current connections on probe fail

[shday]

Is there a way to configure a probe so that if the probe fails allow the server to be taken out of the laodbalance pool, but to allow the current connections to stay connected.

I want to be able to have the probe to fail automatically wait long enough for all the connections to bleed off and then take the server out of loadbalance pools so that I can patch it or update it.

[Daniel Arrondo Ostiz · Accepted solution]

Hi,

The behavior you are describing is the default one for all the load-balancers. When a probe goes down, no new connections will be sent to the failed real server, but, the connections that were already there will remain.

You may also be interested to know that it's also possible to configure the device to automatically transfer all the connections from one server to the remaining ones in the farm in case the probes for that server go down. This is described for the ACE module in the link below, but, you also have similar commands for CSM or CSS.

http://www.cisco.com/en/US/docs/interfaces_modules/services_modules/ace/vA2_3_0/configuration/slb/guide/rsfarms.html#wp1117375

Daniel