01-11-2001 01:46 PM
On the console "connection reset by peer" meesages appear. I was wondering what could be causing this to happen. Any help is appreiciated.
01-12-2001 07:06 AM
#netfmt -f /var/adm/nettl.LOG00 > /tmp/file
Could be caused by PC users not exiting applications gracefully, instead just "x-ing" them.
01-15-2001 02:02 PM
connection reset by peer
is a catch all message. All that this means is that there was a TCP connection to another system that received a RST packet. The cause for this reset is as numerous as the types of applications you run on your system. The fact that this appeared on the console is of interest. First, check to make sure that the system logging deamon (syslogd) is properly running and you are getting messages posted to syslog. The message should log in /var/adm/syslog/syslog.log.
The only reason that this message would log to either the console or the syslog is that the process that controls the socket connection is either a system process (inetd) or the controlling process has previously died.
Is this something to be concerned about? It depends on whether you are getting work done or if this reset is causing work to stop. You can tell this by complaints or by applications failing. No failing applications or complaints... No issue.
I've not seen a duplicate IP address cause this error, but there is always a first time.
The PC application could cause this, but it would have to be one that cared about the network stack. Most don't and don't issue a close on the open socket. In my experience they wait for the keepalive process to clean up after them.