Re: pluto using 800M

"D. Hugh Redelmeier" <[email protected]> Mon, 1 Dec 2003 11:55:19 -0500 (EST)
Newsgroups gmane.network.freeswan.user,gmane.network.freeswan.devel
Message-ID <[email protected]>
-----BEGIN PGP SIGNED MESSAGE-----


| From: Paul Wouters <[email protected]>
| 
| I just had to restart the ipsec subsystem on www.freeswan.org. It is
| running OE. Pluto was using about 800M of memory. I've made a barf
| before restarting it: www.xtdnet.nl/paul/pluto-800M.barf
| Though I don't believe it contains any useful information on the
| memory leak.
| 
| It's version Linux FreeS/WAN 2.03-pre0. I will upgrade it to 2.04.

I admit that I haven't looked at the barf.  It agree that it probably
won't help.

How long was pluto running?  [This might give an idea of how fast the
leak is.]

Can you duplicate the problem easily?  [This makes diagnosis easier.]

If you are willing to recompile pluto, there is a compile-time macro
to select malloc debugging.  See "LEAK_DETECTIVE" in
freeswan/programs/pluto/Makefile.

If you can repeat the experiment with 2.02, it might be help to
determine if the problem was introduced between 2.02 and 2.03.

Thanks for the report.

Hugh Redelmeier
[email protected]  voice: +1 416 482-8253

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3ia
Charset: noconv

iQCVAwUBP8tygsFAuQPManGZAQE9kgQArCs86zvF+CNmHkVSsVDn1SaNPYk6Yshy
2zLX76MDH7Qq1QROl2JtQO7dX1eX+QOY8XEs7MvSnAvadDQzbGGsSAZ1sNsDy+ky
t69QAzIyFGxS5UFof3sJ/lVwPdRdjgqZYtCZZtR8l0DC3ijCVeygXdY9iYZpnSIn
MKsNE8rUyLA=
=jtqH
-----END PGP SIGNATURE-----