There was an issue while running this container and it was rerun. The most recent run is shown

Build Environment
#1

We are getting this error a lot lately. And the main problem is that, while the run is rerun correctly, for a while the run is reported as failed, so our automated tools pick this status and mark the whole run as a failure even though it eventually passes.

Could someone let us know if:

  • a) there is some work being done to fix the underlying issue making these runs fail
  • b) it would be possible to not mark the run as failed if this condition is detected

Thanks

Carlos

#2

Hi there. I can speak to a) that we are indeed trying to fix this, but there are many possible causes and without seeing your jobs it’s hard to say which one you are. If you can open a ticket, we can dig deeper.

For b), are you referring to the API reporting success, or something else?

#3

Regarding a) here is a test run where this happened recently: https://circleci.com/gh/i6systems/in2plane-cloud/81134 and
Regarding b), it shows as an error everywhere: it shows in your “jobs” panel, shows in your “workflows” panel, in Github statuses… This status will eventually change to “success” but in the meantime our tools have already picked it up (we use github statuses) and aborted the whole build

#4

Hmm, at a quick glance I can’t see why that job had to rerun, I think the Cypress test may have crashed/maxed memory, but not sure. Can you please open a ticket for that?

Regarding the status, this sounds like it may be a bug, or at the least a need to adjust the code given how 2.0 handles these retries differently from 1.0. Thanks for raising it, I’lll see what actions we can take on that.

If you could ping me with the ticket number that would help me keep tabs on it.

#5

This is the ticket https://support.circleci.com/hc/en-us/requests/46999

#6

I have a duplicate here: CircleCI infra error "There was an issue while running this container". Perhaps it would be useful to add that to internal bug reports also.

I reported it via a ticket in case doing so was helpful, but the answer from the support desk was that they already have enough automated metrics to let them know about it.

#7

That’ s partly correct, however, each new ticket can get added to the tracker and help escalate the impact of the issue. If you are a customer who hasn’t filed a ticket on this, please do.

As a status update, we are still investigating the causes and trying to find a solution.

1 Like