Possible to Run Two Identical Web2py Instances on Two Computers for Active/Active Failover Purposes?

[Tristan Britt]

Context:
I have a working web2py website that does some internal data visualization for my job, where employees can upload files and visualize different metrics. My boss would like me to take a functional windows server cluster with two active nodes, and have the web2py webapp running on both nodes. Once the data gets uploaded, it doesn't get changed, just visualized. The windows cluster nodes have a shared volume with the web2py code and the databases it creates shared between them.

Goal:
In the event one node goes down there is already another instance of the web2py data visualization webapp running on the other node, and the only failover that has to happen is that the cluster url redirects users to the 2nd nodes web2py instance. 

Is this, or something similar, possible? I've tried running the webapps on the same port only to get errors, and can run the webapp on different ports, but wonder if there might be any issues with two instances accessing the same database?

[Jim S]

It should work just fine, no issues with the database  as long as you aren't doing something silly with it.  But, there are things to think about.

1.  Create a load balancer that will balance the requests to both systems and will transfer requests should one system go down.  We use haproxy

2.  Sessions need to be stored in a shared place accessible to both systems.  Use redis or some caching system or store sessions in the database

I'm on my way out the door or I'd give more details.  Hit me back up next week if you are still stuck.

-Jim