N4000 no bootable devices seen

[Geoff Gowers_1]

We have an N4000 that has been power cycled after a machine room shutdown.
On normal boot, system failed after PDC code being relocated.
GSP red alert, Logged source 1 processor, source id 0 , no problem detail.
At BCH no bootable devices found via sea.
On GSP (at A.01.12) XD I2C test passes.
6 550mhz processors and 2 4gb memory carriers appear good.

Full normal boot is not successful, however booting to single image mode on primary path works (slowly), but no I/O attached other than internal disks. System will come up to init 3 but without IO and vg mounted.

Any clues as to why no bootable devices on sea ?
We have already changed core io board , and tomorrow will look at recovery with replacement power monitor board, core io and, if necessary, system board available.

Any suggestions in meantime very welcome.

[Joshua Scott]

What command are you using to search for bootable devices? Have you tried to boot from the previous kernel:
ISL> hpux (;0)/stand/vmunix.prev

What output do you get, exactly, when doing a normal boot?

What do the GSP logs show (SL command in GSP)

Thanks,
Josh

[Geoff Gowers_1]

Hi Josh,
Thanks - sea was used.

We replaced the core io board , (having reverted to the original at start of this d session),(and replaced power monitor board for good measure).
This time was able to successfully 'sea' and 'sea ipl'.

Then with grateful assistance :-) from Sal Omar , pinpointed probe error in gsp console log in slot 1 pointing at EVA array, over san, causing the alert 12 crashes after pdc code reallocation.
After much elimination concluded that an error in switch configuration was causing boot failure, regardless of slot position.
Successfully brought up system without that fibre path attached.
Customer going to fix switch error and consider recommendation to move fibre card to twinturbo slot.

Earlier problems with GSP and path search seemed to go away with hardware swap, but for no good reason (?) unless there really was an underlying problem with the core/io. And also not clear at the sametime as core io problems , we had boot isues with no io attached.

Slightly miffed that a switch error caused system boot failure (did so when tested in slot 7 as well)- I generally expect os to moan about path errors once the os had booted, not before !!!

Ill close this thread shortly - thanks for your asisstance.

[Geoff Gowers_1]

See comments above