Agent to Gateway Fail over

I have come across some interesting test results with regards to Agent fail over and gateways. I am working with an environment that has server (agents) in different regions around the world (Latin America (LA), Asia (AP) and North America (NA) ). We have placed two Gateways in each region (which are also fully trusted child domains of the larger corporate global forest) for fail over. We have also set up Auto Agent Fail over with in the Admin pane so that LA Agents report to the 2 LA gateways, same in AP, NA etc. What we wanted to find out is what happens when both Gateways go offline (in a fully trusted environment). We already know that in a non-trusted environment where we are using certs and have to deal with Firewalls that the agents will queue their data until the GW come back online. Now in an environment where the domains are all trusted and there are no firewalls it appears that if both GW servers go offline the agents take it upon them selves to search for a report to the RMS.

You can see this by looking for an Event 21017 in the Ops mgr Evt log.

Something to keep in mind when planning fail over strategies.

If you are looking for more great information and the best technical training available on Operations Manager 2007, check out http://www.infrontconsulting.com/events.htm.

Read the complete post at http://david-stclair.spaces.live.com/Blog/cns!112A71B19678F08D!176.entry