Storages: Make DMFile ready for new column indexes/types #8756

JaySon-Huang · 2024-02-05T04:39:50Z

What problem does this PR solve?

Issue Number: ref #6233, close #8768

Problem Summary:

In the near future, we may add new types (vector type) or new indexes. But currently the ColumnStat is not flexible enough for extending new fields for the index size, array size, etc.

What is changed and how it works?

Add new type ExtendColumnStat meta block under DMFile meta v2, which accept protobuf base ColumnStat
Add array_sizes_bytes/array_sizes_mark_bytes for ColumnStat to support vec type
Try to merge array_size(.size0.dat) into X.merged file

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

Release note

None

ti-chi-bot · 2024-02-06T04:06:55Z

PR needs rebase.

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 kubernetes/test-infra repository.

JaySon-Huang · 2024-02-06T13:23:03Z

/run-all-tests

JaySon-Huang · 2024-02-06T14:55:35Z

/run-unit-test

breezewish · 2024-02-07T02:13:36Z

dbms/src/Storages/DeltaMerge/File/DMFile.cpp

+#else
+        // ExtendColumnStat is not enabled yet because it cause downgrade compatibility, wait
+        // to be released with other binary format changes.
+        writeExtendColumnStatToBuffer(tmp_buffer),


How about we write the new format data at the end (and also write old format in the old place)? In this way it may be possible to keep downgrade compatibility as new format will be regarded as extra data and will be discarded.

Good suggestion! The ColumnStat is small enough to ignore its impact on performance and data size. We can keep it until we actually need an incompatible storage format changed by other components.
I'll change it

After some tests, I found the downgrade compatibility can not be satisfied because there is a default branch 😂

tiflash/dbms/src/Storages/DeltaMerge/File/DMFile.cpp

Lines 1055 to 1059 in d9dc96b

default:

throw Exception(

ErrorCodes::INCORRECT_DATA,

"MetaBlockType {} is not recognized",

magic_enum::enum_name(handle->type));

Maybe we can just logging instread of throwing exception here.

Good idea. What kind of downgrade compatibility do we need to have? How about we first land the code that will not throw error (and will not change the format at all) for a few major versions, and then land the compatible format? In this way we will have downgrade compatibility for 1 major version. I think Vector will not land pingcap/tiflash in 6 months, so that there could be enough time for us to make the change.

There is only downgrade compatibility requirement between patch versions under the same minor version. And no requirement across minor or major versions now. But it is good to have.

I've removed the exception thrown branch in this PR because the correctness of bytes read is protected by checksum. I think ignoring unknown meta block types without exception or logging is an acceptable behavior.

@JinheLin @breezewish

dbms/src/Storages/DeltaMerge/File/DMFile.cpp

dbms/src/Storages/DeltaMerge/File/DMFileReader.cpp

dbms/src/Storages/DeltaMerge/File/DMFileWriter.cpp

dbms/src/Storages/DeltaMerge/File/DMFile.h

dbms/src/Storages/DeltaMerge/File/DMFile.cpp

JaySon-Huang · 2024-02-15T05:49:14Z

/run-all-tests

CalvinNeo

lgtm

ti-chi-bot · 2024-02-18T02:46:11Z

[LGTM Timeline notifier]

Timeline:

2024-02-18 02:34:42.321961049 +0000 UTC m=+151771.069584160: ☑️ agreed by CalvinNeo.
2024-02-18 02:46:10.763367893 +0000 UTC m=+152459.510991004: ☑️ agreed by Lloyd-Pottiger.

Lloyd-Pottiger · 2024-02-18T02:46:15Z

/hold

ti-chi-bot · 2024-02-21T03:32:07Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: CalvinNeo, 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~~ [CalvinNeo,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

JaySon-Huang · 2024-02-22T05:16:56Z

/unhold

JaySon-Huang · 2024-02-22T05:17:29Z

/run-all-tests

JaySon-Huang · 2024-02-22T05:59:45Z

/run-integration-test

JaySon-Huang · 2024-02-22T06:13:45Z

/run-unit-test

ti-chi-bot · 2024-02-22T06:14:06Z

@JaySon-Huang: Your PR was out of date, I have automatically updated it for you.

At the same time I will also trigger all tests for you:

/run-all-tests

trigger some heavy tests which will not run always when PR updated.

If the CI test fails, you just re-trigger the test that failed and the bot will merge the PR for you after the CI passes.

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.

ti-chi-bot bot added do-not-merge/needs-linked-issue release-note-none Denotes a PR that doesn't merit a release note. needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files. labels Feb 5, 2024

JaySon-Huang force-pushed the refactor_dmfile_1 branch 2 times, most recently from 5d665a7 to 683de5a Compare February 5, 2024 08:47

ti-chi-bot bot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Feb 5, 2024

JaySon-Huang force-pushed the refactor_dmfile_1 branch from 683de5a to a03986d Compare February 6, 2024 02:09

ti-chi-bot bot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Feb 6, 2024

JaySon-Huang force-pushed the refactor_dmfile_1 branch 2 times, most recently from e19a5ea to 7e3b18c Compare February 6, 2024 10:50

JaySon-Huang changed the title ~~[DNM] Storages: Ready for new column indexes/types~~ Storages: Make DMFile ready for new column indexes/types Feb 6, 2024

JaySon-Huang requested review from breezewish, JinheLin and Lloyd-Pottiger February 6, 2024 13:15

ti-chi-bot bot removed the do-not-merge/needs-linked-issue label Feb 6, 2024

JaySon-Huang removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Feb 6, 2024

breezewish reviewed Feb 7, 2024

View reviewed changes

Lloyd-Pottiger reviewed Feb 7, 2024

View reviewed changes

JinheLin reviewed Feb 7, 2024

View reviewed changes

dbms/src/Storages/DeltaMerge/File/DMFile.h Outdated Show resolved Hide resolved

JaySon-Huang added 7 commits February 7, 2024 15:29

Prepare for new indexes or data types

6e12532

Enhance tests

3ea3c8a

Remove debug logging

e5a3b24

Remove duplicate code

4b2d30b

Fix unstable test

2dd81f1

Fix test

04dcf52

Address comment

bdc1188

ti-chi-bot bot added the do-not-merge/needs-triage-completed label Feb 7, 2024

Merge remote-tracking branch 'upstream/master' into refactor_dmfile_1

dceea42

JinheLin reviewed Feb 7, 2024

View reviewed changes

dbms/src/Storages/DeltaMerge/File/DMFile.cpp Show resolved Hide resolved

JinheLin reviewed Feb 7, 2024

View reviewed changes

dbms/src/Storages/DeltaMerge/File/DMFile.cpp Outdated Show resolved Hide resolved

breezewish self-assigned this Feb 13, 2024

JaySon-Huang added 2 commits February 15, 2024 13:29

Address comment

d381e53

Merge remote-tracking branch 'upstream/master' into refactor_dmfile_1

84b5ee0

ti-chi-bot bot removed the do-not-merge/needs-triage-completed label Feb 15, 2024

CalvinNeo approved these changes Feb 18, 2024

View reviewed changes

ti-chi-bot bot added needs-1-more-lgtm Indicates a PR needs 1 more LGTM. approved labels Feb 18, 2024

Lloyd-Pottiger approved these changes Feb 18, 2024

View reviewed changes

ti-chi-bot bot added lgtm and removed needs-1-more-lgtm Indicates a PR needs 1 more LGTM. labels Feb 18, 2024

ti-chi-bot bot added the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Feb 18, 2024

JinheLin approved these changes Feb 21, 2024

View reviewed changes

Merge branch 'master' into refactor_dmfile_1

c4a4b96

ti-chi-bot bot removed the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Feb 22, 2024

Merge branch 'master' into refactor_dmfile_1

283ba53

ti-chi-bot bot merged commit b0c365c into pingcap:master Feb 22, 2024
6 checks passed

JaySon-Huang deleted the refactor_dmfile_1 branch February 23, 2024 03:08

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Storages: Make DMFile ready for new column indexes/types #8756

Storages: Make DMFile ready for new column indexes/types #8756

JaySon-Huang commented Feb 5, 2024 •

edited

Loading

ti-chi-bot bot commented Feb 6, 2024

JaySon-Huang commented Feb 6, 2024

JaySon-Huang commented Feb 6, 2024

breezewish Feb 7, 2024 •

edited

Loading

JaySon-Huang Feb 7, 2024 •

edited

Loading

JaySon-Huang Feb 7, 2024

JinheLin Feb 21, 2024

breezewish Feb 21, 2024 •

edited

Loading

JaySon-Huang Feb 21, 2024 •

edited

Loading

JaySon-Huang commented Feb 15, 2024

CalvinNeo left a comment

ti-chi-bot bot commented Feb 18, 2024

Lloyd-Pottiger commented Feb 18, 2024

ti-chi-bot bot commented Feb 21, 2024

JaySon-Huang commented Feb 22, 2024

JaySon-Huang commented Feb 22, 2024

JaySon-Huang commented Feb 22, 2024

JaySon-Huang commented Feb 22, 2024

ti-chi-bot bot commented Feb 22, 2024

	default:
	throw Exception(
	ErrorCodes::INCORRECT_DATA,
	"MetaBlockType {} is not recognized",
	magic_enum::enum_name(handle->type));

Storages: Make DMFile ready for new column indexes/types #8756

Storages: Make DMFile ready for new column indexes/types #8756

Conversation

JaySon-Huang commented Feb 5, 2024 • edited Loading

What problem does this PR solve?

What is changed and how it works?

Check List

Release note

ti-chi-bot bot commented Feb 6, 2024

JaySon-Huang commented Feb 6, 2024

JaySon-Huang commented Feb 6, 2024

breezewish Feb 7, 2024 • edited Loading

Choose a reason for hiding this comment

JaySon-Huang Feb 7, 2024 • edited Loading

Choose a reason for hiding this comment

JaySon-Huang Feb 7, 2024

Choose a reason for hiding this comment

JinheLin Feb 21, 2024

Choose a reason for hiding this comment

breezewish Feb 21, 2024 • edited Loading

Choose a reason for hiding this comment

JaySon-Huang Feb 21, 2024 • edited Loading

Choose a reason for hiding this comment

JaySon-Huang commented Feb 15, 2024

CalvinNeo left a comment

Choose a reason for hiding this comment

ti-chi-bot bot commented Feb 18, 2024

[LGTM Timeline notifier]

Lloyd-Pottiger commented Feb 18, 2024

ti-chi-bot bot commented Feb 21, 2024

JaySon-Huang commented Feb 22, 2024

JaySon-Huang commented Feb 22, 2024

JaySon-Huang commented Feb 22, 2024

JaySon-Huang commented Feb 22, 2024

ti-chi-bot bot commented Feb 22, 2024

JaySon-Huang commented Feb 5, 2024 •

edited

Loading

breezewish Feb 7, 2024 •

edited

Loading

JaySon-Huang Feb 7, 2024 •

edited

Loading

breezewish Feb 21, 2024 •

edited

Loading

JaySon-Huang Feb 21, 2024 •

edited

Loading