Help Center/ MapReduce Service/ Component Operation Guide (LTS) (Ankara Region)/ Using Elasticsearch/ Common Issues About Elasticsearch/ What Can I Do If High Read I/O Occurs When Document IDs Are Specified in the Scenario When the Data Written into the Database Reaches a Certain Volume?
Updated on 2024-11-29 GMT+08:00
View PDF
Share

What Can I Do If High Read I/O Occurs When Document IDs Are Specified in the Scenario When the Data Written into the Database Reaches a Certain Volume?

Symptom

When the data written into the database reaches a certain volume in the scenario where the document IDs are specified, the read I/O is high.

Cause Analysis

Run the iotop command on the EsNode. The disk read rate of a large number of Elasticsearch threads is high. Based on the information about the thread stacks, a full data query is required when document IDs are specified for writing requests to determine whether the specified indexes exist in the document IDs in the process of writing the commands for bulk load. The query process occupies a large number of disk read I/Os.

Solution

The logic is modified on the service side to enable that no document ID is specified during data writing. Instead, index fields are specified.