if you care about lost updates you should use transaction or at least an API that support optimistic concurrency control.

Posix File system API don’t offer much for concurrency control other than lock.

In my line of work we use HDFS or AWS S3 as “filesystem” so we are not using the OS file access API at all and most File are append only by a single server.

I don't understand - you suggested using the filesystem, but filesystems don't have good concurrency control. S3's concurrency handling is even worse than filesystems. Something like Kafka is a much better way to do a stream of commands that need to be transformed to eventually produce some kind of result (which is really almost all workloads).

No this was a misunderstanding I said if you need (concurrency control/transaction and index) use a real database not a filesystem, since it’s the whole reason database exist.

But if you don’t need it there are plenty of good distributed file system that will not eat your data.

> No this was a misunderstanding I said if you need (concurrency control/transaction and index) use a real database not a filesystem, since it’s the whole reason database exist.

Just because something is designed for x doesn't mean it's any good at x. Kafka is also designed for handling concurrent events, and it's significantly better at it.

> But if you don’t need it there are plenty of good distributed file system that will not eat your data.

There really aren't any distributed filesystems that are nice to work with. S3 isn't a full filesystem and your writes take a while before they're visible in your reads, HDFS isn't a full filesystem and can be slow to failover, AFS has generally weird performance and locks up every so often. If your needs are simple then anything will work, including distributed filesystems, but it's very rare that a distributed filesystem is an actively good choice, IME.