WIP: Draft implementation for DRBD replication to secondary node (per volume)

[chbmuc]

This is a first draft for volume redundancy with local volumes on multiple nodes. Replication is done using drbd instances per volume.

This is untested and there still are potential issues that could potentially affect data integrity, so DO NOT USE in produktion.

[Gerrit91]

Thanks for this contribution, it's a very cool idea.

For bigger pull requests like this, it would be great to first open an issue and discuss a new feature with people or bring in the idea in our biweekly planning meeting. Maintaining such a complex feature is potentially time consuming and I am currently not sure if we have the capacities to do this. :/

For a review, please make sure that the pull request passes the CI pipeline. Please also provide tests for the feature. As this feature is about failovers, I think that specific failover tests should be tested, too. To accelerate this process, you maybe want to demo this feature and show how it works.

I don't want to put obstacles in the way. It's just quite some people use this csi-driver and we need maintainers who can spare time on this specific topic in order to support this feature in the long run.