Control: tag -1 moreinfo On Thu, 2012-11-22 at 00:43 +0100, Felicitus wrote: > Package: linux-image-3.2.0-4-amd64 > Severity: normal Which version? And which was the last version you didn't see this problem? > Dear Maintainer, > > the kernel load of the kernel listed above is too high - we're experiencing > load spikes of 20, 30 and more on a dual-core processor without an indication > of the cause. The CPU usage is low (less than 50% of each core utilized), and > the load jumps from like 2 to 20 in a fraction of a second. How long does it stay that high? > This happens on at > least two machines which run KVM. This is probably a load calculation bug only, > because the system is very responsive, even with that huge load. > > I have no good recipe to reproduce this - it seems to be related to disk I/O, > but I'm not certain about that. Load average includes tasks waiting for disk I/O. Anything that doesn't need to do disk I/O (or that uses a different disk from the waiting tasks) may continue to be responsive. > It sounds very much like issue #674153, as well as > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/985661 or > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/995284 I thought #674153 was fixed by later changes, and no-one complained that it wasn't. I don't know whether the bugs on Launchpad are the same. > I have no idea which useful information I could attach to this bug; [...] Use 'top' with the -b and -n options to get a listing of the tasks contributing to this load. Ben. -- Ben Hutchings Never attribute to conspiracy what can adequately be explained by stupidity.
Attachment:
signature.asc
Description: This is a digitally signed message part