IBM-AIX users continuously hitting a web application that has many users (for example, greater than 400) through Oracle HTTP Server might experience a large error log being generated. This is most likely a tuning issue that can be corrected by updating certain IBM-AIX system parameters and Oracle HTTP Server tuning parameters, as described as shown in the following examples.
For AIX system parameters:
In /etc/security/limits, should have the following parameters:
nofiles = -1 nofiles_hard = -1
In /etc/rc.net, should have the following parameters:
/usr/sbin/no -o sb_max=6192000 /usr/sbin/no -o tcp_sendspace=4096000 /usr/sbin/no -o tcp_recvspace=4096000 /usr/sbin/no -o udp_sendspace=65536 /usr/sbin/no -o udp_recvspace=655360 /usr/sbin/no -o rfc1323=1 /usr/sbin/no -o ipqmaxlen=150 /usr/sbin/no -o clean_partial_conns=true
You will need to reboot your machine.
For Oracle HTTP Server Configuration:
In mod_wl_ohs.conf, should look like the following: example:
LoadModule weblogic_module "${PRODUCT_HOME}/modules/mod_wl_ohs.so"
# This empty block is needed to save mod_wl related configuration from EM to this file when changes are made at the Base Virtual Host Level
<IfModule weblogic_module>
# WebLogicHost <WEBLOGIC_HOST>
# WebLogicPort <WEBLOGIC_PORT>
# MatchExpression *.jsp
WebLogicCluster <host-name>:<port>,<host-name>:<port>,<host-name>:<port>
ConnectTimeoutSecs 99999 //[Optional parameter]
WLIOTimeoutSecs 99999 //[default value 300]
WLSocketTimeoutSecs 99999 //[default value 2]
MatchExpression *
<Location /diagservlet>
# SetHandler weblogic-handler
WLSRequest On
WebLogicCluster
<host-name>:<port>,<host-name>:<port>,<host-name>:<port>
# PathTrim /weblogic
# ErrorPage http:/WEBLOGIC_HOME:WEBLOGIC_PORT/
</Location>
</IfModule>
In httpd.conf, set the following:
<IfModule mpm_worker_module>
-
-
-
MinSpareThreads 200 [default value 25]
MaxSpareThreads 800 [default value 75]
-
-
-
-
</IfModule>