Data Center Orchestration Part 1: Add Resource

Hi! This is Joel Hendrickson and I’m going to show a demonstration of how the F5 Management Pack can be used to respond to a traffic spike by adding server resources.

I have System Center Operations Manager up on my screen showing a couple of different views. In the top pane, I’m looking at the diagram view of a load balancing pool with three pool members.

In the middle pane, I’m looking at a graph of the current number of connections for each pool member with the graph legend on the bottom. Two of my pool members are handling the traffic load, while a third pool member is a standby server. The standby server has been setup in a lower priority group on the big IP, so that it will only receive traffic if the high priority group of servers encounters a problem.

Let’s look at virtual machine manager. I’ve got two web servers currently running and serving traffic while a third web server has shutdown until it’s resources are needed. I’m now going to spike the traffic to the virtual server using a traffic simulation program. This will cause the big IP to start sending some traffic to the standby server. I’ve setup a trigger using the F5 management pack to respond to this scenario by using virtual machine manager to start a new server followed by adding that new server into the big IPs primary load balancing pool.

Let’s look back at Virtual Machine Manager. The virtual machine is now starting up. In operations manager, the graph shows that the connections have spiked and some connections are being routed to the standby server. Virtual Machine Manager shows that the virtual machine is now fully started. In operations manager, I’ll refresh the diagram view. This shows that another pool member has been added to this pool.

Notice that the pool member marked in a critical state is our standby server. I’ll refresh the performance graph and we can see that the standby server’s connections dropped back to zero as it is no longer needed. The original pool members are now able to fully handle the traffic. This causes the health state of the standby server to return to a green or successful state.


Transcription by: Scribe4you Transcription Services