Hi,
We have a web server(apache 2.0.52, mod_jk/1.2.28) on Red Hat Linux, connect with JBoss (JBoss [Trinity] 4.2.3.GA). After the server ran for few hours, all the AJP connections piled up with CLOSE_WAIT status. At the same time, the JBoss threads also increased. Finally, the apache server response status 503 "service unavailable". I got the thread dump, most threads are like this:
"ajp-0.0.0.0-8113-93" daemon prio=1 tid=0x0000002bc6be0680 nid=0x4aab in Object.wait() [0x000000004ba08000..0x000000004ba08e30]
at java.lang.Object.wait(Native Method)
- waiting on <0x0000002ac6e7d918> (a org.apache.tomcat.util.net.JIoEndpoint$Worker)
at java.lang.Object.wait(Object.java:474)
at org.apache.tomcat.util.net.JIoEndpoint$Worker.await(JIoEndpoint.java:415)
- locked <0x0000002ac6e7d918> (a org.apache.tomcat.util.net.JIoEndpoint$Worker)
at org.apache.tomcat.util.net.JIoEndpoint$Worker.run(JIoEndpoint.java:441)
at java.lang.Thread.run(Thread.java:595)
If we restart JBoss, the site can run for a while. Then the problem would happen again while the AJP connections pile up.
This is workers.properties file:
worker.node2.port=8113
worker.node2.host=potato
worker.node2.cachesize=100
worker.node2.type=ajp13
worker.node2.connection_pool_timeout=300
worker.node2.connect_timeout=5000
worker.node2.prepost_timeout=5000
worker.node2.socket_timeout=10
worker.node2.socket_keepalive=True
worker.list=node2
server.xml in JBoss:
<Connector port="8113" address="${jboss.bind.address}" protocol="AJP/1.3"
emptySessionPath="false" enableLookups="false" redirectPort="8111"
maxThreads="600" connectionTimeout="300000"/>
Any suggestion to resolve this issue ? Thanks,
Jimy
There is no workaround for this bug except using APR AJP or the patch https://jira.jboss.org/jira/secure/attachment/12316577/patch.txt