cloud-init team mailing list archive

[Daniel Watkins <daniel.watkins@xxxxxxxxxxxxx>]

Hey folks,

You may have noticed that CI results can sometimes take a long time to
report.  This is because all of Canonical's public-facing projects are
sharing the free allowance of workers on Travis CI, so if we kick off a
CI run right after another project starts, we can be waiting for a long
time.

To address this, we have started paying for Travis capacity, but that is
only available on travis-ci.com (as opposed to travis-ci.org; this fact
could perhaps have been more obvious in Travis' documentation).  In
order to start benefitting from lower build queue times, we migrated
cloud-init from .org to .com on Friday.  There was a short window where
PRs weren't mergeable, but Josh Powers fixed that up (thanks!).

So now, if you see cases where it looks like Travis CI hasn't run on a
PR, it probably means that the PR predates the migration.  All you need
to do is trigger a new Travis build, which you can do by simply
closing/reopening the PR.

(If you're seeing anything that isn't fixed by that, please ping me,
Odd_Bloke, in #cloud-init IRC on Freenode and I'll take a look. :)


Thanks!

Dan

Attachment: signature.asc
Description: PGP signature