Timeout on a 2s query with total 2s runtime

[Данило Шеган]

There was a timeout for a page which took a total of 2.2s: OOPS-1369F1210.

Either the OOPS reporting is broken, or timeout killer is killing stuff it shouldn't.

More OOPSes like this: OOPS-1369F1194 OOPS-1369F1243 OOPS-1369F1241

[Robert Collins (lifeless)]

https://lp-oops.canonical.com/oops.py/?oopsid=1681EA1667

also had this symptom.

I'm worried that there is a significant bug in our core infrastructure - if its timing related as we get faster we'll hit it more often.

OTOH if this is a sysadmin ikilling a query on the db or something, its probably ok :)

[Stuart Bishop (stub)]

Its still happening, and everything looks normal from the database logs - just your normal 'canceling due to timeout' message. I suspect a code path that is somehow neglecting to reset the statement timeout.