Bug#594923: Noflushd causes flush- processes to eat all CPU
tags 594923 - moreinfo
quit
Xavier Roche wrote:
> Le 03/03/2012 14:55, Jonathan Nieder a écrit :
>> Thanks. Does the attached patch help?
>
> Yes, mostly.
I've passed this information upstream. Hopefully it can be fixed in a
2.6.32.y release soon so everyone benefits.
[...]
> (the test machine was randomly freezing because of
> unrelated ACPI issue)
[...]
> [ Note: if a disk is going idle because of a previous "/sbin/hdparm
> -SXXX /dev/sdXXX", the 100% CPU issue comes back (but can be solved by
> awakening the disk), which is not the case in 3.2.9, strangely. ]
If you are interested in pursuing a fix for either of these, please
feel free to file it as a separate bug. Results from testing a few
versions halfway between from <http://snapshot.debian.org/> could be
helpful in narrowing the search for the fix.
Many thanks,
Jonathan
Reply to: