Sun, 29 Sep 2002 22:17:25 -0700 (PDT) "Craig R. McClanahan" <[EMAIL PROTECTED]> said:
> Regarding restarting a webapp without restarting Tomcat, you should read
> up on the Manager servlet that comes with Tomcat 4.0 and 4.1:
>   http://jakarta.apache.org/tomcat/tomcat-4.1-doc/manager-howto.html
If the manager servlet can remove a webapp with a servlet that
tries to use 100%CPU and never comes back, then that is it. Perfect.
I more than happily withdraw the idea of using Apache as server side
proxy to access several independent Tomcats via port 80 in order
to reduce downtime.
I am sorry I overlooked that until now. Thank you for the pointer.

On dealing with rogue (runaway, stuck, ..) servlets:

> what you mean by this) either way.  The only answer is to correct the bugs
> in your app that are causing the servlet not to complete its responses.
Almost agree. But this looks only at the servlet programmers side.
As server administrator I also want to minimize the damage
a single broken servlet can do. A local problem should stay local.
This not a problem in a perfect world. Reality is not always perfect.
Recently I had a servlet that worked fine for 9996 times out of 10000.
In the other 4 cases it would use all CPU it could get until somebody killed
Tomcat.

That is why
> > I searched quite a while but failed to find a way to
> > monitor and catch individual runaway servlet threads.
Being lazy as I am, I would like to be able to set resource ceilings
on servlets, forcing servlets to behave and programmers to act, not
the server administrator.

Thank you again
Oskar
--
Dr. Oskar Bartenstein                 [EMAIL PROTECTED]
IF Computer Japan                  http://www.ifcomputer.com


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to