Best practices for running behind a load balancer

Archive
1

Are there guidelines on how to run Grav behind a load balancer? I don’t mind if the Admin doesn’t have redundancy, but the content generated for users visiting to my site needs more resiliency than a single machine can provide.

Could we do something as simple as rsync a set of directories on a machine running the Admin to set of machines running core?

Also, are there caveats for running Grav core behind a load balancer? I ran across an issue on the Grav git page stating that sticky sessions are necessary to avoid cache conflicts between servers. Has anyone run into any other issues?

2

As Grav is file based, then as long as the files are the same on both machines, you should have no problems. If there’s ‘dynamic’ stuff going on that would mean data is going to be modified by the end users, then you will need to have some sort of rysnc going on to ensure the files stay in sync. That’s pretty much it.

3

That’s great to hear, man. I’m planning on a one way sync of the grav/user directory from an admin machine to multiple core machines.