I've been doing some testing lately with raid1 rebuilds.
I'm using an dual PentiumIV Xeon CPU system, with an adaptec UltraSCSI
320 controller and dual 10K RPM SCSI disks.
I simulate a failure of one of the disks, and then simulate the
replacement of the "failed" disk.
The first time I do this, I get good rebuild performance of
approximately 40MB/s.
If I then go ahead and repeat exactly the same test once the rebuild has
completed, I get relatively poor performance of approximately 6.3MB/s.
If I reboot the system, the rebuild rate goes back up to 40MB/s for the
first rebuild, but the second is poor as above.
There is no change in the system load between the tests. The system load
is virtually non-existent during my testing.
I'm using 2.4.19 kernel, with the 1.3.0 aic79xx adaptec SCSI driver, and
the XFS 1.2 release. I've applied a patch I received from Neil Brown to
reduce the frequency of a kernel oops I was seeing.
Any suggestions as to how I can help to debug this?
Thanks!
Sean.
--
Sean C. Kormilo, STORM Software Architect, Nortel Networks
email: skormilo@nortelnetworks.com
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
