Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 17 Jun 2002 22:05:56 -0400
From:      Bosko Milekic <bmilekic@unixdaemons.com>
To:        "Scot W. Hetzel" <hetzels@westbend.net>
Cc:        FreeBSD-Stable <FreeBSD-Stable@FreeBSD.ORG>
Subject:   Re: System running out of mbufs
Message-ID:  <20020617220556.A97210@unixdaemons.com>
In-Reply-To: <011101c2165e$cbc5f400$11fd2fd8@ADMIN00>; from hetzels@westbend.net on Mon, Jun 17, 2002 at 07:26:32PM -0500
References:  <011101c2165e$cbc5f400$11fd2fd8@ADMIN00>

next in thread | previous in thread | raw e-mail | index | archive | help

 Can you try something other than the ET Inc. driver?  They may be
 right, but you should still try to make sure.

 --Bosko
 
On Mon, Jun 17, 2002 at 07:26:32PM -0500, Scot W. Hetzel wrote:
> We are currently running:
> 
> FreeBSD ns0.westbend.net 4.6-RC FreeBSD 4.6-RC #1: Wed Jun 12 11:39:27 CDT
> 2002     root@mail.westbend.net:/usr/obj/usr/src/src4/sys/GENERIC-ET  i386
> 
> The kernel configuration is:
> 
> # cvs diff -u GENERIC
> Index: GENERIC
> ===================================================================
> RCS file: /home/ncvs/src/sys/i386/conf/GENERIC,v
> retrieving revision 1.246.2.43
> diff -u -r1.246.2.43 GENERIC
> --- GENERIC     23 May 2002 17:04:01 -0000      1.246.2.43
> +++ GENERIC     17 Jun 2002 23:56:02 -0000
> @@ -56,6 +56,23 @@
>  options                ICMP_BANDLIM            #Rate limit bad replies
>  options        KBD_INSTALL_CDEV        # install a CDEV entry in /dev
> 
> +options         IPFIREWALL              #firewall
> +options         IPFIREWALL_VERBOSE      #enable logging to syslogd(8)
> +options         IPFIREWALL_FORWARD      #enable transparent proxy support
> +options         IPFIREWALL_VERBOSE_LIMIT=100    #limit verbosity
> +#options         IPFIREWALL_DEFAULT_TO_ACCEPT    #allow everything by
> default
> +options         IPV6FIREWALL            #firewall for IPv6
> +options         IPV6FIREWALL_VERBOSE
> +options         IPV6FIREWALL_VERBOSE_LIMIT=100
> +#options         IPV6FIREWALL_DEFAULT_TO_ACCEPT
> +
> +# RANDOM_IP_ID causes the ID field in IP packets to be randomized
> +# instead of incremented by 1 with each packet generated.  This
> +# option closes a minor information leak which allows remote
> +# observers to determine the rate of packet generation on the
> +# machine by watching the counter.
> +options                RANDOM_IP_ID
> +
>  # To make an SMP kernel, the next two are needed
>  #options       SMP                     # Symmetric MultiProcessor Kernel
>  #options       APIC_IO                 # Symmetric (APIC) I/O
> @@ -150,7 +167,7 @@
>  device         npx0    at nexus? port IO_NPX irq 13
> 
>  # Power management support (see LINT for more options)
> -device         apm0    at nexus? disable flags 0x20 # Advanced Power
> Management
> +#device                apm0    at nexus? disable flags 0x20 # Advanced
> Power Management
> 
>  # PCCARD (PCMCIA) support
>  device         card
> @@ -163,6 +180,11 @@
>  device         sio2    at isa? disable port IO_COM3 irq 5
>  device         sio3    at isa? disable port IO_COM4 irq 9
> 
> +options                BREAK_TO_DEBUGGER       # a BREAK on a comconsole
> goes to
> +                                       # DDB, if available.
> +options                CONSPEED=115200         # speed for serial console
> +                                       # (default 9600)
> +
>  # Parallel port
>  device         ppc0    at isa? irq 7
>  device         ppbus           # Parallel port bus (required)
> @@ -177,6 +199,10 @@
>  device         em              # Intel PRO/1000 adapter Gigabit Ethernet
> Card (``Wiseman'')
>  device         txp             # 3Com 3cR990 (``Typhoon'')
>  device         vx              # 3Com 3c590, 3c595 (``Vortex'')
> +
> +# ETinc
> +device         eth0
> +#device                bw0 at isa ?
> 
>  # PCI Ethernet NICs that use the common MII bus controller code.
>  # NOTE: Be sure to keep the 'device miibus' line in order to use these
> NICs!
> 
> The problem we are experiencling is that an unusual high number of mbufs are
> being used by the system, this problem started about 2 weeks ago with an
> April 25th 4.5-STABLE kernel, we have tried, upgrading to 4.6-RC, and
> downgrading to 4.4-RELEASE, without any success in solving this problem.
> 
> I had posted about this problem before (see "Run away mbufs"), which I
> received only one response too.  That response had pointed me to a utility
> called minfo (by Ian Dowse).  The minfo utility had shown quite a few "Free
> mbuf not on free list" and "refs missing" in it's output.
> 
> 
> We have been using a crontab script to reboot the server when ever the peak
> mbufs in use is > 75% (checked at 5 min intervals), currently the server
> seems to stay up for only 1Hr 45min, before we need to reboot it.
> 
> Mon Jun 17 16:45:00 CDT 2002
> 0% mbufs in use
> 1717/1936/240000 mbufs in use (current/peak/max):
>         1717 mbufs allocated to data
> 220/278/60000 mbuf clusters in use (current/peak/max)
> 1040 Kbytes allocated to network (0% of mb_map in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 0 calls to protocol drain routines
> 
> :
> :
> 
> Mon Jun 17 18:15:00 CDT 2002
> 62% mbufs in use
> 150769/150960/240000 mbufs in use (current/peak/max):
>         150762 mbufs allocated to data
>         7 mbufs allocated to packet headers
> 212/318/60000 mbuf clusters in use (current/peak/max)
> 38376 Kbytes allocated to network (21% of mb_map in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 0 calls to protocol drain routines
> 
> Mon Jun 17 18:20:00 CDT 2002
> 66% mbufs in use
> 160437/160496/240000 mbufs in use (current/peak/max):
>         160427 mbufs allocated to data
>         10 mbufs allocated to packet headers
> 223/318/60000 mbuf clusters in use (current/peak/max)
> 40760 Kbytes allocated to network (22% of mb_map in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 0 calls to protocol drain routines
> 
> Mon Jun 17 18:25:01 CDT 2002
> 72% mbufs in use
> 172748/172864/240000 mbufs in use (current/peak/max):
>         172739 mbufs allocated to data
>         9 mbufs allocated to packet headers
> 216/318/60000 mbuf clusters in use (current/peak/max)
> 43852 Kbytes allocated to network (1% of mb_map in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 0 calls to protocol drain routines
> 
> Mon Jun 17 18:30:00 CDT 2002
> 76% mbufs in use
> 182638/182848/240000 mbufs in use (current/peak/max):
>         182638 mbufs allocated to data
> 212/318/60000 mbuf clusters in use (current/peak/max)
> 46348 Kbytes allocated to network (2% of mb_map in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 0 calls to protocol drain routines
> 
> We have also sent messages to ET Inc, and they have advised that the problem
> is not with their driver, since we are seeing mbufs starvation, and not mbuf
> cluster starvation occuring.
> 
> We have tried disabling sendmail, amavisd, named, and haven't been able to
> figure out what could be causing the mbuf not to be freed properly.
> 
> Any ideals as to what else we should do to trouble shoot this problem?
> 
> Scot
> 
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-stable" in the body of the message
> 

-- 
Bosko Milekic
bmilekic@unixdaemons.com
bmilekic@FreeBSD.org


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020617220556.A97210>