Description of the impact of adding cop cache on the execution plan

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

Original topic: 增加cop cache对执行计划影响的描述

| username: h5n1

As the title suggests, it is recommended that the official documentation for cop cache include descriptions of its impact on execution plans, such as total_process_keys, etc.:

| username: Billmay表妹 | Original post link

Received, feedback provided!

| username: Billmay表妹 | Original post link

It seems that this information already exists:

Note:
Another available optimization solution in the above example is the coprocessor cache. If you cannot add an index, you may consider enabling the coprocessor cache feature. Once enabled, as long as the Region has not changed since the last execution of the operator, TiKV will return values from the cache. This also helps to reduce most of the computation costs of the TableFullScan and Selection operators.

| username: h5n1 | Original post link

It’s not this, it’s the relationship between the cache and the TiKV scan within the execution plan.

| username: ajin0514 | Original post link

Observe the cache.

| username: Billmay表妹 | Original post link

When the cache is hit, the data is directly obtained from TiDB without needing data from the TiKV side. However, the TiKV scan operator is still required as it needs to access TiKV to determine whether it is safe to use the cache.

| username: h5n1 | Original post link

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