Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 13 Oct 1998 23:32:58 -0400 (EDT)
From:      <ratbert@phoenix.aye.net>
To:        Pecsenyanszky Istvan <pisti@c3.hu>
Cc:        freebsd-isp@FreeBSD.ORG
Subject:   Re: kernel panic
Message-ID:  <Pine.BSF.3.96.981013225740.9809A-100000@phoenix.aye.net>
In-Reply-To: <Pine.SGI.4.05.9810132035350.19435-100000@visio.c3.hu>

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

Does your box have lots of processes, lots of network traffic, or
both? Increasing maxusers will increase mbuf clusters something like 
(512 + (maxusers * 16)) which is 8704 -- kind of high. Each mbuf
cluster takes 2k so with 8704 mbuf clusters you are using 16 meg
plus, and there is only 32 meg of kvm. Increasing maxusers
will cause other kernel data structures besides mbuf clusters to
grow (those need kvm too), so you may want to shrink maxusers
down and twiddle with "options  NMBCLUSTERS=xxxxx" in your
kernel config file.

If you have a lot of processes running (a thousand or so) you
may want to search the archives for the patches David Greenman
applied to wcarchive (requires a buildworld).


Watch the mbuf cluster usage with a "netstat -m" and see if
it gets around 8000. That is really kind of high. My squid
box get 14 new connections a second and it rarely gets above
1000.

Is your box really really busy?

On Tue, 13 Oct 1998, Pecsenyanszky Istvan wrote:

> Hi!
> 
> I'm running 2.2.6-RELEASE, and lately it crashes frequently with the
> following message:
> 
>  Oct 13 15:50:05 fmback /kernel: Out of mbuf clusters - increase maxusers!
>  
>  
>  Fatal trap 12: page fault while in kernel mode
>  fault virtual address   = 0x1c
>  fault code              = supervisor write, page not present
>  instruction pointer     = 0x8:0xf014fac8
>  stack pointer           = 0x10:0xefbffc6c
>  frame pointer           = 0x10:0xefbffdc8
>  code segment            = base 0x0, limit 0xfffff, type 0x1b
>                          = DPL 0, pres 1, def32 1, gran 1
>  processor eflags        = interrupt enabled, resume, IOPL = 0
>  current process         = 126 (nfsd)
>  interrupt mask          = net
>  panic: page fault
>  
>  syncing disks... 26 23 11 11 11 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 giving up
>  Automatic reboot in 15 seconds - press a key on the console to abort
>  Rebooting...
> 
> 
> I tried to increase maxusers, but now it is 512, and if I try to set it
> to a higher value, I get warning message at the kernel configuration.
> 
> What should I do?
> 
> Thank you,
> Istvan
> 
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-isp" in the body of the message
> 


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



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.981013225740.9809A-100000>