-
Notifications
You must be signed in to change notification settings - Fork 323
Closed
Description
Problem Description:
- Node replicas cannot continue syncing when the batch message sync write process partially fails.
Cause Analysis:
- The write of the batch replica fails partly. For example, the replica index is 1-4, index=1-3 is successfully written, but index=4 fails.

- The master node sync request timed out , and resync the replicas with index=1-4.

3.The slave node rewrites the batch replicas, but at this time writeIndex=3, request.getFirstEntryIndex() = 1, the check throws an exception, but the exception is not handled, causing the master node to time out again. Repeat the above process.

Metadata
Metadata
Assignees
Labels
No labels
