High Availability/Failover question

litechforce

I was just thinking of this scenario and wanted some input.

If I created another instance of ITFlow, but had mysql replication enabled on both databases, would that work? The idea is that if the server ITFlow was running on needed to be updated, but the business requirement is that ITFlow is always running (24/7 support center, etc). So if ITFlow instance A went down, Instance B could just take over.

Thoughts?

wrongecho_hiatus

Hey,

You could put the file uploads on an NFS share and then have config.php set to use the clustered database IP address. Then each ITFlow app instance is just a dumb app worker behind a load balancer (ideally with session affinity).

You'd have to recreate the golden image for each update though as the files change and you probably wouldn't wanna serve the core files over NFS due to latency.

The challenge then becomes maintaining the load balancer, network storage, and database cluster. How deep are your pockets and how many 9s do you actually need? If you run this in something like AWS, they offer ELB, RDS and EFS which are essentially managed HA versions of these.

Edit: I haven't actually tried any of this, but it works okay when we do it with WordPress

litechforce

wrongecho_hiatus

My pockets are quite shallow. I was just thinking of the scenario. I don't actually have to do this, though I was toying with the idea. I didn't even think about the client files there, so good call. The load balancer could be a simple round robin DNS entry or I could use something like HAProxy

wrongecho_hiatus

https://fastgood.cheap/

Bigbug

wrongecho_hiatus a like isn't enough…

🤣🤣🤣