Hi list,
We are currently in the process of setting up a production environment
with resin 4.0.x on our Linux servers. Currently we are testing resin
4.0.7.
First of all I would like to mention that the watchdog documentation
http://caucho.com/resin-4.0/admin/starting-resin-watchdog.xtp
is repeating point 1 - 7 in point 8 to 14.
Thanks, that will be fixed with the next doc release.
We have experience with resin 3.0.x from earlier, so the biggest change
for us is the watchdog.
Experiments with a single, shared watchdog has however left us with some
questions:
1) With a single, shared watchdog we manage to specify where the
watchdog places its logs with the -log-directory argument to resin.jar,
however that also results in the jvm-logs for each resin application
will be placed in the same directory. Especially on our development
server it would be convenient to have these located along with the logs
of the application, like we get when running a watchdog for each
application. Is it possible to specify in watchdog.xml where the jvm for
each application should be placed?
You can change the log's location in resin.xml configuration file using
log-hanlder.
log-handler name= level=all path=/var/logs/${server.id}.log
timestamp=[%y-%m-%d %H:%M:%S.%s] {%{thread}} /
2) In addition to experimenting with the watchdog we have been testing
/resin-admin which might turn quite useful for some of the applications
we are running. However, we were wondering if it is possible to run one
single /resin-admin in watchdog.xml in order to monitor all our
applications? If that is not possible I guess we have to sett up a
/app-admin for each applications we want to monitor.
For historical reasons we have a setup where each applications has its
own resin.xml file, and we would like to keep it like this for the
future as well since we find it easier to administrate.
The watchdog's purpose is to keep resin instances running, (re)starting them
when necessary (failures). It doesn't run any apps, but the /resin-admin
deployed on any node in the cluster will track every server in the cluster.
3) What is the gain of having only one watchdog?
Needing fewer resources – each watchdog would be a separate JVM otherwise.
Different physical machines will have separate instances of watchdog.
Thanks,
Alex
That is, do we use less
resources by having only one watchdog with several applications, than
having one watchdog per application?
--
Halvor Utby
___
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
___
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest