1995-12-08 - Strange Unix behavior dealing with utmp - anyone recognize this?

Header Data

From: fc@all.net (Dr. Frederick B. Cohen)
To: cypherpunks@toad.com
Message Hash: 113bb6e90a54057a001736e30a3c01e378c21c970d5378e4eac454587682df7b
Message ID: <9512080330.AA27098@all.net>
Reply To: N/A
UTC Datetime: 1995-12-08 03:33:00 UTC
Raw Date: Thu, 7 Dec 95 19:33:00 PST

Raw message

From: fc@all.net (Dr. Frederick B. Cohen)
Date: Thu, 7 Dec 95 19:33:00 PST
To: cypherpunks@toad.com
Subject: Strange Unix behavior dealing with utmp - anyone recognize this?
Message-ID: <9512080330.AA27098@all.net>
MIME-Version: 1.0
Content-Type: text


I was sending out a few hundred 25K emails over a period of a minute or so
when my systemn encountered this thrilling sequence:

Dec  7 22:11:56 all vmunix: file:<3>zs1: silo overflow
Dec  7 22:11:57 all syslogd: /etc/utmp: File table overflow
Dec  7 22:11:56 all vmunix:  table is ful<3>zs1: silo overflow

It appears to be some sort of file table overflow related to the syslog
daemon.  The message repeats numerous times and I think it has to do
with the number of messages sent over a short time, however:

	I regularly send far more messages without this error
	It lost the processes that were generating these messages
		so I had to redo part of the mailing.

Does anyone recognize the symptom, and does this open a potential for
attack by stress?  (e.g., if you can get something to fail in this way
can the failure cause something to be missed?)

Just thought I'd mention it.

-> See: Info-Sec Heaven at URL http://all.net/
Management Analytics - 216-686-0090 - PO Box 1480, Hudson, OH 44236




Thread