We're having problems with a production Glassfish cluster that is load balanced via apache/mod_jk. During moderately heavy load, the AJP ports stop listening on the GF instances, killing the site. We have to restart the nodeagents, kill some running GF processes, and restart the nodeagents to recover. This system has been running for months, and has been fairly stable, but the last few weeks we've had 3 outages caused by this.
We're running Glassfish 2.1 Patch 04 with tomcat 5.5.17 tomcat-ajp.jar
on Solaris 10 Sparc
Latest tomcat is 5.5.29
and
Apache/2.2.14 (Unix) DAV/2 mod_jk/1.2.28.
Latest mod_jk is 1.2.29
Has anyone had similar problems?
[Message sent by forum member 'dgulino']
http://forums.java.net/jive/thread.jspa?messageID=470805