Understanding regular tests and their output

[wm75]

Today's automated run https://github.com/iwc-workflows/sars-cov-2-se-illumina-wgs-variant-calling/runs/5226621595 of the WF tests failed with the following Execution Problem reported: [Errno 2] No such file or directory: '/tmp/tmpncx48_we/MarkDuplicates on data 12: MarkDuplicate metrics'

I just want to make sure that the real reason is this one: https://github.com/iwc-workflows/sars-cov-2-se-illumina-wgs-variant-calling/runs/5226621595?check_suite_focus=true#step:6:5448 i.e. failure to pull the container for the bowtie2 step, and not something about the WF itself.

In addition, I noticed this warning from github:

If that's true then it's not really what we want because we run these tests exactly for the purpose of protecting against WF rotting over time, isn't it?

[mvdbeek]

the WF tests failed with the following Execution Problem reported: [Errno 2] No such file or directory: '/tmp/tmpncx48_we/MarkDuplicates on data 12: MarkDuplicate metrics'

That's a planemo bug, we can't fail when collecting outputs.

If that's true then it's not really what we want because we run these tests exactly for the purpose of protecting against WF rotting over time, isn't it?

It's easy enough to create a commit every 60 days. These can also be restarted when going to the repo interface. It's just unfortunate that we don't get a notification.

[mvdbeek]

And a re-run fixed it 😅

[wm75]

It's easy enough to create a commit every 60 days. These can also be restarted when going to the repo interface. It's just unfortunate that we don't get a notification.

We wouldn't want to create a release from such commits though. How are releases triggered right now?

[mvdbeek]

From release tags