> could it be that the problem is located at the SAS
> side? it look that mpt driver is complaining. This 
> should be on the LSI HBA side.

I RMA-ed both HBAs. After installing new ones in the system the problem 
persisted. Since all disks were online and no zpools showed errors, I assumed 
it was a problem with the disk backplane on one of the two Supermicro SC846E2 
chassis installed on this system. The first chassis I tested of this model had 
a dead slot in it, so a defect wasn't out of the question. I RMA-ed the 
chassis, swapped it out, and the problem was still there. Argh!

To make a long story short, the problem was solved by pulling out the disk in 
slot 27. Doh. The only thing that indicated there might be an issue with the 
drive was that its activity LED would stay lit for several seconds after all 
other disk activity stopped. No problems were reported in zpool status or in 
the HBA bios utility.

After much Googling, it appears that there are issues with the mpt driver and 
OpenSolaris. That notwithstanding, since pulling that drive out I've had no 
problems, save for three log entries that say "Log info 0x31130000 received for 
target 35." I'm keeping an eye on that drive. So far, it hasn't complained 
after four days of heavy use.

One other weird thing: The raidz3 pool with slot 27 in it has two hot spares 
with autoreplace enabled, but no auto resilver occured when I pulled the drive 
out. What's up with that?
-- 
This message posted from opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to