DevOps, CI/CD and Automation

I am supporting an app server for a Forms & Reports application that uses run_report_object to run reports on a Reports Server. This is on 10gR2 (10.1.2.0.2) on Solaris & Linux.

Currently running multiple standalone Reports Servers on a single server (for 3 different projects). Need to prevent one project from affecting another as much as possible (yes, I know they are on the same box....).

I have a requirement to ensure that a hanging Reports Server doesn't prevent other jobs from running. I see that Reports Server clustering is no longer an option. The Reports documentation about how to implement HA basically say just read the AS docs and figure it out. The problem is that using the AS J2EE HA means that you'd HAVE to use the Reports Servlet. Additionally, I don't see how the reports servlet could be called so that more than one Reports server would be used.

I keep hearing about an impending doc about implementing HA on Reports. When is it coming?

Future environments will be multiple AS installs on separate boxes accessing a RAC DB.

So, what's the scoop? How can I implement this?

So, here are my questions:

1. (Two Servers) I believe I could do this on two servers if I had the same Reports Server name on each. If so, could I have two OC4J containers on each server both with the same mount point for the rwservlet application? What else would I need to do to tell OHS to load balance between both since they have the same name.

2. (Two Servers) If 1 above is okay, would it be possible to have two OC4J containers point to two different standalone Reports Servers?

3. (One Server) Can I start up multiple Reports Servers on same box with the same name (I'm pretty sure this is no :-) )

4. (One Server) Like 2 above except on a single server. Two independent OC4J containers accessing two different servlets. Each has a default Reports Server setup. URL would look the same from each.

I'm kind of graping for straws here on how to move ahead. I'm surprised the Reports Server clustering was taken away, but I suspect it is because the underlying Visibroker architecture was changed and there weren't the resources required to redo this. I'm not sure that is a good thing.

Any help would be appreciated.

:-) Steve