[TiDB Usage Environment] Production Environment
[TiDB Version] v4.0.10
[Reproduction Path] Operations that led to the issue
Performed unsafe downscale operations on 2 TiKV nodes
After deleting the region cache, the issue persisted
[Encountered Issue: Symptoms and Impact]
Cluster write operations are slow, DM to TiDB write operations are abnormal, with error message Error 9005: Region is unavailable. There are 10 regions without a leader, as shown in the image
It is estimated that these 10 regions need to perform unsafe recovery. First, check if all replicas of these 10 regions have 2 replicas in the TiKV you took offline. If so, use unsafe recovery to restore them. Data loss may occur.