Sorry Servers & Recovery

brian.cunningham · ‎05-17-2007

According to the document below when the primary services in a content rule recover after an outage, traffic will not automatically get routed from a sorry server back to the primary services. The document states that you either have to suspend the service of the sorry server or shutdown the sorry server. Is there a way around this?

http://www.cisco.com/warp/public/117/css_sorry_server.html

We'd like the following scenario to occur for an outage:

1) The two services in the content rule go down.

2) Traffic gets routed automatically to the configured sorry server(s).

3) The two services (from #1) come back online.

4) The CSS automatically reroutes all traffic for the content rule to the two services, instead of the sorry server.

RODRGUTI · ‎05-17-2007

Hello Brian,

Basically, when we use a sorry server, the Content rule is going to send traffic to that server just when all the "production" services are down.

Now, what happen when the "production" services are alive again?

R/ The content rule is going to stop sending new connections to the sorry server, but the current connections that are been handled by the sorry server are going to be finalized normally, the sorry service is going to wait until the connection ends.

The sorry service wont get new connections, but wont kill or re-route the connections to the production services.

There is no way to make the sorry server to re-route the connections to any other service.

Hope this help.

- Rodrigo.

brian.cunningham · ‎05-17-2007

That helps. Thanks, Rodrigo!

The Cisco document was a little confusing the way it was worded. I don't think the Cisco manuals even cover the topic of what happens after the production services are alive again? What you described is what I'd expect from sorry server functionality. We'll start adding that functionality to our CSS strategy.