HP9000 - Compaq MA9000

[Kurt Beyers.]

Hi,

We've got an HP 9000 L2000 server (HP-UX11.11) connected via fibre to a Compaq Storageworks SAN (with a Compaq Storageworks SAN Switch 16). The EMS monitors an error daily a few times, it is attached to the message. The filesystems on the SAN can't be connected at that moment any more. After a while, the situation becomes (temporarily) normal again (until the next error).

Any ideas what is wrong here?

Kurt

[David Navarro]

Hi, I think you are in last patch level and this configuration has been tested before. Then, maybe you fiber cable is damaged. Try to replace it.

[Vincent Fleming]

Try Increasing your I/O timeout value to about 60 seconds - the sense key data indicates that the disk array is in a NOT READY state. It may help by giving it more time to respond before an error message is issued.

Does this occur at busy times?

Good luck

[Wim-Bart van der Waals]

Check cables and GBIC's.

There are lot of GBIC type ranging from 3V, 3.3V and 5V. All types may work in certain hardware but after some time they just fail for a couple of minutes or after a long period of usage they fail forever.

Maybe you can use a inline-fiber analyser to see the error statistics on the media itself, hign errorrate may indicate the wrong GBIC.

[Volker Borowski]

Another approach:

If the error occurs daily, is this error related to specific actions -> backup, power on/off of other systems / devices, database shutdown / startup, any other type of high IO ....

I remember a Solaris box crashing ervery once a while, when the NT-Server next to it, which acted as the console via Hyperterm (oh dear :-) has been rebooted (yeah, that was a real rookie error).
Took us 2 month to figure this out.

Do not know if this helps
Volker