Hi Solaris Experts,

I am wondering whether it is possible to clear a network socket port 17005 left by Apache Tomcat/6.0.20 after having to terminate it forcefully, since it has run away due to remote JDBC resource contention on another server. A normal Tomcat stop / shutdown proves ineffective from then on. This listening port below would hang around for quite some time before disappearing:

localhost.17005            *.*                0      0 49152      0 LISTEN

The suggestions I have may make it better, however you cannot get rid of some wait states and still have tcp configured correctly. I am assuming it goes to TIME_WAIT.

This is a programming error - which you cannot fix. Most times programs call setsockopt() with SO_REUSEADDR so that if the program is forced to exit then the socket can be reused right away.

What you can do is to check a TCP parameter, especially if you really are waiting one half hour. You should only have to wait 4 minutes or so.
As root:

ndd -get  /dev/tcp time_wait_interval

ndd -set /dev/tcp tcp_time_wait_interval 60000

kill <pid>

#!/bin/ksh
pids=$(/usr/bin/ps -ef | sed 1d | awk '{print $2}')

if [ $# -eq 0 ]; then
   read ans?"Enter port you would like to know pid for: "
else
   ans=$1
fi

for f in $pids
do
   /usr/proc/bin/pfiles $f 2>/dev/null | /usr/xpg4/bin/grep -q "port: $ans"
   if [ $? -eq 0 ]; then
      echo "Port: $ans is being used by PID:\c $f"      
   fi
done

Hi Jim mcnamara,

Thank you very much for your detail response and suggestion.

I tried the following command but had no luck with it:

root@braveheart # /usr/sbin/ndd -get /dev/tcp time_wait_interval
name is non-existent for this module
for a list of valid names, use name '?'

Is it supported on Solaris 10 (5.10 Generic_139556-08 i86pc i386 i86pc)?

Thanks so much again,

George