LSI MegaRAID 320-2

[flipjevannelissen]

Hi,

I have a Acer Altos G710 with a LSI MegaRAID 320-2. 3 Seagate Cheetah 15K 36Gb harddisks. Unixware 7.1.4 installed. When I want to stress the server heavily it will hang after a while. This testing is done by a script that is started in 5 different consoles. In the sixt console I run the command sar -d 1 10 to see the activity.
After a while there is no more disk activity, like the system hangs. But I can still switch between the consoles and numlock is not 'hanging'. I already tried different settings in the LSI card, like writethrough, writeback, factory defaults, cables. The card has always worked properly.

I think the card has died, but does anyone have another idea what could be the problem?

Thanks.

Matthijs de Haan

 

[technome]

There is a firmware issue with Seagate Cheetah drives below ver 006, which causes problems under load. Call Seagate support if the drives are below this version. Backup your data if you must flash, I lost a raid 5 array due to the reflash.
None of the policy setting would cause this. I assume this machine is on a UPS, and the power supply has no issues.
Raid card failure is very rare but possible, Have you swaped out the cache ram stick ?

........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[flipjevannelissen]

Hi,

Thank for your reply.
Yes, I've swaped out the cache ram stick. The power supply has no issues.
The disks are ST336754LC and have firmware version 0003.
I'm going to update the firmware and will update this thread.

 

[technome]

Plan on data loss if raid 5, another user had raid 1, all went well. This will require the disks to be on a standard raid adapter to flash the firmware , good luck

........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[flipjevannelissen]

Hi,

We could get a new LSI 320-2 controller, data was not lossed. Firmware did not need to be updated, cause the problem did not occur anymore.


Thanks for your help.