TiDB latency not stable and TiKV OOM if running cdc changefeed when titan is on and average region size is large #17696
Description
Bug Report
What version of TiKV are you using?
v8.4.0
31aa015
What operating system and CPU are you using?
X84, K8S
Steps to reproduce
- Deploy v8.4.0 cluster with 6 TiKV node (16C64G), titan is on.
- br restore 4000 tables (~10TB)
- start log backup task
- create one CDC kafka changefeed to replicate the workload to Kafka
- run workload
What did you expect?
TiDB latency should be stable, when running workload
What did happened?
TiDB latency not stable, and TiKV OOM seen.
