# Problem: Slow Region Replication after Node Failure


![`smart`](./student-guide/diagram/SlowRegionReplicationAfterNodeFailure_aa_001_20221017.png)
----------------------------------------------------------------------------------------------------------
- Replication of a failed replica of a Store node is slow
  - Observe the build execution of the PD monitoring analysis operator
  - If no corresponding operator schedule is generated, there are possible reasons
    - Offline scheduling is turned off, or replica-schedule-limit is set to 0
    - Unable to find a node to move the region. If the capacity of the replacement node with the same label is greater than 80%, PD will stop scheduling in order to avoid the risk of explosion
  - If the schedule generates normally, it's just slow
    - Adjust replica-schedule-limit appropriately according to business pressure
    - Consider adding evict leader to the offline store
    - Observe if there is competition from other scheduling tasks, or if TiKV is under too much pressure