[Openswan Users] Gradual fail of openswan, memory goes 300MB+

[Paul Wouters]

On Sat, 11 Nov 2006, ted leslie wrote:

> I am running a openswan , with about 1000 x.509 certs in it,
> it runs fine for about a week, 1000's of connections are brought up and
> down on it.
>
> The memory usage grows about 300-400 MB  (on a 756 MB machine),

Not good. Please recompile pluto with -DLEAK_DETECTIVE enabled in
programs/pluto/Makefile. Let it run for a while, until you got a
sieable memory increase, and then "nicely" restart openswan using
the initscript. At shutdown, pluto will log a lot of memory debugging
info that should help us trace this memory leak.

One memory leak I know about is if you have a phase 1 up, but phase2
keeps retrying with some failure.

Paul
-- 
Building and integrating Virtual Private Networks with Openswan:
http://www.amazon.com/gp/product/1904811256/104-3099591-2946327?n=283155