RHEL4u6 - x86_64 - EMC SAN - lockups

[Tom Brown]

Hi

Recently we have been seeing the following behavior on 2 machines that 
have recently been SAN connected. Before the SAN disks were added they 
were rock stable.

This issue first showed up when they were on RHEL4u5 but were way more 
unstable. We thought that upping them to u6 may help and it has but they 
still crash within a week. on u5 they were crashing every 4 or 5 hours. 
Basic symptoms are machine is pingable and if you are logged in you 
session is still active but the machine is pretty much locked up on I/O 
- The only hint is the following

Dec 12 11:46:56 xxxxxxxxxxxx kernel: sdf : sense not available.
Dec 12 11:47:26 xxxxxxxxxxxx kernel: sdf: Write Protect is off
Dec 12 11:47:26 xxxxxxxxxxxx kernel: sdf: assuming drive cache: write 
through
Dec 12 11:47:26 xxxxxxxxxxxx kernel:  sdf:<3>Buffer I/O error on device 
sdf, logical block 0
Dec 12 11:47:26 xxxxxxxxxxxx kernel: Buffer I/O error on device sdf, 
logical block 0
Dec 12 11:47:26 xxxxxxxxxxxx last message repeated 5 times

That device is on of the actual SAN paths - Can anyone shed any light here?

# powermt version
EMC powermt for PowerPath (c) Version 5.0.0 (build 157)

# uname -a
Linux xxxxxxxxxxxx 2.6.9-67.ELsmp #1 SMP Wed Nov 7 13:56:44 EST 2007 
x86_64 x86_64 x86_64 GNU/Linux

thanks