From: Dragos Stoichitescu (Dragos.Stoichitescu@connex.ro)
Date: Thu Aug 11 2005 - 08:46:54 EDT
Hi,
I've got some Java Technology-Based Server Applications running on
solaris8 (SunFire 880). Both applications are running on the same server
(which is part of a three nodes Sun Cluster - version 3.0) .
>From time to time (but starting from a few days more and more often)
when an application tryes to get data from the other it exists ; in the
application logs it reports TCP errors;
I don't have dropped packets, interface errors, collisions ... ; also
the number of used sockets is below 400. The only thing that grows when
the errors begin is "tcpAttemptFails" (when I do a "netstat -sP tcp") .
I've tried to increase some TCP related parameters ("tcp_xmit_hiwat" ,
"tcp_recv_hiwat" ) and I set "tcp_wscale_always" to 1 but
"tcpAttemptFails" keeps growing .
Does any of you know what TCP parameter(s), if changed, will make
""tcpAttemptFails" go down and hopefuly keep the applications running ??
Thanks in advance
regards
Dragos Stoichitescu
UNIX System Administrator
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:31:17 EDT