[Disk Manager] Implement filesystem backups

[qkrorlqr]

This issue is mostly about the Disk Manager part. Filestore checkpointing is out of scope here - in this issue we need to implement filesystem backups by simply reading the current state of the filesystem. Yes, it will not be consistent, but:

1. it's much better than nothing
2. after we add checkpoints we will simply switch the implementation that is going to appear in this issue from reading the current state to making a checkpoint and reading the state from that checkpoint

We need to keep in mind that we expect to have filesystems up to several hundreds of TiB in size (maybe even up to 1PiB).

[qkrorlqr]

There is actually an alternative approach for backups: create a VM, attach Filestore to it and do a simple rsync from Filestore to a S3-based FS.