Solaris SPARC Operating System
 
Sun Enterprise 10000 Server
 

GCS>Sun Microsystems>Servers>High-End Servers
 

There are occasions where it is useful to know exactly what was typed at a
netcon session, especially if there are multiple administrators and
suspected watchdog resets.
Netcon logging is integrated into SSP 3.1.1 and beyond. To enable logging
for SSP 3.1.1 and higher, edit the /etc/syslog.conf file and uncomment the
local1.debug line, then restart the SSP.
# Uncomment to enable SSP netcon session logging:
local1.debug           /var/opt/SUNWssp/.ssp_private/machine_server_fifo
The output for each domain's console is stored in $SSPVAR/adm/<domain>/netcon.
SSP 3.0 and 3.1 netcon logging use the following 3 programs and 1 crontab
entry:
/etc/rc2.d/S99netcon_logger
===========================
su ssp -c sh <<END_OF_LOGGER >/dev/null
. /etc/opt/SUNWssp/ssp_env.sh
$SSPOPT/bin/netcon_loggerd &
END_OF_LOGGER
/opt/SUNWssp/bin/netcon_loggerd
===============================
#!/bin/sh
while [ `ps -ef | grep netcon_server | grep -v grep | wc -l` = 0 ]
do
sleep 10
done
for domain in `domain_status | tail +2 | awk '{print $1}'`
do
SUNW_HOSTNAME=$domain
export SUNW_HOSTNAME
netcon < /dev/null 1>>$SSPLOGGER/$domain/netcon.log 2>&1 &
joblist="$joblist $!"
done
sleep 5
trap 'kill -9 $joblist ; exec $0' 18 1
wait
/opt/SUNWssp/bin/newnetconlog
=============================
#! /usr/bin/sh
#
# Copyright (c) 1998 by Sun Microsystems, Inc.
# All rights reserved.
#
###############################################################################
#
# This script archives netcon.log messages files.  Normally, it should be
# run once per week.  The number of logs to be kept is controlled by
# the shell variable called "MAX" below.
#
###############################################################################
MAX=8
fatal()
{
	echo "ERROR!  $*" >&2
	echo "ABORTING!" >&2
	exit 2
}
ENV=/etc/opt/SUNWssp/ssp_env.sh
[ -f ${ENV} ] || fatal "Couldn't find ${ENV}!"
. ${ENV}
DOM=/var/opt/SUNWssp/.ssp_private/domain_config
[ -f ${DOM} ] || fatal "Couldn't find ${DOM}!"
DOMAINS=`awk -F: '{print $1}' ${DOM}`
LOG=netcon.log
for domain in $DOMAINS
do
	LOGDIR=$SSPLOGGER/$domain
	if [ ! -d $LOGDIR ]
	then
		rm -f $LOGDIR
		mkdir -p $LOGDIR
	fi
	cd $LOGDIR
	i=$MAX
	while [ $i -gt 0 ]
	do
		j=`expr $i - 1`
		test -f ${LOG}.$j && mv -f ${LOG}.$j ${LOG}.$i
		test -f ${LOG}.$j.Z && mv -f ${LOG}.$j.Z ${LOG}.$i.Z
		i=$j
	done
	test -f $LOG && mv -f $LOG ${LOG}.$i
	( sleep 5 ; compress ${LOG}.$i ) &
	touch $LOG
	chown ssp $LOG
	chgrp staff $LOG
done
pid=`/usr/bin/ps -ef | /usr/bin/grep [n]etcon_loggerd | awk '{print $2}'`
if [ "$pid" != "" ]
then
	kill -HUP $pid
else
	sh /etc/rc2.d/S99netcon_logger start
fi
entry to be added to root's crontab
===================================
5 4 * * 6 /opt/SUNWssp/bin/newnetconlog >/dev/null 2>&1 # archives
netcon.log
		NOTES
		=====
S99netcon_logger starts netcon_loggerd at boot time.
netcon_loggerd records a log of netcon into a file called netcon.log in the
same directory as the messages files for each domain,
i.e. /var/opt/SUNWssp/adm/<domainname>/netcon.log
If the netcon process which is recording the log dies or is killed,
netcon_loggerd will restart it, unless an exclusive netcon session is used.
To restart recording after an exclusive session has been used, send a
SIGHUP signal to netcon_loggerd:
# kill -HUP <pid-of-netcon_loggerd>
If this is forgotten, then newnetconlog will send the SIGHUP each week anyway.
Finally newnetconlog archives the old log files and compresses them where
possible.

This solution has no attachment