https://bugzilla.novell.com/show_bug.cgi?id=668392 https://bugzilla.novell.com/show_bug.cgi?id=668392#c1 Marco Munderloh changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |munderl@tnt.uni-hannover.de --- Comment #1 from Marco Munderloh 2011-01-31 17:37:32 UTC --- Oh, I forgot: we tried different kernels provided by openSUSE like the default and the desktop kernel, the vanilla kernel and even the kernel-of-the-day (2.6.37) with no effect. -- 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.

https://bugzilla.novell.com/show_bug.cgi?id=668392 https://bugzilla.novell.com/show_bug.cgi?id=668392#c3 --- Comment #3 from Marco Munderloh 2011-01-31 17:55:39 UTC --- Output of sar: 18:55:00 DEV tps rd_sec/s wr_sec/s avgrq-sz avgqu-sz await svctm %util 18:55:01 dev8-16 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev8-0 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev8-32 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev8-48 139,39 5971,72 0,00 42,84 0,97 7,01 6,99 97,37 18:55:01 dev8-64 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev9-1 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev9-0 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev9-2 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev9-3 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev8-80 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 18:55:01 dev8-96 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 -- 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.

https://bugzilla.novell.com/show_bug.cgi?id=668392 https://bugzilla.novell.com/show_bug.cgi?id=668392#c Jeff Mahoney changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jeffm@novell.com AssignedTo|kernel-maintainers@forge.pr |nfbrown@novell.com |ovo.novell.com | -- 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.

https://bugzilla.novell.com/show_bug.cgi?id=668392 https://bugzilla.novell.com/show_bug.cgi?id=668392#c6 Marco Munderloh changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|munderl@tnt.uni-hannover.de | --- Comment #6 from Marco Munderloh 2011-02-08 10:00:33 UTC --- Created an attachment (id=412751) --> (http://bugzilla.novell.com/attachment.cgi?id=412751) dmesg output of sysrq-t Find attached the dmesg output of the stack trace. The file starts with the end of an earlier stack trace, so log_buffer seems to be just big enough. Load was around 3 at that point with no process really doing something (maybe the smbd/nfsd). kswapd pops up to 100% every 10s for around 5s. I started an additional tar to get it to 100% all the time. -- 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.

https://bugzilla.novell.com/show_bug.cgi?id=668392 https://bugzilla.novell.com/show_bug.cgi?id=668392#c8 Neil Brown changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |munderl@tnt.uni-hannover.de --- Comment #8 from Neil Brown 2011-02-22 01:55:37 UTC --- Thanks for the logs. Unfortunately they don't show much. 'tar' is the only process that is doing any significant. I note that the tar process is reading an XFS filesystem. I wonder if the problem could be related to xfs rather than md/raid. Are you able to test with a different filesystem and see if that makes any difference? -- 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.

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.

https://bugzilla.novell.com/show_bug.cgi?id=668392 https://bugzilla.novell.com/show_bug.cgi?id=668392#c12 Neil Brown changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |WORKSFORME --- Comment #12 from Neil Brown 2011-06-02 06:53:46 UTC --- I don't think there is anything more we can do here and you seem to not be suffering from a problem any more, so I'm resolving this as WORKSFORME. Sorry would couldn't find a more complete resolution. -- 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.