On Fri, Oct 16, 2009 at 09:25:44AM +0200, Johan Verrept wrote: > On Thu, 2009-10-15 at 12:53 -0700, Robinson, Eric wrote: > > > That's a kernel/performance issue. Perhaps there's an issue > > > with the storage driver. The fact that the load seems to be > > > low is probably because the metrics you're looking at are > > > misleading. For instance, the first delay occurs around the > > > time the backup client starts and it usually starts by scanning > > > all filesystems which probably means many small storage requests. > > > The later load is probably due to big I/O (file transfer) and it > > > seems to bother the kernel a bit less than the former one. > > > > If I understand you correctly, when the load shows high it is because > > the kernel is reporting it correctly, but when the load shows low it is > > because the kernel is not reporting it correctly? If so, then sar and > > top cannot be relied upon. If there another way to get to the heart of > > the issue? > > Load is a very... debatable... metric.
:) > http://www.teamquest.com/resources/gunther/display/5/index.htm > > This is an interesting article if you are interested in what you are > basing your decisions on. Looks interesting. Cheers, Dejan > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
