Then we rely on variable substitution in server.xml for items which might need to be specific such as ip addresses, port numbers, and worker names.
We never call startup.sh directly - we have wrapper scripts which set the appropriate -D variables for us.
-Tim
Ryan Daly wrote:
On Mon, 2004-11-29 at 14:53, Shapira, Yoav wrote:
Hi, I do it this way all the time. The biggest pro is complete separation. If one app gets whacky (heavy load, OutOfMemoryError, malicious attack), others are not affected at all. If one app needs to be upgraded or a bug fix shipped, or its server restarted, again, others are not affected at all.
How do you handle the server.xml and the other configuration? Do you start Tomcat with an option to read a different config file, or do you have separate installations of Tomcat for each and every webapp?
Cons include inability to run these apps on the same port (this can be overcome by putting a common front end, which I usually do with another Tomcat running only the balancer app), and increased monitoring/maintenance needed (which again can be mitigated with tools like nagios).
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
