From nobody  Tue May 19 12:46:23 1998
Received: (from nobody@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id MAA00582;
          Tue, 19 May 1998 12:46:23 -0700 (PDT)
          (envelope-from nobody)
Message-Id: <199805191946.MAA00582@hub.freebsd.org>
Date: Tue, 19 May 1998 12:46:23 -0700 (PDT)
From: coutinho@dextra.com.br
To: freebsd-gnats-submit@freebsd.org
Subject: Network hangs, with "No buffers available" message
X-Send-Pr-Version: www-1.0

>Number:         6694
>Category:       kern
>Synopsis:       Network hangs, with "No buffers available" message
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Tue May 19 12:50:01 PDT 1998
>Closed-Date:    Fri Mar 19 03:42:44 PST 1999
>Last-Modified:  Fri Mar 19 03:43:29 PST 1999
>Originator:     Bill Coutinho
>Release:        2.2.6-RELEASE
>Organization:
Dextra Informatica
>Environment:
FreeBSD server.domain.com.br 2.2.6-RELEASE FreeBSD 2.2.6-RELEASE #0: Tue Apr  7 11:36:27 EST 1998
root@server.domain.com.br:/usr/src/sys/compile/SERVER  i386

>Description:
We are using a FreeBSD box to run socks5 server. This machine has one 
3Com Etherlink III (3c5x9) card.

When someone using socks-enabled client in a client machine tries to 
FTP something, the server's net adapter hangs. The main symptom is:

# ping 192.168.0.1
ping: sendto: No buffer space available
ping: sendto: No buffer space available
^C
# netstat -m
246 mbufs in use:
        109 mbufs allocated to data
        128 mbufs allocated to packet headers
        8 mbufs allocated to protocol control blocks
        1 mbufs allocated to socket names and addresses
74/266 mbuf clusters in use
562 Kbytes allocated to network (31% in use)
0 requests for memory denied
0 requests for memory delayed
0 calls to protocol drain routines

It seems that the only way to recover this problem is re-booting the 
machine.

>How-To-Repeat:
FTP'ing something with a FTP client under sockscap (win95). It does not
happen every time, but most of the times.

>Fix:

>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->feedback 
State-Changed-By: phk 
State-Changed-When: Tue May 19 22:31:55 PDT 1998 
State-Changed-Why:  
Please give us your kernel config file, and the output of netstat -an 
netstat -s, vmstat -m, netstat -in and ifconfig 

From: dag-erli@ifi.uio.no (Dag-Erling Coidan =?iso-8859-1?Q?Sm=F8rgrav?= )
To: coutinho@dextra.com.br
Cc: freebsd-gnats-submit@FreeBSD.ORG
Subject: Re: kern/6694: Network hangs, with "No buffers available" message
Date: 24 May 1998 17:39:14 +0200

 coutinho@dextra.com.br writes:
 > # ping 192.168.0.1
 > ping: sendto: No buffer space available
 > ping: sendto: No buffer space available
 > ^C
 > # netstat -m
 > 246 mbufs in use:
 >         109 mbufs allocated to data
 >         128 mbufs allocated to packet headers
 >         8 mbufs allocated to protocol control blocks
 >         1 mbufs allocated to socket names and addresses
 > 74/266 mbuf clusters in use
 > 562 Kbytes allocated to network (31% in use)
 > 0 requests for memory denied
 > 0 requests for memory delayed
 > 0 calls to protocol drain routines
 > 
 > It seems that the only way to recover this problem is re-booting the 
 > machine.
 
 This is a known problem with the ep driver. The immediate workaround
 is to ifconfig the interface down and up again. A more permanent
 workaround is to set NMBCLUSTERS to some large value, such as 4096 or
 8192, in your kernel configuration file. The Correct Solution (tm)
 would of course be to identify and correct the error in the ep driver.
 
 -- 
 Noone else has a .sig like this one.
State-Changed-From-To: feedback->closed 
State-Changed-By: sheldonh 
State-Changed-When: Fri Mar 19 03:42:44 PST 1999 
State-Changed-Why:  
Fixed in 2.2.6-RELEASE. 
>Unformatted:
