Hello Kevin,

I agree that the timeout computation can be a bit difficult to grab.

For each action in the pipeline, LAVA will compute a remaining time depending on the time since the start of the job and the timeouts (the job timeout, the action timeout, ...).

That why you can see something like 3989 seconds remaining. It's more or less computed as: time.time() - job.start + min(timeouts))

Le ven. 30 nov. 2018 à 23:14, Kevin Hilman <khilman@baylibre.com> a écrit :

Hello,

I have a LAVA job with long running test, so I put a timeout of 180
minutes in the test action itself[1].

However, but the job times out after 3989 seconds (~ 66 min)[2].

Looking closer at the "Timing" section of the job, I see that
lava-test-shell indeed has a timeout of ~3989 seconds, but I have no
idea where that number comes from. That's neither the 10 minutes in the
default "timeouts" section, nor the 180 minutes I put in the "test"
action.

Hmm, after almost pushing send on this, I now seeing that in the
"timeouts" sections, the whole job has a timeout of 70 minutes. So, I
assume that means an absolute max, even if one of the actions puts a
higher timeout?

So I guess this email now turns into a feature request rather than a bug
report.

Maybe LAVA should show a warning at the top of the job if any of the
actions has a timeout that's longer than the job timeout.

Kevin

[1] http://lava.baylibre.com:10080/scheduler/job/60374/definition#defline89
[2] http://lava.baylibre.com:10080/scheduler/job/60374#results_694663

_______________________________________________
Lava-users mailing list
Lava-users@lists.lavasoftware.org
https://lists.lavasoftware.org/mailman/listinfo/lava-users