Investigating the Root Cause of Increased Pending Compaction Bytes Accumulation Due to Continuous Scaling

[translator_bot]

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: 续扩容引起pending compaction bytes累计增加排查根因

| username: robert233

---

[TiDB Usage Environment]

• Production

[TiDB Version]

• v4.0.12

[Encountered Issues]

• Original post: TiKV 扩容过程中，Pending compaction bytes 一直在增加 - TiDB 的问答社区
• Issue encountered: During scaling, pending compaction bytes keep accumulating

[Problem Phenomenon and Impact]

• Adjusted rocksdb.max-background-jobs from 8 to 12 and restarted the cluster for recovery.

• Matching metrics, rocksdb cpu remained at 100% during scaling
  

  

• View rocksdb cpu metric description as follows:
  

  

• Raised a few questions:
  (1). Is rocksdb single-threaded, and which parameters are related to resource limitations?
  (2). After adjusting rocksdb.max-background-jobs from 8 to 12, why did the CPU increase so much?
  (3). This metric was not found in the official template for corresponding alerts, is it necessary?
  raft store cpu
  

  
  
  rocksdb cpu
  
  

[translator_bot]

| username: h5n1 | Original post link

---

When was this adjustment made? RocksDB is not single-threaded, and TiKV has parameters to control compaction and flush.

[translator_bot]

| username: system | Original post link

---

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.