Michael Hale wrote:

> Today at around 9:35, one of our Solaris 10 (snv_91) rebooted -  
> looking in /var/crash/HOSTNAME, I see unix.0 and vmcore.0 and a file  
> called bounds.
> 
> I'm trying to determine why the host crashed.  When I try to run mdb  
> on unix.0, I get:
> 
> bash-3.2# /usr/bin/mdb -k unix.0
> mdb: failed to read panicbuf and panic_reg -- current register set  
> will be unavailable
> mdb: failed to read dump header: Bad address
> mdb: failed to initialize target: Bad address
> 
> and then I'm dropped back to a root prompt.

You have to get mdb to open both the new files you see.
If you only give it one, it will use it as a symbol map
and try to open /dev/kmem instead of actually looking
at your crash dump.  You could use this:

/usr/bin/mdb -k unix.0 vmcore.0

Or this simpler version:

/usr/bin/mdb 0

(which opens up the .0 pair).

Once that's done, '::status' and '$C' commands should
tell you something.

Rob T
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to