Re: Generating kernel crash dumps

• Subject: Re: Generating kernel crash dumps
• From: Jose Commins <email@hidden>
• Date: Wed, 2 Jun 2004 13:44:39 +0100

Hi everyone,

I'd like to know whether it's possible to generate kernel crash dumps on
panic() that can later be analysed.
What I mean is a process similar to that on FreeBSD or Solaris, where
you set the dump device plus a kernel variable to tell the kernel to
dump on panic and have the system call the savecore utility when it
comes back up.
You then have a vmcore file in the directory of your choice (given as a
parameter to savecore) which you can then debug with gdb -k.

For my purposes, the recommended two-machine debugging utility described
in the Kernel Programming document is very cumbersome. I've read
Technical Note TN 2063 "Understanding and debugging kernel panics" which
is more along the lines of what I want, but is there any way of actually
creating and then analysing a vmcore?

Thanks very much for your help

Cheers,
Gerald

• Follow-Ups:
  • Re: Generating kernel crash dumps
    • From: Gerald Heinig <email@hidden>

• Prev by Date: Re: Generating kernel crash dumps
• Next by Date: Re: Generating kernel crash dumps
• Previous by thread: Re: Generating kernel crash dumps
• Next by thread: Re: Generating kernel crash dumps
• Index(es):
  • Date
  • Thread