Hi All,
Excuse me for my language/semantic problems.
We have a frequent problem which we have been encountering on one of our glassfish cluster (currently two nodes) from quite sometime now.
Here is the description of the problem.
We have a system which on an average processes more 8 concurrent transaction at a given where the request number hitting web server(loadbalancers) -> app servers are quite high.
For past few days our one or both nodes in ending up in hung state and basically we are prompted with error message reached the below are few sample logs we can supply
[i]#|2009-11-12T13:18:46.146+0000|WARNING|sun-appserver9.1|javax.enterprise.system.container.web|_ThreadID=1400;_ThreadName=SelectorReaderThread-38080;_RequestID=0d2c82c2-7322-4917-b1a7-863b5091f13c;|Maximum Connections Reached: 8192 -- Retry later|#][/i]
and unfortunately oflate only way by we are solving this is by restarting the node/cluster
We can confirm that there was enough of buffer available yet the below was available.
the below are http-service configurations which were applied avilable on domain.xml
[code]<config dynamic-reconfiguration-enabled="true" name="cluster_xxx-config">
<http-service>
---------------------------------------------
---------------------------------------------
---------------------------------------------
<request-processing header-buffer-length-in-bytes="8192" initial-thread-count="16" request-timeout-in-seconds="30" thread-count="256" thread-increment="2"/>
<keep-alive max-connections="300" thread-count="4" timeout-in-seconds="30"/>
<connection-pool max-pending-count="8192" queue-size-in-bytes="8192" receive-buffer-size-in-bytes="8192" send-buffer-size-in-bytes="8192"/>
<http-protocol default-response-type="AttributeDeprecated" default-type="text/html; charset=iso-8859-1" dns-lookup-enabled="false" forced-response-type="AttributeDeprecated" forced-type="text/html; charset=iso-8859-1" ssl-enabled="true" version="HTTP/1.1"/>
<http-file-cache file-caching-enabled="true" file-transmission-enabled="false" globally-enabled="true" hash-init-size="0" max-age-in-seconds="30" max-files-count="1024" medium-file-size-limit-in-bytes="537600" medium-file-space-in-bytes="10485760" small-file-size-limit-in-bytes="2048" small-file-space-in-bytes="1048576"/>
<property name="accessLoggingEnabled" value="true"/>
</http-service>[/code]
I understand there are a lot backend work our application does where the response time sometime goes beyond 20 sec.
However, looking at the long term we wondering if someone suggest in terms of fine tuning these configuration it'd be great.
Thanks in advance !!!
Regards,
Rahul
[Message sent by forum member 'rahul_akkina' ]
http://forums.java.net/jive/thread.jspa?messageID=371716