Storage: Fix page_id being mis-reuse when upgrade from cluster < 6.5 (#9041) (release-7.1)

[JaySon-Huang]

This is an automated cherry-pick of #9041

What problem does this PR solve?

Issue Number: close #9039

Problem Summary:

#9039 (comment)

The root cause is:

1. When the cluster upgrade from version lower than v6.5.0 to higher or equal to v6.5.0, TiFlash will try to migrate the data from PageStorage V2 to PageStorage V3.
2. When the table has no data in the delta layer, it will run under PageStorage V3 mode
3. But we get a wrong max_id for this situation, we must use max_data_page_id = std::max(data_storage_v2->getMaxId(), data_storage_v3->getMaxId()); to avoid the page_id from being mis-reused. Otherwise it will cause data corruption and make tiflash fails to restart in the next restart

tiflash/dbms/src/Storages/DeltaMerge/StoragePool.cpp

Lines 518 to 520 in eb6272c

	max_log_page_id = log_storage_v3->getMaxId();
	max_data_page_id = data_storage_v3->getMaxId();
	max_meta_page_id = meta_storage_v3->getMaxId();

What is changed and how it works?

• In StoragePool::restore, after "meta" and "data" transformed into PSV3, update the max_id by the maximum of that id in PSV2 and PSV3
• Review other codes that call raiseXXXPageIdLowerBound()
• Add unit test case

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
  1. Deploy a cluster with tiflash replica and load some data, the cluster version < 6.5.0
  2. Before upgrade, alter table ... compact tiflash replica to make the table has no data in the delta layer (this raises the possibility of the bug happening)
  3. Upgrade the cluster to a version >= 6.5.0 and write more rows into table, then alter table ... compact tiflash replica
  4. Restart the tiflash instance
• 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

Fix the issue that metadata may be corrupt and make tiflash panic after upgrading from a cluster less than v6.5.0

[JaySon-Huang]

/run-all-tests

[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:

• 2024-05-16 05:54:42.701195361 +0000 UTC m=+1719036.458330937: ☑️ agreed by Lloyd-Pottiger.
• 2024-05-16 06:29:33.000156269 +0000 UTC m=+1721126.757291867: ☑️ agreed by JinheLin.

[JaySon-Huang]

/cherry-pick release-6.5

[ti-chi-bot]

@JaySon-Huang: new pull request created to branch release-6.5: #9056.

In response to this:

/cherry-pick release-6.5

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the ti-community-infra/tichi repository.