Description
Use cases:
-Taking periodic snapshots of a cluster's DB files and being able to restore exactly to that point in time
-Transferring data from one cluster to another (of the same size) either for the purposes of node expansion, data center recovery, setting up a test cluster, etc
The two high level areas that need to be addressed are:
-vbucket states need to be written into the database files (or alternately to another location that will need to be kept in sync, but it seems better to put it directly into the DB files)
-upon restoration, the Erlang process needs to understand the vbucket states that have been written and use those to restore the same states in the new cluster
Attachments
Issue Links
- duplicates
-
MB-3001 Better coping with changing IP addresses for cloud based deployments
- Closed