From nobody  Fri Mar 13 02:14:26 1998
Received: (from nobody@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id CAA01572;
          Fri, 13 Mar 1998 02:14:26 -0800 (PST)
          (envelope-from nobody)
Message-Id: <199803131014.CAA01572@hub.freebsd.org>
Date: Fri, 13 Mar 1998 02:14:26 -0800 (PST)
From: daniel@consol.de
To: freebsd-gnats-submit@freebsd.org
Subject: System hangs with NFS send error 55 (out of buffer space?)
X-Send-Pr-Version: www-1.0

>Number:         5995
>Category:       kern
>Synopsis:       System hangs with NFS send error 55 (out of buffer space?)
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Mar 13 02:20:02 PST 1998
>Closed-Date:    Tue Apr 14 12:00:59 PDT 1998
>Last-Modified:  Tue Apr 14 12:01:39 PDT 1998
>Originator:     Daniel Lang
>Release:        3.0-current SMP
>Organization:
ConSol
>Environment:
FreeBSD beatles 3.0-CURRENT FreeBSD 3.0-CURRENT #0: Fri Mar 13 10:59:53 GMT 1998     root@beatles:/usr/src/sys/compile/BERND  i386

current from 12.3.98
>Description:
Heavy NFS traffic causes the system to hang with that
error (NFS send error 55, seems to be out of buffer space)
on a SMP kernel. Using the non-SMP kernel doesn't show that
Problem. Could be related to the panic described earlier.

>How-To-Repeat:
ls -lR over NFS on a FreeBSD 3.0-current SMP Box


>Fix:
don't use SMP ?
>Release-Note:
>Audit-Trail:

From: daniel@consol.de
To: freebsd-gnats-submit@freebsd.org
Cc:  Subject: Re: kern/5995: System hangs with NFS send error 55 (out of buffer space?)
Date: Fri, 13 Mar 1998 17:18:59 +0100

 Funny thing happening in this situation:
 Even if the system doesn't hang completely, the network performance
 becomes extremely sluggish. 
 Then I decided to ping the machine, or to ping any other
 machine from it, and what happens is, that rtt's start
 with ~ 1000ms but decreases by 10 ms for each packet sent!!
 
 looks something like this:
 PING beatles: 64 byte packets
 64 bytes from 155.208.6.51: icmp_seq=0. time=1000. ms
 64 bytes from 155.208.6.51: icmp_seq=1. time=990. ms
 64 bytes from 155.208.6.51: icmp_seq=2. time=980. ms
 64 bytes from 155.208.6.51: icmp_seq=3. time=970. ms
 64 bytes from 155.208.6.51: icmp_seq=4. time=960. ms
 64 bytes from 155.208.6.51: icmp_seq=5. time=950. ms
 64 bytes from 155.208.6.51: icmp_seq=6. time=940. ms
 [..]
 64 bytes from 155.208.6.51: icmp_seq=97. time=30. ms
 64 bytes from 155.208.6.51: icmp_seq=98. time=20. ms
 64 bytes from 155.208.6.51: icmp_seq=99. time=10. ms
 64 bytes from 155.208.6.51: icmp_seq=100. time=0. ms
 64 bytes from 155.208.6.51: icmp_seq=101. time=1000. ms
 ..
 
 Note that ot starts again from 1000ms after reaching zero.
 The machines are in a local LAN.
 
 I've never seen anything like that before. 
 Comments ?
 
 Regards,
  Daniel
 -- 
 IRCnet: Mr-Spock                      - May His Shadow fall upon thee - 
 RL: Daniel Lang * dl@leo.org * +49 89 8540017 * http://www.leo.org/~dl/
State-Changed-From-To: open->closed 
State-Changed-By: phk 
State-Changed-When: Tue Apr 14 12:00:59 PDT 1998 
State-Changed-Why:  
This is some other problem, the ramping ping times indicate 
interrupt delivery problems. 
>Unformatted:
