[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: no buffer space available




Op 27-mrt-2007, om 12:05 heeft Timur Irmatov het volgende geschreven:

On 3/20/07, Maarten Vink <vink@interstroom.nl> wrote:
What's the output of /proc/slabinfo in a "normal" and "error" situation? That would give you some indication of the amount of memory allocated for,
among other things, the ARP cache. Also, what kernel version are you
running?

I have attached three files, slabinfo-bad.txt - it is contents of
/proc/slabinfo when machine starts experiencing this errors again,
slabinfo-fresh.txt - slabinfo just after reboot, slabinfo-other.txt -
slabinfo on other server, that is running for three days..

According to these files, the arp cache contains 300 entries after the reboot, and over 2000 when your problems start. I'd recommend increasing the values in proc/sys//net/ipv4/neigh/default/gc-thresh even further; if that doesn't solve your problem try flushing the arp table when the problem starts, and see if that fixes things without rebooting.

Maarten



Reply to: