this happens very rare. it is mostly seen on systems with full bgp table
combined with a massive scan on diverse networks caused by worms/troyans/etc.
you can tuneup your systems' max route cache size by
echo sane-big-value > /proc/sys/net/ipv4/route/max_size
if you happen to get a permanent attack, you may also need to twiddle with
garbage collection values in /proc/sys/net/ipv4/route/gc_* to ensure that the
route cache will not grow more than the max value for the period of gc.
> ----- Original Message ----
> From: Andrew Morton <akpm@xxxxxxxx>
> To: netdev@xxxxxxxxxxx
> Sent: Sun, 20 Mar 2005 16:37:25 -0800
> Subject: Fw: dst cache overflow messages
> Begin forwarded message:
> Date: Sun, 20 Mar 2005 22:51:08 +0200 (EET)
> From: buakaw@xxxxxxxxxxxxxxxxxxxx
> To: linux-kernel@xxxxxxxxxxxxxxx
> the system become very laggy.
> i use kernel 2.6.10/126.96.36.199 on p4 2.8, msi915p, 3 x 3com905 boomerang.
> and the cpu usage normally be about 12% and after something happens it
> boost to 100% and it is used mostly by ksoftirqd/0, and a little bit by
> migration/0 and event/0. and in syslog i found thies lines
> Mar 20 22:21:09 buakaw kernel: printk: 5543 messages suppressed.
> Mar 20 22:21:09 buakaw kernel: dst cache overflow
> what can cause this?