Storages: Update the length of Block queue and the number of read threads (#8702)

[ti-chi-bot]

This is an automated cherry-pick of #8702

What problem does this PR solve?

Issue Number: close #8703

What is changed and how it works?

• The length of Block queue:

  • For on-premise deployment, adjusting the length of the Block queue to 1.5 times of the current default value is expected to reduce the time required for simple queries by 3-6%.
  • For disaggregated deployment, adjusting the length of the Block queue to 1.5 times of the current default value is expected to reduce the time required for simple queries by 2-20%. The specific performance improvement depends on data caching. The less cached data, the more significant the improvement.

• The number of read threads:

  • For disaggregated deployment, if most of the data cannot hit the local disk cache, the network will become a bottleneck, and increasing the number of threads can significantly improve performance. If more than half of the data is expected to hit the local cache, adjusting the number of threads to 2 times of the current default value is enough.

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

None

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: JinheLin, Lloyd-Pottiger

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [JinheLin,Lloyd-Pottiger]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[ti-chi-bot]

[LGTM Timeline notifier]

Timeline:

• 2024-01-19 01:45:10.587059532 +0000 UTC m=+493152.151357239: ☑️ agreed by JinheLin.
• 2024-01-19 02:22:41.470859062 +0000 UTC m=+495403.035156761: ☑️ agreed by Lloyd-Pottiger.

[JinheLin]

/run-all-tests

[JinheLin]

[2024-01-19T03:12:16.920Z] fullstack-test/expr/agg_pushdown.test: Running

[2024-01-19T03:12:34.964Z]   File: fullstack-test/expr/agg_pushdown.test

[2024-01-19T03:12:34.964Z]   Error line: 29

[2024-01-19T03:12:34.964Z]   Error: set @@tidb_isolation_read_engines='tiflash'; select substr(c, 2), count(1) from test.t group by substr(c, 2) order by substr(c, 2)

[2024-01-19T03:12:34.964Z]   Result:

[2024-01-19T03:12:34.964Z]     ERROR 1105 (HY000) at line 1: DB::TiFlashException: Failed to register MPP Task MPP<gather_id:<gather_id:1, query_ts:1705633947849904343, local_query_id:481, server_id:1094, start_ts:447121702997196805, resource_group: default>,task_id:2>, reason: task is already registered

[2024-01-19T03:12:34.964Z]   Expected:

[2024-01-19T03:12:34.964Z]     +--------------+----------+

[2024-01-19T03:12:34.964Z]     | substr(c, 2) | count(1) |

[2024-01-19T03:12:34.964Z]     +--------------+----------+

[2024-01-19T03:12:34.964Z]     |              |       64 |

[2024-01-19T03:12:34.964Z]     | BC           |       64 |

[2024-01-19T03:12:34.964Z]     | EF           |       64 |

[2024-01-19T03:12:34.964Z]     +--------------+----------+

[JinheLin]

/run-integration-test