Thanks go out to Dan Harrington and Udo de Boer for their aid.
After talking to our developers the action taken is to be decided upon. I
have made the following suggestions as I examined some of the suggestions
given. Although I can't yet give the results of each...
1.) Modify the socket listen queue limits - opening up more sockets
2.) Enable keepalive functionality - this cleans up sockets that do not exit
cleanly when the keepalive interval expires.
*3.) Decrease the TCP Connection Context Timeout limit - shortens the life
of a wait state for the tcp connection
**4.) Dis-able the PMTU discovery - the PMTU(path maximum transmission unit)
if a system has many remote clients this is suggested as the system creates
routing information for each client network.
* needs to be tested...
**not really related to the problem, but could improve performance on the
system.
Cheers,
Todd M. McGuinness
Systems Engineer
____________________________
NAGRAVISION SA
Kudelski Group
____________________________
Tel. : +41 21 732 03 11
Direct: +41 21 732 0623
Fax : +41 21 732 03 00
E-mail : mcguinness_at_nagra.com
Received on Tue Jan 22 2002 - 14:52:24 NZDT