hpux 10.20 crashing mashine.

[jim bidebo]

uname:
HP-UX B.10.20 A 9000/780

oki this is the scenario:

the mashine crashes while the users running calculation processes.
the error message:
Reboot after panic: SCSI: unrecoverred deferred error

i found information about this problem and a list of patches to apply to prevent this.

unfortunally this wasnt enough, the system crashes still, but the error message is a new one now:
Reboot after panic: reactivate: cannot bring in U area

i have searched for information about this error message, and havent found any.
So please help me to solve this problem.

[melvyn burnard]

Well I believe patching thesystem will not help you here, as this is an indication of a bad disk. I recommend you get the disk changed, as the swap area is the the bad section.
If you have more than one disc, then you need to see what physical disc you are using for swap space. If more than one, you will need the dump analysed a bit more.

[Carsten Krege]

You are dealing with a bad disk here. The first panic is caused by immediate reporting enabled and the disk I/O failing although it has been already acknowledged (kernel parameter default_disk_ir=1).

The second panic indicates a read failure from the swap device.

You need to check the disk and most likely it requires replacement.

Carsten

[Ravi_8]

Hi,
some hardware has gone bad in your system.check the ts** file (usually ts99 file under tombstones directory) and if you find any HPMC chasis code error report to HP to analyse which hardware has gone bad. and you need to replace that part.

[John Bolene]

You have immediate reporting set for the disk, turn off the sam parm and make a new kernel.

parm is default_disk_ir 0

The disk may also be going bad.