[Planetlab-devel] dst cache overflow

[Sapan Bhatia]

Hi Stephen,

Yes, this means that the machine is out of dst cache entries, which are
required to route packets.
If the no. of spare entries drops to zero, then the machine loses
connectivity.

One way to address the problem, which will work as long as the kernel is not
leaking dst structs
is running the following command:

echo 262144 > /proc/sys/net/ipv4/route/max_size

Sapan

On 7/17/07, Stephen Soltesz <soltesz at cs.princeton.edu > wrote:
>
> Has anyone seen errors like these before on the PL kernel?
>
> planetlab1.rbg.informatik.tu-darmstadt.de recently came backup, and the
> contact there reported that these messages were on the console during boot
>
> up.  The node has successfully booted. But, based on the logs this has
> occurred for *a while.*  Monitor identified it as down, so I'm concerned
> that these messages are related to making the machine inaccessible.
>
> Thank you,
> Stephen.
>
> Jul 17 16:06:06 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:11 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:16 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:21 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:25 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:31 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:36 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:41 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:46 planetlab1 kernel: dst cache overflow
> Jul 17 16:06:51 planetlab1 kernel: dst cache overflow
>
> _______________________________________________
> Devel mailing list
> Devel at lists.planet-lab.org
> https://lists.planet-lab.org/mailman/listinfo/devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.planet-lab.org/pipermail/devel/attachments/20070717/23836ab7/attachment.html