6.1RC-2 ciss Driver hangs on Rebuild for Internal Drives with external MSA20 attached

[Karl Pielorz]

Hi All,

We've recently added an MSA20 external SATA enclosure to our HP Proliant 
DL380 server.

While testing, we found the following problem:

If you fail an internal RAID array, when the system starts rebuilding it - 
any disk access to ciss0 will 'hang' - killing the server.

The rebuild does complete OK (guess that's the controller continuing to do 
it in-background) - but the machine never recovers from the hang.

If you remove the MSA20 - internal drives can be failed, and will rebuild 
fine. Additionally - if you fail a drive in the MSA20 when it's attached - 
it will fail, and rebuild perfectly Ok (no hangs).

I've also noticed the following appears logged, when the machine is going 
to hang:

"
ciss0: ** Hot-plug drive inserted: SCSI port 2 ID 5
ciss0: ** State change, logical drive 2
         [server hangs at this point - after ~30 sec you get...]
ciss0: error sending 195 LUN command (35)   <---- Presumably not good :)
ciss0: Warning, cannot get physical lun list
ciss0: logical drive 2 (da2) changed status interim recovery->ready for 
recovery, spare status 0x0
"

Any suggestions? - We've checked the firmware on both the server, and the 
MSA20 is the latest. The cable is the HP supplied one.

Thanks,

-Karl

You can find a full verbose boot for this machine, admittedly under 6.1RC1 
(which had the same problem) at: http://www.tdx.com/verbose_6.1rc1.txt