2 instances/machine; redhat init script

[alpheratz-jb]

I have looked at $JBOSS_HOME/bin/jboss_init_redhat.sh.

This seems inadequate if one wants to have 2 JBoss instances/machine...

It can easily be copied/hacked to run $JBOSS_HOME/bin/run.sh -c config for each separate instance.

BUT I can't see how the code for stop/status can relate to WHICH instance to stop/report status on.

Is this a bug/need for enhancement?

Has there anyone out there who has solved this and is willing to contribute their solution?

I have a REALLY BAD/hacky solution (can't post now...code is at work, I'm at home) I use twiddle.sh/probe.sh.

I probably could use something like kill `ps axw | grep config | cut ...` for 'stop' but not for 'status'...

This applies to 4.0.5GA and 4.2...., as far as I can see.

Thoughts/comments/suggestions gratefully accepted.

Cheers,

Alph

[areplogle]

If you've setup multiple instances per machine and you know your rmi port per instance (I use bindings.xml to specify this) then you can do shutdown.sh -s jnp://localhost:

ie. shutdown.sh -s jnp://localhost:1099 && shutdown.sh -s jnp://localhost:1199 etc


Hope this helps.


Andrew

[jaraco]

I ran into this problem also, but I'm only running one instance. I want to run it behind Apache, and I have other Tomcat instances running on the machine, so using the recommendation from http://wiki.jboss.org/wiki/ConfiguringMultipleJBossInstancesOnOneMachine, I set up my instance to run on an internal IP address. Unfortunately, as you encountered, the scripts don't handle status or stop when the instance IP is specified.

I fixed the issue with 'stop' by including the -s parameter to the Shutdown call.

In particular, I added two lines during the setup phase of jboss_init_redhat.sh:


#if JBOSS_HOST specified, use -b to bind jboss services to that address
JBOSS_BIND_ADDR=${JBOSS_HOST:+"-b $JBOSS_HOST"}

#if JBOSS_HOST is specified, the stop command also requires a server parameter
JBOSS_STOP_SERVER=${JBOSS_HOST:+"-s $JBOSS_HOST"}

#define the classpath for the shutdown class
JBOSSCP=${JBOSSCP:-"$JBOSS_HOME/bin/shutdown.jar:$JBOSS_HOME/client/jnet.jar"}


and then I included ${JBOSS_STOP_SERVER} in the JBOSS_CMD_STOP following --shutdown:

JBOSS_CMD_STOP=${JBOSS_CMD_STOP:-"java -classpath $JBOSSCP org.jboss.Shutdown --shutdown ${JBOSS_STOP_SERVER}"}


This works. Apparently, if the RMI port is never specified, it need not be included in the shutdown URI.

I suggest this be added to the released scripts, as it does not affect users who do not specify JBOSS_HOST and it only behaves better if the user does specify JBOSS_HOST.

Regards,
Jason

[brian.stansberry]

Can you open a JIRA for this?