I'd agree with Larry.
- For single instance case, you can use something like SMF which will restart the domain when it abruptly fails (e.g. see asadmin create-service command).
- For cluster case, the node-agent will take care of restarting failed instances.
The only slight problem is I don't know of a way of restarting the instance/domain only if it fails and the JVM process exits with a known exit-code. The above two ways will restart the entities regardless of exit-code the JVM exits with.
Maybe v3 has better support for this, because v3 can be started with "java -jar glassfish.jar"
and then you can monitor the returned process's pid and monitor it in a shell script.
Hope that helps.
-Kedar
[Message sent by forum member 'km' (km)]
http://forums.java.net/jive/thread.jspa?messageID=345972