are you using postgresql on it ? or any other application that uses 
semaphores ?

On Monday 05 April 2004 20:29, Wojciech Puchar wrote:
> i have this problem for a long time using current - still exist with
> 1 month old current
>
> it's stable if there is a lot of memory and swapping is rare.
>
> in system where swapping is normal (Pentium 133, 256MB RAM, 200MB
> swap at normal load) - but still not heavy, system crashes after 2-3
> days average.
>
>
> it happens (mostly observed) when squid -k rotate is started and
> 150MB squid process travels it's all VM space to write cache
> directory to /var/spool/squid/swap.state.
>
>
> it's not complete crash, i still can ping machine, routing/NAT still
> works, but anything else.
>
> connecting to any port succeed but then it waits forever, or
> disconnect immediately (depends).
>
> locally - all that can be done is console changing with CTRL-ALT-F*.
>
> nothing else.
>
>
> any idea how to fix it?