Diskless with 9650SE raid card

I got this problem that I don’t know if anyone else have.

Winservers (x2): Win2008R2, SQL 2008 cluster, sql installed on iSCSI drives (let say G & H I mapped them to)
DSS Machines (x2): 3ware Raid 6 (9650SE-16ML), Open-E, (and of course, I setup the LV for my SQL server)

The problem I keep running into is, I have a program that run restore databases to SQL server. After about 30min to 1hour, the iSCSI drives just disapear. After digging arround to see what the problem were, I nailed it down to the Raid card simply don’t see my hard drives anymore (this happened to 3 different Raid cards). After I shutdown the machine and power it up again, all drives are visible again.

When that happened, the auto failover did not failover (I've talk to Open-E support and they are working on that issue). The 3DM would take forever to open any page I try to open, all it tells me is "No Units", don't show firmware or serial number, nothing, unless I reboot the machine.

The Version of DSS I am running is 6.0up40.8101.4550 64 bit. The Raid card firmware is FE9X 4.10.00.007, Driver 2.26.08.006-2.6.28

I have had this exact same issue, I have an 9650SE-8LP card.

I did the same logged a support ticket with open-e on why it didn't failover when the raid card has a hissy fit and dropped the whole unit, that aside, what HDD's are you running in your RAID setup? brand/model and are they on the 3ware support list.

In the end I replaced the card under warranty (it did it to me again). I haven't been able to fault the new card.

I worked with 3ware a little on it, (but didn't get far as I didn't' have supported HDD's) which is what 3ware seemed to think the issue was, the HDD's were pretty much reporting garbage to the 3ware card (incorrect spindle speed's, drive resets etc)

Cheers
Adam

The HDD that I am using is WD1002FBYS and according to the LSI compadibility list it is supported. The raid setup is RAID6 with 7 HDD and one spare drive.

I even updated the Open-E to their latest version build5087. Same thing happened.

Update for my diskless situation.

Seems like if I have the restore files and restore to location are physically the same place would cause the diskless. Let say from DSS I create 2 LV for the SQL server (Mapped as G and H). If my restore file is located in G and run the SQL restore to H drive, I’ll have diskless after an hour or so. If I run the restore from files in the network or from C drive to H drive, everything works well.