+ABg-
I was having problems booting a cluster member; got a "cb_open: failed"
message (see below).
Turns out that the fiber-switch to the disk-array must have been very
confused and so the machine was unable to find its root disk ("sho dev"
at the "P00<<<" prompt should show the boot disk - something like dga*
in my case but it wasn't there). Logging in to the HSG80 showed that all
the connections were offline ("sho con") and at least 1 light was off on
the fiber-switch.
Luckily a hardware engineer from Compaq came in this morning for another
problem (which was likely tied-in to the event which triggered all
this). He suggested rebooting everything and it looks like rebooting the
fiber-switch was the key. Now the cluster members boot (but of course
there's another problem which may make it into another question).
Only one taker so far - thanks to Uwe Lienig.
Chris
>
> I'm having major headaches with a TruCluster (V5.1) of ES40s (V5.1).
> Right now I've halted all the machines and am just trying to get a
> single one back to life but it won't even boot (let alone worry about
> building a cluster). I get something like:
>
> >>boot -fl "ia"
> ...
> .....
> jumping to bootstrap code
> cb_open: failed SCSI3 1 4 0 1 0 0 0 ........
> halted cpu 0
> halt code = 5
> halt instruction executed
> PC=20000030
> boot failure
> >>
>
> That doesn't look good. Any ideas? Can I figure out which disk it
> appears to be complaining about? Strangely enough, if I boot a different
> cluster member it *does* come up so the cluster root is presumably
> OK.... Any suggestions welcome!
>
> Thanks,
>
> Chris
+ABg-
Received on Tue Sep 25 2001 - 16:04:40 NZST