[Date Prev] [Date Index] [Date Next] [Thread Prev] [Thread Index] [Thread Next]

Lingering port binding after killing conserver on Sol8

Trevor Fiatal trevor@seven.com
Wed, 27 Jun 2001 18:47:17 -0700 (PDT)


I'm running conserver on Solaris 8 (SPARC), patchlevel MU4.
I compiled conserver as a 32-bit binary using gcc 2.95.3.

Overall, conserver is great.  It fscking rocks.  However,
it does have one major annoying behaviour in my environment:
when you kill the conserver daemon, either via in-band kill
from a 'console' client or via the 'kill' command, it does
NOT correctly relinquish the TCP port it had been listening
on.  In my case, it's tcp/782.

This is really ugly because the port hangs around in TIME_WAIT
for a long time (>5 minutes on my systems) and I can't restart
the conserver daemon until the timeout expires.  Clearly, this
makes restarting conserver something you don't want to do
unless you *really* have to.  It probably goes without saying,
but being able to 'kill -HUP' the daemon to cause a reread of
the config file would be a Very Good Thing.

Has anyone else seen this behaviour, or is it just me?


Trevor Fiatal -- trevor@seven.com -- http://www.seven.com/