When an http request is processed through the Apache web server it typically forks a brand new process unless of course the first is using something similar to fastcgi.

My question relates to "synchronised demands" when utilizing fastcgi.

If I am using fastcgi and I've got a tree-like datastructure in primary memory, should i be worried about concurrent read/write accessibility tree?

Or can one just depend on the truth that demands are processed within the order they arrive.

Let's say one request attempts to access the disk also it blocks? Would be the other demands processed or will they stand in a queue?

If I am not using fastcgi, things appear clearer since I must reload the tree data structure from the database to control it after which write it to a database - no concurrency needed.

Basically, should i be worried about multiple visitors/creates to my primary memory data structures with Apache?

Thanks ahead of time.

When an http request is processed through the Apache web server it typically forks a brand new process

No, usually among pre-forked processes accepts the bond and executes it. There isn't any fork-per-request

In case your FastCGI application is single threaded, you shouldn't be worried about concurrecy, also should you run in mod-prefork. But when you manage your computer data structures in shared memory you should be worried about concurrency.