On Thu, 20 Jan 2000, David desJardins wrote:
> Willy Tarreau <wtarreau@yahoo.fr> writes:
> > David, you can use the KMSGDUMP tool I developped for this a while
> > ago. Compile your kernel with it compiled in and you can automatically
> > dump all your messages on a diskette previously inserted in the drive
> > on a kernel panic. You can also get into a simple text interface in
> > which you can scroll your messages, and decide to dump them or print
> > them via the parallel port.
>
> These machines don't have floppy drives or keyboards either. Only a few
> have video cards (which is one reason we only have a few dumps). If we
> can't get it solved another way, perhaps we can set up machines with
> more hardware from which we can get more crash data. For now, it's not
> clear that collecting more dumps that look the same will help, though?
Since the shit seems to hit the fan above the ethernet layer... What about
putting the printk output into packets (e.g. one per line) and sending
them to logging host? Disclaimer: I'm a VFS hacker, not a networking one,
so it may be a total BS. Looks like a tweaked copy of dev_queue_xmit()
might be a decent starting point... Alan, your comments? Related question:
how realistic it would be (in case of panic) to shove the dump into 802.2
packet and push it ahead of everything else?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Jan 23 2000 - 21:00:25 EST