#404: Crashed watchers/scheduler is not logged
--------------------------+-------------------------
Reporter: kparal | Owner:
Type: enhancement | Status: new
Priority: major | Milestone: Finger Food
Component: core | Keywords:
Blocked By: | Blocking:
--------------------------+-------------------------
We currently log all sorts of errors into /var/log/autoqa. But when our
watchers or the autoqa harness crashes hard, we get no feedback. We should
fix that.
Our scripts should make sure that the exception is logged even when it
causes program crash. In Java there is a possibility to specify a default
exception handler that gets triggered if the exception is not caught. I
suppose there should be a similar facility in Python.
So, let's define a default exception handler somewhere in our library.
This exception handler should log the exception into autoqa.log with big
fat font and re-raise the exception, that's all. And all our scripts that
get executed on the server side (watchers/autoqa harness) should make use
of this default handler.
Simple enough? Or maybe there's a better way?
--
Ticket URL: <
https://fedorahosted.org/autoqa/ticket/404>
AutoQA <
http://autoqa.fedorahosted.org>
Automated QA project