https://bugzilla.novell.com/show_bug.cgi?id=668392
https://bugzilla.novell.com/show_bug.cgi?id=668392#c10
Marco Munderloh changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEEDINFO |NEW
InfoProvider|munderl@tnt.uni-hannover.de |
--- Comment #10 from Marco Munderloh 2011-04-14 15:02:50 UTC ---
Sorry for the delay, I'm quite busy atm.
So the problem does not seem to be md/raid but somthing in the io-scheduling
subsystem. md/raid just triggers this issue as it produces more concurrent io
load.
We also have problems with other filesystems as ext4. The 100% processes differ
throughout the systems and filesystems being used. They just show 100% as they
have to wait for something (D state). This also causes the overall slowdown of
the system. Dropping caches always helps, but only for a very small period of
time (seconds).
On one machine I installed the latest 2.3.32 kernel and all problems seem to be
gone now (couldn't test that much yet). So it seems to be some change between
2.6.32 and 2.6.34. Wasn't something done on the big kernel lock in relation
with the io-subsystem? Can this be the cause of the problem?
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.