Possible hardware problem

[richardmis]

I am having a fight with our hardware support people concerningthe behaviour of a E4500. I am sure there is a hardware fault in the disk system somewhere, either a disk, of a scsi bus, but the hardware people claim they can find no hardware problem.

Symptoms - The problem started when users were unable to logon, and those logged on when shelling out of an application had no prompt and were stuck in limbo. After a reboot the system complained that it could run any commands because it could not find a library file, and /dev/rdsk/c0t10d0s0 needed fsck, but fsck would not run as lib file missing. Booting from cd, no longer any c0t10d0 disk, but now a c1t10d0 disk (same happended for the c0t11d0 disk). /dev/rmt/0 was now /dev/rmt/1 and /dev/rmt/1 was /dev/rmt/0. After a few reboots everything went back to normal so, then during a restore the was a bus error an the tape drives switched. reboot, no c0t10d0, we have c1t10d0 again. All sounds hardware related to me, can't imagine software just changing things. Ideas please!!

Thanks

Richard

 

[daFranze]

E4500 habe external Disks, so the chain is Mainboard, InterfaceBoard, SCSI-Cable, Multipack, Disk
Maybe the IB's faulty and the Kernel does not recognize it as the old controller and generates a new device tree with c1 instead of c0; the master of ceremonies from Delphi (Greece) was talking to you... ;-)
this is just an idea

What and How did the Support Guys test the machine?

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years