This sounds very similar to the mbuf leaks back in pre 1.2
	(What version are you running?)
	If its 1.2 or -current it looks like there might be another leak.
	Can you check with netstat as to the amount of data queued in/out
	when the mbuf usage is high?

		David/abs

  -- "Its hard to say it... I hate to say it... but its probably me..." --

On 17 Aug 1997 explorer@flame.org wrote:

> 
> I am getting this message (and the lockup associated) on my home machine.
> I really think I am getting it too often...  I am not really putting
> the machine through its paces enough to make me think it should be
> using 256 pages for network mbufs:
> 
> 317 mbufs in use:
>         315 mbufs allocated to data
>         2 mbufs allocated to socket names and addresses
> 256/256 mapped pages in use
> 551 Kbytes allocated to network (100% in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 6403 calls to protocol drain routines
> 
> This is on the following setup:
> 
> Lockup #1:
> 	Running X, with one ssh session to a remote host, with emacs
> 	displayed from that machine (X forwarding).  Also, NFS to another
> 	local machine.  That machine is also the ppp machine for my home
> 	systems.
> 
> Lockup #2:
> 	Running X, with no ssh sessions, but just started Netscape.
> 
> In the first case, I don't know how long the machine was up.  In the second,
> it was up for no more than 8 hours, and without much load at all.
> 
> Here is the output of netstat -m from the system right now, with an
> emacs-over-ssh connection and a netscape running:
> 
> 37 mbufs in use:
>         35 mbufs allocated to data
>         2 mbufs allocated to socket names and addresses
> 32/56 mapped pages in use
> 116 Kbytes allocated to network (58% in use)
> 0 requests for memory denied
> 0 requests for memory delayed
> 0 calls to protocol drain routines
> 
> Has anyone else seen anything like this?
> 
> --Michael
>