# Tikv Server Oom Impact

+ Unsuccessful request on TiKV caused an abnormal abort
+ Re-election of regional leader 
  + Raft Group starts elect a new region leader
  + The new region leader reports information to the PD Server
+ Regions cache is updated frequently
  + When accessing the region cache of TiDB Server, an error related to TiKV RPC occurs
  + Automatic backoff retries in the background
  + PD return the latest region leader information to the region cache of the TiDB Server
------------------------------------------------------------
TiKV Server is a data storage node in a TiDB cluster. When the TiKV Server restart due to some abnormal reasons, such as OOM, the following symptoms occur:
- A request failure on the TiKV server causes an abnormal abort, and a corresponding error is return to the application
- Re-election of regional leader
  - Raft Group begins to elect a new region leader
  - New Region Leader reports information to PD Server
- Region cache is updated frequently
  - When a new request accesses the region cache of the TiDB Server to obtain a region, an error related to TiKV rpc may occur
  - Backoff is automatically retried in the background after a related error occurs in the region cache
  - PD return the latest region leader information to the region cache of TiDB Server
It can be seen that when an OOM condition occurs on the TiKV server, the original business request will fail, while the new business request will increase the request time due to the elect of the region and the backoff retry.