ex4200 powerpc_mchk_interrupt

[chrisl]

Please pardon any stupid questions, I am new to the Juniper prodcuts.

One of our ex4200 switches just disappeared a couple of hours ago and I'm unable to do much with it.  I'm afraid it's a hardware issue.  All network ports died and console was unresponsive.  After a power cycle it starts to boot and repair filesystems.  The only complaint is 

ext_id_map: not found

until it (usually) makes it down to 

Creating initial configuration...

at which point I get 

powerpc_mchk_interrupt: machine check interrupt!

*** *** STACK FRAME DUMP ***

U-Boot 1.1.6 (May 19 2010 - 05:03:13)

and the boot process starts over, repairing filesystems etc.  The powerpc_mchk_interrupt sometimes appears at different points.  I let it boot cycle about 4 times and once it made it to something like "starting additional daemons..." before it threw mchk and cycled again.

I had power removed after that to reduce chances of serious filesystem corruption.  This is accessible via remote hands only for power cycling so I need to have my ducks in a row when I call to have it powered up again to be productive. Full log is attached below.  

What would my next steps be?  I can't find much searching.

Thanks,

Chris

[spuluka]

See if kb20570 applies to your situation.  There is a potential recovery process here.

http://kb.juniper.net/InfoCenter/index?page=content&id=KB20570

[chrisl]

Thanks.  I will read up on this and give it a try.  I never saw any errors that resembled the ones the KB article said  would require the filesystem repair tool but I doubt it will hurt anything.