OC4j container load balancing is not happening in Clustered OC4J and one node is getting loaded with active requests
Issue :- Mod_oc4j not able to connect to OC4J .
Below error in apache logs(error logs)
#########
Sun Feb 7 10:55:58 2010] [error] [client 100.1.2.101] [ecid: 1265525758:172.1.2.3:18412:0:1429,0] File does not exist: /
$OH/Apache/Apache/htdocs/error/HTTP_BAD_GATEWAY.html.var
[Sun Feb 7 10:56:00 2010] [warn] [client 100.1.2.101] oc4j_socket_recvfull timed out
[Sun Feb 7 10:56:00 2010] [error] [client 100.1.2.101] [ecid: 1265525456:172.1.2.3:25882:0:294,0] mod_oc4j: request to OC
4J ab.testworld.com :12543 failed: recv failed (errno=4)
[Sun Feb 7 10:56:00 2010] [warn] [client 100.1.2.102] oc4j_socket_recvfull timed out
[Sun Feb 7 10:56:00 2010] [error] [client 100.1.2.102] [ecid: 1265525458:172.1.2.3:15032:0:1931,0] mod_oc4j: request to O
C4J ab.testworld.com:12506 failed: Connect failed
#########
Solution in metalink :-
Troubleshooting MOD_OC4J_0080 MOD_OC4J_0058 MOD_OC4J_0035 MOD_OC4J_0121 MOD_OC4J_0013 Errors In the HTTP Server error_log file [ID 329456.1]
The MOD_OC4J* errors basically convey that the HTTP Server (through mod_oc4j) is unable to communicate with the OC4J instance in question. Some of the possible causes are :
1. Internet Explorer (KeepAlive) Bug
2. Firewalls
3. Connection timeouts between HTTP Server/mod_oc4j & OC4J
4. OC4J instance running short on memory
5. OC4J JVMs loaded
6. Application(s) blocking up OC4J
Connection timeouts between HTTP Server/mod_oc4j & OC4J
In $ORACLE_HOME/Apache/Apache/conf/httpd.conf, increase the Timeout directive value to, say, 600 seconds. (This should be smaller than the Firewall timeout setting, if any)
In $ORACLE_HOME/Apache/Apache/conf/mod_oc4j.conf below thetag add the following:
Oc4jCacheSize 0
Oc4jConnTimeout 600 (This is again in seconds &, again, should be smaller than the Firewall timeout setting, if any)
Update the configuration as follows:
$ORACLE_HOME/dcm/bin/dcmctl updateconfig -ct ohs
$ORACLE_HOME/opmn/bin/opmnctl restartproc process-type=HTTP_Server
Below error in apache logs(error logs)
#########
Sun Feb 7 10:55:58 2010] [error] [client 100.1.2.101] [ecid: 1265525758:172.1.2.3:18412:0:1429,0] File does not exist: /
$OH/Apache/Apache/htdocs/error/HTTP_BAD_GATEWAY.html.var
[Sun Feb 7 10:56:00 2010] [warn] [client 100.1.2.101] oc4j_socket_recvfull timed out
[Sun Feb 7 10:56:00 2010] [error] [client 100.1.2.101] [ecid: 1265525456:172.1.2.3:25882:0:294,0] mod_oc4j: request to OC
4J ab.testworld.com :12543 failed: recv failed (errno=4)
[Sun Feb 7 10:56:00 2010] [warn] [client 100.1.2.102] oc4j_socket_recvfull timed out
[Sun Feb 7 10:56:00 2010] [error] [client 100.1.2.102] [ecid: 1265525458:172.1.2.3:15032:0:1931,0] mod_oc4j: request to O
C4J ab.testworld.com:12506 failed: Connect failed
#########
Solution in metalink :-
Troubleshooting MOD_OC4J_0080 MOD_OC4J_0058 MOD_OC4J_0035 MOD_OC4J_0121 MOD_OC4J_0013 Errors In the HTTP Server error_log file [ID 329456.1]
The MOD_OC4J* errors basically convey that the HTTP Server (through mod_oc4j) is unable to communicate with the OC4J instance in question. Some of the possible causes are :
1. Internet Explorer (KeepAlive) Bug
2. Firewalls
3. Connection timeouts between HTTP Server/mod_oc4j & OC4J
4. OC4J instance running short on memory
5. OC4J JVMs loaded
6. Application(s) blocking up OC4J
Connection timeouts between HTTP Server/mod_oc4j & OC4J
In $ORACLE_HOME/Apache/Apache/conf/httpd.conf, increase the Timeout directive value to, say, 600 seconds. (This should be smaller than the Firewall timeout setting, if any)
In $ORACLE_HOME/Apache/Apache/conf/mod_oc4j.conf below the
Oc4jCacheSize 0
Oc4jConnTimeout 600 (This is again in seconds &, again, should be smaller than the Firewall timeout setting, if any)
Update the configuration as follows:
$ORACLE_HOME/dcm/bin/dcmctl updateconfig -ct ohs
$ORACLE_HOME/opmn/bin/opmnctl restartproc process-type=HTTP_Server
Comments