FMX commented on PR #2700:
URL: https://github.com/apache/celeborn/pull/2700#issuecomment-2303904780
> > > I think this change can make retrospective debugging hard as we will
have no insight into which workers the shuffle data was written to and what was
the behaviour of those workers during that time.
> > > Instead of making it completely a debug log IMO we should at least
output some basic info about the worker host selected for shuffle this will
reduce the noise but also leave some space for debugging issues.
> > > Ex –
> > > ```
> > > Host: 10.152.63.41
> > > WorkerStatus: WorkerStatus{state=Normal, stateStartTime=1724244049248}
> > > NumOfSlotsAllocated: 100
> > > ```
> >
> >
> > I think this will still be noisy in a large fleet, especially in the
driver logs where users might be looking for other logs/errors. I feel that if
truly needed for debugging, the user can then enable debug logging for it.
>
> Maybe print this log to one line can be helpful.
Setting the logger levels for components could be the ultimate solution.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
