Merge/Tombstone clean coordination example

[danthegoodman1]

Support coordination for concurrent merges to prevent duplicate data.

Maybe we make a class to implement that will handle the locking and such, with examples?

The class would need:

• Lock files
• Release lock

Example options:

• Redis
• etcd
• Zookeeper/CH Keeper
• Postgres/CRDB
• C*/Scylla with LWT

Merging locks should be by table

[danthegoodman1]

Should we just have examples of how to do this?

[danthegoodman1]

with the new s3 conditional writes, this can even be done in pure S3!