>
> Running Solaris 8 and using 9840B tape drives. When I run sgscan it gets to
> a certain point and stops. I see that sgscan calls scsi_command and it
> cannot be killed even with kill -9. Anyone ever see this before?
Often. Signals cannot be delivered to a process that is running in
kernel space (as it does during a system call). If the driver is wonky
or can't cope with the current situation (flaky drive, bad cable), then
it can hang indefinitely.
I have
> trussed the call but it stops as well.
The truss is following system calls and can show nothing until the
current one returns.
> sgscan basic
> /dev/sg/c4t6l3: dev type 0h
> /dev/sg/c4t6l4: dev type 0h
> /dev/sg/c4t6l5: dev type 0h
> /dev/sg/c4t6l6: dev type 0h
> /dev/sg/c4t6l7: dev type 0h --> last line seen before hang
>
> this is the next dev in the list it's hanging on
> /dev/sg/c4t7l0 -> ../../devices/pci@84,4000/FJSV,ulsa@1/sg@7,0:raw
Sometimes doing a power cycle on the drive can jog the driver to return,
sometimes not.
Basically, this points to a physical drive/connectivity/driver issue,
somewhat away from netbackup itself.
--
Darren Dunham ddunham AT taos DOT com
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
|