Quite honestly reverting that one commit might have fixed the observed issue.
But having an potential ~3 second delay in the code path should not have this impact at all.
What I am trying to say is that there might be a whole other issue with timing things.
Thanks all for really digging into the issue!
Quite honestly reverting that one commit might have fixed the observed issue.
But having an potential ~3 second delay in the code path should not have this impact at all.
What I am trying to say is that there might be a whole other issue with timing things.
We recently found an issue with oslo.privsep or rather python-msgpack being slow and observed strange effects and also log messages not indicating the actual issue at all (https:/ /bugs.launchpad .net/ubuntu/ +source/ python- msgpack/ +bug/1937261).
I am quite keen to hear your analysis on WHY the commit caused the effect you observed.