Subject: IPF state table handling
To: None <tech-net@netbsd.org>
From: Richard Braun <syn@sceen.net>
List: tech-net
Date: 06/28/2005 20:12:55
--2oS5YaxWCcQjTEyO
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Hello,

I'm having problems with the IPF state table. I have incremented
IPSTATE_SIZE and IPSTATE_MAX to 32749 and 22921 respectively on a router.
However, here is the output of ipfstat -s after some hours of activity :

IP states added:
        232489 TCP
        124194 UDP
        4372 ICMP
        75831299 hits
        555435 misses
        0 maximum
        0 no memory
        0 max bucket
        0 maximum
        0 no memory
        379 bkts in use
        12506 active
        0 expired
        0 closed
State logging enabled

State table bucket statistics:
        379 in use
        1.16% bucket usage
        0 minimal length
        1 maximal length
        1.000 average length

379/0.0116 =3D 32672.41, so I guess the new values are correctly used. What
I don't understand however is that there are only 379 buckets used, for
12506 entries in the table ?! I'm using a tricky NAT configuration which
could be the origin of a bug (if the values I report seems unusual for
other people too, it may be interesting to investigate this, since, on
other routers which have almost, but not the same configuration, I don't
have this problem).

If anyone has good links about details of the IPF state table implementatio=
n,
I would appreciate.

Thanks

--=20
Richard Braun

--2oS5YaxWCcQjTEyO
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (NetBSD)

iD8DBQFCwZMnBlWsEPLYRi8RAmvcAJ9+ur47PuObp1mqHqwUiiGmS8GJPwCgr5s2
YJ7ub1dzqP5xo06IurGXWlQ=
=ZqYL
-----END PGP SIGNATURE-----

--2oS5YaxWCcQjTEyO--