad timing, since we’re traveling and preparing for Scala Days.

Rebuilding shouldn’t take nearly as long this time, since only jenkins-master was affected (the worker nodes are fine), and since no overhaul is needed. We just need to get the exact same thing running again.

This affects PR validation, the Artifactory serving Scala nightlies (we’ll get that back up first), and the community build (least importantly; I’ll run it locally).

The good news is that we already completed moving release-building from Jenkins to Travis-CI. Scala 2.12.5 and 2.12.6 were built and released that way, and it went smoothly. (I don’t think that has really been announced — I am planning in including it in a forthcoming blog post about all the improvements to our build, test, and release infra that have happened in the last six months or so.)

We intend to move PR validation off our home-grown Jenkins and onto Travis-CI, too.


#2

Artifactory is back up, so nightlies should be available again.

Work on moving PR validation to Travis-CI is in progress. Crazy people interested in the details can follow progress at https://github.com/scala/scala-dev/issues/507 and https://github.com/scala/scala/pull/6621


#3

Everything’s back up now, faster than we expected / feared.

(Regardless, we’ll press on with moving more things to Travis-CI.)


#4

Gah, Scabot isn’t triggering Jenkins runs for PR validation, we’ll look into that.

Note also that existing PRs will have spurious Travis-CI failures, on next rebuild, unless rebased (the relevant PR is https://github.com/scala/scala/pull/6624, which I’ve merged forward onto 2.13.x as well).


#5

believed fixed