When I access the site, and get the message the site is offline for maintenence, whats going on? I would think that HA in both unplanned (crash) but also planned should be a goal of developing j2EE applications.
Is there something as a container vendor we can do to make customers sites more available?
Internet Applications Division
If the unavailable page is displayed, it is for one of several reasons:
1) We are actually performing maintenance (so Weblogic was taken down to update some code, etc).
2) Apache timed out waiting for the WL servlet engine to respond to web request.
This can happen due to:
a) Too much load on the box (Weblogic can't reply in time).
b) Weblogic has hung (this almost never happens).
3) Weblogic is down (no one to respond to Apache web requests). This also rarely happens, since we have scripts set up to check for this and restart the server.
In response to your question, I would say that a container vendor should offer adcanced failover/clustering support, features that mission critical sites require to be fool-proof.