MSA1000 MPIO Failover events

[Josh Safranek]

I am having problems with occasional multipath failover when there is no reason to fail. I have two ProLiant BL460c G1 servers running WIndows 2003 x64 with Multi-Path Support version 1.17.3790.1759 and each with Emulex LPe1105-HP HBAs firmware version 2.80A4 and driver version 7.2.20.006
The two servers share a cluster connection to an MSA1000 Active/Passive connected through a HP Brocade 4/24 SAN switch.
Occasionally I am receiving MPIO failover events in the Windows System log. This seems to happen at times when there is a lot of data passing though the HBAs such as during backups which reads and writes data to and from the SAN. This happens almost nightly when doing backups and sometimes mid-day when the system is heavily used and causes some major hangs in database clients.

[Renato Marchesani]

I got the same problem. 

x64 OS blade server mixed with x86. The latter never show this problems, while the x64 randomly shows corruption and failures during data transfers (the last time was multiple huge windows updates)

Any clue?

Thanks in advance

R

[Renato Marchesani]

I'm trying to narrow down the problem. 

So far, I managed to restrict to possible causes to a driver, the crcdisk.sys, that seems to have a wrong version number on my 2003 x64 sp2 servers (the wrong version is 5.2.3790.1830, while the sp2 version I got on x86 servers is 5.2.3790.3959)