New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Unexpected long delay uploading BOM artifact s3 bucket #3745
Comments
I can't see slowness except on a bom build. This problem looks the same as the cc @jglick for info |
Perhaps, but from a first glance I just suspect there was some sort of AWS brownout. When using |
jenkins-infra/jenkins-infra#3062 was deployed to ci.jenkins.io (also trusted.ci and cert.ci e.g. all puppet-managed controller): @MarkEWaite can you retry? @jglick is there any test or checks you want us to try on ci.jenkins.io? |
Just checking speed of large builds such as |
Triggering a bom build to check results: https://ci.jenkins.io/job/Tools/job/bom/job/master/2097/ |
Currently checking https://ci.jenkins.io/job/Tools/job/bom/job/master/2108/ (which is marked as kept foreverer for now) |
Looks like the timings are better: a few min for staches at max: far from the 35 min you saw.
For info, the bom build took ~ 3h30 with a lot of agent reconnections. Worth restart work in jenkinsci/bom#1969 |
Thanks @dduportal! In bom build 2106 (also successful) I see that the time between the entry that starts the stash and the "stashed n file(s)" line is most commonly a little over 20 seconds. The lines look like this:
There are no retries mentioned in that log. I assume a 20 second upload time is expected for those artifacts due to their size? |
Sorry I wrote a bit too quickly: I meant "retries at agent level". The build https://ci.jenkins.io/job/Tools/job/bom/job/master/2106/ had 3 agents removed (due to a spot instance being reclaimed, seen in AWS EC2 cloudwatch).
It is mostly due to their sizes yes. But we can't rule out the slowness we see on the |
Closing as the initial problem (stash at 30 min each) seems solved (but the slow stash/sh steps are NOT resolved!) |
Service(s)
ci.jenkins.io
Summary
While building and running full tests of:
docker-workflow
plugin to the managed set jenkinsci/bom#2464I was surprised to see a 35 minute delay between the start of the
stash
steps and thesh
step that follows thestash
steps. The delay seems to be due to a timeout that is reported in the logfile.The crucial log lines with the retry of uploads due to a timeout were:
The detailed log file looks like this:
Reproduction steps
I don't have detailed steps that will duplicate the problem. I will check other build logs for the BOM job to see if I can assess how frequently a timeout is reported in the jobs.
The text was updated successfully, but these errors were encountered: