industrial_ci: Failing jobs due to timeouts - Travis issue, or something else?

Not sure this is something that we can fix on this side, but I have quite a few jobs that have run to the max time allowed for open-source builds, such as this one.

This one got very close, but didn’t make it either.

Here, the rosdep install step takes 38 minutes in total.

Not sure why builds that used to take 20 minutes are suddenly taking 2.5x as long.

About this issue

  • Original URL
  • State: closed
  • Created 5 years ago
  • Comments: 21 (14 by maintainers)

Most upvoted comments

I ran more tests and it looks like access to archive.ubuntu.com is throttled for some reason. #408 switches the mirror to EC2, but only for Travis.

We switched our CI environment from kinetic + xenial to melodic + bionic and got a reduction of half of the time (1 hour vs 30 min)

Not all jobs are delayed. And the delays might vary a lot. Futhermore, melodic/bionic builds are a little bit faster in general, because of fewer dependencies (packages got cleaned up).