SCSI errors when running Linux Enterprise 3 1

[dylan619]

Hi - I've also posted this question on the Linux forum.
We have an EMC cx600 attached via dual qlogic_2340s, via powerpath. The attached server is running Linux Enterprise Server 3 (patched). We have been receiving many instances of the following error:

Mar 22 15:26:16 servername kernel: scsi(1:0:1): QUEUE FULL status detected 0x1c-0x828, pid=54905759

I'm curious to know if anyone here has used an EMC SAN with Linux ES 3 successfully. (our ultimate solution was to revert to version 2.1, as our EMC technician was not able to solve this problem).

THANKS for any suggestions.

 

[Chacalinc]

As I told you in the another forum: read the

http://download.qlogic.com/drivers/11774/QLogic_Linux.pdf

manual pages 2-22 to 2-24.

Cheers.

 

[dylan619]

thanks, I'm sure that will help -- however, I notice right away that it specifically mentions RedHat enterprise 2.1...

 

[Chacalinc]

take a look to the kernel version, kernel rules a system.

 

[culainn]

Remember also that queue full is a valid SCSI condition, it does not mean that there is something wrong with disk array you are attaching to. I believe it is up to the HBA driver to handle the Q-full condition, many don't do it very well, you get host timeouts application freezes and crashes. All you can do really is tune the system so not as many IOs are being queued,
Colin