Moving from single dedicated server to cloud based multiple node solution
I currently use Wowza on a single dedicated server. Simple install and maintenance. I’m now moving to a cloud solution and need a little advice about the software and implementation.
On the cloud solution I have 2 web server nodes, Web01 and Web02. Wowza will be installed on Web01 and will run on that single node unless the node fails and Wowza will be started on Web02. Now considering the configuration, application data etc for Wowza will be identical for each node, I wanted to find out if it’s safe to share configuration and application data between nodes. For example I thinking of having [install-dir]/conf and [install-dir]/applications directory shared between both Wowza installs. There’s a chance that both Wowza instances could be running at the same time or that the first instance will not be shut down correctly before the second node is fired up. Is this safe to do? What are best practices in this situation?
Also I don’t want to be paying double subscription as even though I have 2 instances, they are technically exactly the same and I only ever need one running at one moment in time. What are best practices to avoid paying for 2 instances in my situation?
It should be possible to have two similar Wowza instances available and to flip between one or the other. You would need to ensure that they are properly replicated, i.e. if making a change on node1 then ensuring that is available on node2. If the instance is fairly static then the most robust method would be to apply any config changes to node 1 via the engine manager, and then replicate this on node 2 by carrying out the same steps.
The main thing to be aware of with regards to the billing aspect as if both nodes are active then you would get changed for each one. So you would need to have robust methods in place to prevent that if you wish to avoid incurring extra billing.
Unfortunately there are no simple solutions for this as you do need to work on your side to ensure both nodes aren't active at the same time and also that configuration/assets etc are replicated properly, which may be best done as a manual task on both nodes.