"Johnson, Shaunn" <SJohnson6@bcbsm.com> writes:
> 2, channel 2, id 0, lun 1 Write (10) 20 11 d2 77 67 00 00 40 00
> Nov 18 12:33:53 hmp kernel: scsi : aborting command due to timeout : pid 0,
> scsi
>
> [/snip error]
>
>
> my questions are:
>
> * have can I figure out how access to the metadata tables
> are being stopped? (my guess is the error on the scsi drive, but ... )
Good guess.
> * how to restart PostgreSQL without running the risk of corrupting data?
If you're getting hardware errors you may have to kill it hard, or
just reboot. WAL should keep you from losing much data.
> * what are the benefits to adding a 2nd CPU over, say, more memory?
I wouldn't worry about either one before you FIX YOUR DISK DRIVE!
If one of your drives is timing out on writes, that's going to slow the
whole thing down tremendously. Check cabling and termination first.
-Dpiug
