[KUDU-1330] Add tool to unsafely recover from loss of majority (or all) tablet replicas

• Type: New Feature
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.7.0
• Fix Version/s: 1.4.0
• Component/s: ops-tooling
• Assignee: Dinesh Bhat
• Reporter: Todd Lipcon
• Created: 16/Feb/16 03:05
• Updated: 19/Jan/18 01:58
• Resolved: 19/Jan/18 01:58

Description

bruceSz ran into this issue: if you have a table with replication set to 1, and you permanently lose a node, the table is stuck in a bad state. It would be nice to allow the operator to accept the data loss and replace the lost tablet with a new (empty) one.

Similarly, I've had a few people ask about the scenario where you lose 2 of 3 replicas and you are willing to accept data loss, and force recovery from the remaining one replica.

We should add a tool (or tools) for these recovery scenarios.

Issue Links

• is blocked by: KUDU-1721 Expose ability to force an unsafe Raft configuration change
• is duplicated by: KUDU-1504 Add a tool to force a Raft config change