01-11-2013 07:05 AM
Hello Guys,
I'm currently trying to develop a way to Load Balancing Multiple Applications running in one server.
I have three groups of applications running on a server listening on dfferents ports. For example: 7777 7778 7779
The main idea is to shut down a service in that port that stop listening for connections and not the whole server.
When Configuring probes in the server farm, if a port fails, for example port 7777. The ACE shuts down the whole server and load balances the load to the other server. I want the ACE to shut down that port in particular and not the whole server....
Is there a way to LOAD BALANCING multiple applications in the ACE.
Thanks in advance. I'm looking everywhere for information but can't find any......
Regards
José Rodríguez
Solved! Go to Solution.
01-11-2013 07:22 AM
If I am understanding your question correctly, you want to be able to use a single server to handle requests for a subset of TCP ports. But do not want the server taken out service for other ports, if it stops responding on a single port. If this is the case you can build a separate serverfarm for each TCP port. Apply the specific TCP port probe to the serverfarm and not under the rserver. This way the rserver will only be taken out of service for that particular serverfarm and not take the rserver out of service globally when it fails the probe.
For example one way to do this might be:
Match for VIP A and TCP port 1
Serverfarm configured to handle requests for port 1 and under the serverfarm configure a probe for port 1
Match for VIP A and TCP port 2
Serverfarm configured to handle requests for port 2 and under the serverfarm configure a probe for port 2
Etc.
01-11-2013 07:53 AM
Yes, that way the probe that you associate with the serverfarm would be tailored to the application running on the specific port. When the rserver fails the probe associated with the serverfarm, it would be taken out of service only for that serverfarm. It would remain inservice for the other serverfarms.
01-11-2013 07:22 AM
If I am understanding your question correctly, you want to be able to use a single server to handle requests for a subset of TCP ports. But do not want the server taken out service for other ports, if it stops responding on a single port. If this is the case you can build a separate serverfarm for each TCP port. Apply the specific TCP port probe to the serverfarm and not under the rserver. This way the rserver will only be taken out of service for that particular serverfarm and not take the rserver out of service globally when it fails the probe.
For example one way to do this might be:
Match for VIP A and TCP port 1
Serverfarm configured to handle requests for port 1 and under the serverfarm configure a probe for port 1
Match for VIP A and TCP port 2
Serverfarm configured to handle requests for port 2 and under the serverfarm configure a probe for port 2
Etc.
01-11-2013 07:35 AM
Thanks for your answer
The real deployment is usually two servers running multiple mirrored applications.
But I think that this could be done using your logic, defining a match for the VIP and the ports running the application.. Also defining multiple server farms to separate the application running on each farm.
Am I right?
01-11-2013 07:53 AM
Yes, that way the probe that you associate with the serverfarm would be tailored to the application running on the specific port. When the rserver fails the probe associated with the serverfarm, it would be taken out of service only for that serverfarm. It would remain inservice for the other serverfarms.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide