Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
3.6.3
-
None
-
None
-
None
Description
I am using the latest stable release 3.6.3. Here is the minimal workload to reproduce the bug:
- There is a zookeeper cluster with three nodes: node1, node2 and node3;
- Start the cluster from scratch;
- node2 becomes the leader;
- Stop node2;
- node3 becomes the new leader;
- Client requests to create a znode;
- Restart node2;
- node2 receives SNAP from node3;
- node2 receives the NEWLEADER message from node3;
- node2 takes current snapshot;
- node2 crashes before writing to the current epoch file;
- Restart node2 but it fails since "Unable to load database on disk".