Re: OOM after isdn connection kills 2.3.29

Mike Galbraith (mikeg@weiden.de)
Mon, 6 Dec 1999 06:42:56 +0100 (CET)


On Sun, 5 Dec 1999, Oliver Neukum wrote:

> Hi,
>
> i have a reproducable method to lock up 2.3.29.
>
> After an isdn connection producing an OOM will cause extreme swapping and then
> lock up the system. Even alt-sysreq does not work.
> Should I try to run ikd ?

Maybe. As I stated earlier though, the trick is going to be stopping
the leak before you are completely dead so that you can get the data
from memleak out of /proc and save it. If you have a serial console
or lp console and can still cat a file at OOM time, you're golden.
(hmm.. maybe I should add a SysRq-dump-memleak-data button)

If it's userland exausting memory, memleak won't help at all. Set
resource limits on userland processes first, and watch ps output like
a hawk for root processes. If OOM happens without anyone in userland
being visibly guilty, then its time to look for a possible kernel leak.

-Mike

-
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/