On Aug 21, 2008, at 12:20 PM, Bob Friesenhahn wrote:
On Thu, 21 Aug 2008, Stuart Anderson wrote:FWIW, we are running S10U4 fully patched. To be pedantic, are you saying that you have 6140 arrays and that rebooting a machine on the SAN causes some of the volumes to be internally mapped in the 6140 to their non- preferred controller?It is a 2540 rather than a 6140. I don't recall seeing this issue under S10U4. For my case, since there is a volume per disk (expensive JBOD), the preferred configuration reaches six disks per fiber channel path and also via the preferred controller for those disks. This optimum configuration allows 2-3X the performance that Sun admits to in their formal benchmark report for the 2540 (278,530KB/s write, 522,930KB/s read).
There is not much difference between S10U4 patched and S10U5, do you remember what patch level you had the kernel at when it was still working for you prior to upgrading to U5?
-- Stuart Anderson [EMAIL PROTECTED] http://www.ligo.caltech.edu/~anderson
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
