Support S3 bucket as a storage provider for multi-instance setup

[ScribblerCoder]

The documentation mentions that ReadWriteMany PVC must be created to be able to deploy multi-instance GZCTF setup. The problem with ReadWriteMany volumes backed by NFS is that they manage multiple write operations poorly. This is solved by Object storage e.g. S3 buckets... For example CTFd (one of the most popular established CTF platforms) supports using S3 as a storage provider.

Why S3?

• Supports concurrent writes
• Scalable for large amounts of unstructured data (e.g., media files, static files, challenge files/attachments)
• Supports versioning for each object, allowing you to recover previous versions or recover objects if accidentally deleted or changed
• Lower cost than PVCs
• Supports high availability by default

If the user doesn't wish to depend on cloud providers for S3 (local setup). They can deploy their own:

• MinIO
• SeaweedFS

I can create a feature request if you like

[GZTimeWalker]

I know what is s3, this feature request will be considered in develop group. Thanks for the suggestion.

But currently, with lots of competitions were held, we found the single instance gives a very high performance, so we not suggest to deploy multi-instance in competition with less than 5000 teams / 250 challenges.

[hez2010]

This is a feature we have in our mind.

I can create a feature request if you like

I'm converting this to an issue for better tracking.

[ScribblerCoder]

Done in #365 thanks @GZTimeWalker . Will close this