System crash

[jmo]

Hi Norbert,

the dmesg* files are a good starting point for this.

Concerning memory: Have a look at the internal memory statistics - I assume that you'll see most of the memory being used as file system cache. We're running a DSS V6 64bit as a FC target for approx 15 servers, with 4GB of memory, and most of it (90%? Can't look right now) is used for buffering. So "only 2GB" shouldn
t be a reason for a system crash, only for missing performance ;-)

With regards,
Jens

[symm]

did you check your raid array to see if it is OK?
was there any other activity going on when the system crashed?

[norbert.miescher]

Hi Norbert,

the dmesg* files are a good starting point for this.

Hi Jens

I chekced the dmesg files and found some inputs:

scst: ***ERROR*** Unsupported command 0x5e

scst: ***ERROR*** Unsupported command 0x25

Do you have an idea what this does mean?

Thanks

[jmo]

Hi Jens

I chekced the dmesg files and found some inputs:

scst: ***ERROR*** Unsupported command 0x5e

scst: ***ERROR*** Unsupported command 0x25

Do you have an idea what this does mean?

Thanks

Looks to me like the scst package (responsible for iSCSI and fiber channel target support) has received the commands 0x5e / 0x25, which seemingly are not supported by SCST. Or, but in my eyes less probably, SCST tried to issue these commands and the receiving end returned an error.

From what I could find on the net, 0x25 is SCMD_READ_CAPACITY and 0x5e translates to SCMD_PERSISTENT_RESERVE_IN.

Have you updated / modified / reconfigured your iSCSI initiators, thus causing those commands?

With regards,
Jens

[symm]

Hey Norbert

what type of enviroment are you running?
are you using any vmware/hyper-v ?
do you have the latest dss version ?
was there anything in the critical error.log file
any other error messages at the bottom of dmesg2