I just pushed a series of security fixes to the RHQ server that disallow the
CLI scripts running on the server (i.e. the alert scripts) from doing local
JNDI lookups (other than through the "blessed" interface provided to the
scripts using the *Manager predefined variables).
This prevents the scripts from accessing the RHQ server's internals like local
SLSBs, entity manager, raw database connection, etc.
If you see any unexpected security exceptions while running the master build,
shout at me :)
The only time you should see a security exception is when your alert script
tries to do something illegal (exit JVM, do a JNDI lookup inside the RHQ
server, ...). You should NEVER see a security exception nor any change in
functionality if the alert scripts are not involved.
Lukas
Show replies by date