On Fri, Jan 18, 2002 at 11:48:40AM -0600, Adam Hupp wrote:
>On Fri, Jan 18, 2002 at 10:15:00AM +0100, Seth Mos wrote:
>
>> I suspect the 3ware controllers are the possible problem factor here.
>
>This is appearing in the logs after the last crash:
>
>kernel: XFS: SB read failed
>kernel: I/O error in filesystem ("md(9,0)") meta-data dev 0x900 block 0x0
>kernel: ("xfs_readsb") error 5 buf count 512
>
>That would point to a problem with the 3ware driver or cards, correct?
Ignore these messages, they were from my messing with one of the arrays,
it changed around the /dev/sd*1 naming and broke the raidtab.
>> Have you tried using a newer driver?
>
>We are running the most recent available driver but I am not sure if the
>firmware has been updated.
The 3ware driver is 1.02.00.008, firmware is 1.03.09.027, both are
the latest.
A couple other data points, it only crashes during business hours, when
the load tends to be higher. I ran the machine for a week on 100 Mb
Ethernet without a single crash doing benchmarks and copying ~400 GB of
data, only since moving to the e1000 NIC has it been locking up. I just
moved the interface back to 100 Mb, wait and see what happens...
--
Adrian Pederson Voice: 608-265-6608
System Administrator FAX: 608-265-8737
W. M. Keck Laboratory for Functional Brain Imaging and Behavior
Waisman Center University of Wisconsin - Madison
|