SUMMARY: Overflowing disks hang machine

From: Colin McNeill (colin@gtl.com)
Date: Mon Sep 16 1996 - 04:41:44 CDT


The original message was this:

----- Begin Included Message -----

>From sun-managers-relay@ra.mcs.anl.gov Fri Sep 13 11:03:17 1996
Sender: sun-managers-relay@ra.mcs.anl.gov
Date: Thu, 12 Sep 96 13:23:07 BST
From: Colin McNeill <colin@gtl.com>
Reply-To: Colin McNeill <colin@gtl.com>
Followup-To: Colin McNeill <colin@gtl.com>
To: sun-managers@eecs.nwu.edu
Subject: Overflowing disks hang machine
Cc: colin@gtl.com
Precedence: bulk
Content-Length: 831
X-Lines: 24

Dear All

I have an NFS server here running SunOS 4.1.3_U1. It is actually an Axil 311 (Sparcstation 10 clone) but I don't think that is at fault.

What happens is this:

When an NFS mounted filesystem is filled by a client multiple (anything up to a 1000) file system full errors are generated. This grinds the server to a halt. If the client process which is filling the disk is found and stopped then the machine might recover in about 5 minutes. However, sometimes the machine panics with a:

segvn soft_unlock problem.

Yesterday I applied the NFS Jumbo Patch 102177-04 hoping that this would sort the problem, but although it hasn't paniced again it has ground to a halt.

HELP

Unix systems should be able to handle such abuse.

Any suggestions gratefully received

Thanks in Advance

Colin McNeill

Colin.McNeill@gtl.com

----- End Included Message -----

The replies I had basically said upgrade to SunOS 4.1.4 or Solaris 2 and the problem will go away. I'll try this as soon as practicable.

Thanks to

rloftin@engsys.mc.xerox.com (Ron Loftin)

Anthony Goonetilleke <agoon@arbld.unimelb.edu.au>

for the pointers.

Colin McNeill



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:10 CDT