Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 11 Jun 2003 17:12:41 -0400
From:      "MFW" <mwilliams@2goons.net>
To:        =?ISO-8859-1?Q?Jo=E3o?= Assad <jfassad@parperfeito.com.br>, <freebsd-isp@freebsd.org>
Subject:   Re: ipf/ipnat no memory problem
Message-ID:  <hgc5l5.n4xcf4@admin.2goons.net>
In-Reply-To: <020201c3304a$8fc5bd80$0402a8c0@joao>
References:  <020201c3304a$8fc5bd80$0402a8c0@joao>

next in thread | previous in thread | raw e-mail | index | archive | help
I will take a quick stab at it.

First off, I would add this to your kernel:

options         NMBCLUSTERS=3D16384

That will give you more memory for network traffic. Looks like this server =
is
reaching its limits. you might want to give that a shot now and see if you =
are
running into the same problem tonight. I would also write a script that run=
s
at peak time to capture 'netstat -m'. Also, if this box is running any gig
interfaces, I would crank the number above up to 32768 (Maximum value for
NMBCLUSTERS).

Just my 2 cents.

Matt

Jo=E3o Assad <jfassad@parperfeito.com.br> wrote:

> Hello guys,
>=20
> Does anybody have a clue on how to solve this problem ?
>=20
> firewall# ipfstat -s
> IP states added:
>         8950710 TCP
>         24299 UDP
>         4134 ICMP
>         1592473870 hits
>         3165269525 misses
>         6 maximum
>         650 no memory
>         9215 bkts in use
>         11005 active
>         29606 expired
>         8939070 closed
>=20
> firewall# ipnat -s
> mapped  in      913470782       out     1028719022
> added   59149802        expired 59056159
> no memory       129676  bad nat 0
> inuse   93643
> rules   38
> wilds   0
> firewall#
>=20
> I am getting "no memory" in both ipf and ipnat.
>=20
> CPU: Pentium III/Pentium III Xeon/Celeron (802.72-MHz 686-class CPU)
> real memory  =3D 134217728 (131072K bytes)
> avail memory =3D 127221760 (124240K bytes)
>=20
> ---------Relevant configurations----------
> In /usr/src/sys/contrib/ipfilter/netinet/ip_state.h :
> # define        IPSTATE_SIZE    30011
> # define        IPSTATE_MAX     21011   /* Maximum number of states held =
*/
>=20
> Kernel options:
> maxusers        0
> options         IPFILTER
> options         IPFILTER_LOG
> options         IPFILTER_DEFAULT_BLOCK
> options         IPSTEALTH
> options         VM_KMEM_SIZE_SCALE=3D"2"
>=20
> I dont have the netstat -m output of my peak time which is when the probl=
em
> occurs, but right now its:
>=20
> firewall# netstat -m
> 269/912/6016 mbufs in use (current/peak/max):
>         269 mbufs allocated to data
> 265/594/1504 mbuf clusters in use (current/peak/max)
> 1416 Kbytes allocated to network (31% of mb_map in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 0 calls to protocol drain routines
>=20
>=20
> I would appreciate it if someone can give me some help in this issue, Im
> completely in the dark right now.
>=20
> Best regards,
>=20
> --
> Jo=E3o Assad
> ParPerfeito Comunica=E7=E3o LTDA
> http://www.parperfeito.com.br/
>=20
>=20
>=20
> _______________________________________________
> freebsd-isp@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-isp
> To unsubscribe, send any mail to "freebsd-isp-unsubscribe@freebsd.org"
>=20

_________________________________________________________
This mail sent using V-webmail - http://www.v-webmail.org



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?hgc5l5.n4xcf4>