Skip to content

Storage: Adapt to ingesting snapshot with irregular region range#10151

Merged
ti-chi-bot[bot] merged 5 commits intopingcap:masterfrom
JaySon-Huang:fix_irregular_region_range
May 9, 2025
Merged

Storage: Adapt to ingesting snapshot with irregular region range#10151
ti-chi-bot[bot] merged 5 commits intopingcap:masterfrom
JaySon-Huang:fix_irregular_region_range

Conversation

@JaySon-Huang
Copy link
Contributor

@JaySon-Huang JaySon-Huang commented May 8, 2025

What problem does this PR solve?

Issue Number: close #10147

Problem Summary: As the issue describe

What is changed and how it works?

  • Adapt to irregular region range with any suffix, not only "\x00"
  • Refine the irregular region range parsing warning that we can get more information for better diagnosing
    • Turn RowKeyValue(bool is_common_handle_, HandleValuePtr value_) to RowKeyValue::fromHandleWithSuffix(bool is_common_handle_, HandleValuePtr value_)
    • Move the parse warning from parsing RowKeyValue to parsing RowKeyRange::fromRegionRange
Storage: Adapt to ingesting snapshot with irregular region range
If the key format is "t${tableID}_r${handleID1}{AnySuffix}", and `AnySuffix` is not empty, `"t${tableID}_r${handleID1}{AnySuffix}" > "t${tableID}_r${handleID1}"` according to the the comparison semantics of "Key" on TiDB/TiKV.

For example,
If a table is non-clustered, that is with Int64 as the HandleID type.
And the range of a Region is `["xxx_r{handleId1}{AnySuffix}", "xxx_r{handleId2}{AnySuffix}")`, where handleId1 = 100, handleId2 = 200.

If `AnySuffix` is empty, then Region contains the left-closed-right-open range of `[handleID1, handleID2)`
If `AnySuffix` is not empty, then Region contains the range `(handleID1, handleID2]`, which is actually the left-closed-right-open range of `[handleID1+1, handleID2+1)`.

Check List

Tests

  • Unit test
  • Integration test
  • Manual test (add detailed scripts or steps below)
create table if not exists t (a bigint primary key, b int);
insert into t values (1,2),(2,3),(3,4),(4,5),(20, 50), (21, 51),(200002,2),(200003,3),(200004,4), (9223372036854775807, 101),(9223372036854775806, 100),(-9223372036854775808, -101),(-9223372036854775807, -100);

// The keys is related to the table_id. For these verification cases, the table_id=471

tiup ctl pd -u {pd_ip}:{pd_port} -i
# split with table_id=471, row_id=20, suffix='\x00'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F728000000000FF0000140000000000FB
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=20, suffix='\x01'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F728000000000FF0000140100000000FB
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=20, suffix='\x0130FF'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F728000000000FF0000140130FF0000FD
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=-9223372036854775808, suffix=''
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F720000000000FF0000000000000000FA
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=-9223372036854775808, suffix='\x01'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F720000000000FF0000000100000000FB
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=9223372036854775807, suffix=''
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F72FFFFFFFFFFFFFFFFFF0000000000FA
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=9223372036854775807, suffix='\x01'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F72FFFFFFFFFFFFFFFFFF0100000000FB
# then add tiflash replica
alter table set tiflash replica 1;
  • 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 TiFlash might panic when handling snapshot with irregular region range

@ti-chi-bot ti-chi-bot bot added release-note-none Denotes a PR that doesn't merit a release note. size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels May 8, 2025
Signed-off-by: JaySon-Huang <tshent@qq.com>
@JaySon-Huang JaySon-Huang force-pushed the fix_irregular_region_range branch from b36f3d4 to dc991fe Compare May 8, 2025 03:37
@ti-chi-bot ti-chi-bot bot added release-note Denotes a PR that will be considered when it comes time to generate release notes. size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. and removed release-note-none Denotes a PR that doesn't merit a release note. size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels May 8, 2025
@JaySon-Huang JaySon-Huang force-pushed the fix_irregular_region_range branch from 27775b4 to 1f6f775 Compare May 8, 2025 06:53
@JaySon-Huang
Copy link
Contributor Author

/test pull-integration-test

@JaySon-Huang JaySon-Huang force-pushed the fix_irregular_region_range branch from 90c29e8 to 5a114c5 Compare May 8, 2025 14:02
Signed-off-by: JaySon-Huang <tshent@qq.com>
Signed-off-by: JaySon-Huang <tshent@qq.com>
Signed-off-by: JaySon-Huang <tshent@qq.com>
@ti-chi-bot ti-chi-bot bot added approved needs-1-more-lgtm Indicates a PR needs 1 more LGTM. labels May 9, 2025

assert(!is_common_handle_);
// According to the tidb encoding rule, the int handle must be larger than 8 bytes.
RUNTIME_CHECK_MSG(
Copy link
Member

@CalvinNeo CalvinNeo May 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So we allow(I mean RUNTIME_CHECK_MSG) an encoded handle longer than sizeof(int64) here, but will still panic if we meet an encoded handle shorter than sizeof(int64)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, according to @tangenta this case should not happen. If it happens, then tidb-server would panic.

Copy link
Contributor Author

@JaySon-Huang JaySon-Huang May 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

And if the value_size is less than 8 bytes, then the DB::DecodeInt64 will likely access random invalid address, which will cause random panic.
Adding a check here so that we can get explicit error rather than random panic.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've try to generate a split key like

  • t{tableID}_r\x30
  • t{tableID}_r\x00\x00\x00\x00\x00\x00\x01

TiKV will refuse to split, with logging like

[2025/05/09 17:41:36.242 +08:00] [INFO] [peer.rs:6460] ["invalid split request"] [source=pd] [peer_id=2078] [region_id=2077] [err="KeyNotInRegion([116, 128, 0, 0, 0, 0, 0, 1, 255, 255, 215, 95, 114, 0, 0, 0, 0, 255, 0, 250, 95, 114, 0, 0, 0, 0, 255, 0, 0, 1, 0, 0, 0, 0, 0, 250], id: 2077 start_key: 7480000000000001FFD75F730000000000FA end_key: 7480000000000001FFD800000000000000F8 region_epoch { conf_ver: 26 version: 217 } peers { id: 2078 store_id: 1 } peers { id: 2166 store_id: 132 role: Learner })"] [thread_id=377]

Signed-off-by: JaySon-Huang <tshent@qq.com>
@ti-chi-bot ti-chi-bot bot added size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files. and removed size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. labels May 9, 2025
@JaySon-Huang
Copy link
Contributor Author

/hold

@ti-chi-bot 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 May 9, 2025
Copy link
Member

@CalvinNeo CalvinNeo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@ti-chi-bot
Copy link
Contributor

ti-chi-bot bot commented May 9, 2025

[APPROVALNOTIFIER] This PR is APPROVED

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

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

The pull request process is described here

Details Needs approval from an approver in each of these files:
  • OWNERS [CalvinNeo,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 ti-chi-bot bot added lgtm and removed needs-1-more-lgtm Indicates a PR needs 1 more LGTM. labels May 9, 2025
@ti-chi-bot
Copy link
Contributor

ti-chi-bot bot commented May 9, 2025

[LGTM Timeline notifier]

Timeline:

  • 2025-05-09 02:45:46.330926904 +0000 UTC m=+67612.478888122: ☑️ agreed by Lloyd-Pottiger.
  • 2025-05-09 05:45:40.966560053 +0000 UTC m=+78407.114521271: ☑️ agreed by CalvinNeo.

@JaySon-Huang
Copy link
Contributor Author

/unhold
Verified that tiflash/tikv can handle the region with split key using int64_max correctly

@ti-chi-bot 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 May 9, 2025
@ti-chi-bot ti-chi-bot bot merged commit b35176b into pingcap:master May 9, 2025
5 checks passed
@JaySon-Huang JaySon-Huang deleted the fix_irregular_region_range branch May 9, 2025 07:44
@ti-chi-bot ti-chi-bot bot added needs-cherry-pick-release-8.5 Should cherry pick this PR to release-8.5 branch. needs-cherry-pick-release-8.1 Should cherry pick this PR to release-8.1 branch. needs-cherry-pick-release-7.5 Should cherry pick this PR to release-7.5 branch. labels May 9, 2025
ti-chi-bot pushed a commit to ti-chi-bot/tiflash that referenced this pull request May 9, 2025
Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>
@ti-chi-bot
Copy link
Member

In response to a cherrypick label: new pull request created to branch release-8.5: #10157.
But this PR has conflicts, please resolve them!

ti-chi-bot pushed a commit to ti-chi-bot/tiflash that referenced this pull request May 9, 2025
Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>
@ti-chi-bot
Copy link
Member

In response to a cherrypick label: new pull request created to branch release-8.1: #10158.
But this PR has conflicts, please resolve them!

@ti-chi-bot
Copy link
Member

In response to a cherrypick label: new pull request created to branch release-7.5: #10159.
But this PR has conflicts, please resolve them!

ti-chi-bot pushed a commit to ti-chi-bot/tiflash that referenced this pull request May 9, 2025
Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>
ti-chi-bot bot pushed a commit that referenced this pull request May 9, 2025
) (release-8.5) (#10157)

close #10147

Storage: Adapt to ingesting snapshot with irregular region range
If the key format is "t${tableID}_r${handleID1}{AnySuffix}", and `AnySuffix` is not empty, `"t${tableID}_r${handleID1}{AnySuffix}" > "t${tableID}_r${handleID1}"` according to the the comparison semantics of "Key" on TiDB/TiKV.

For example,
If a table is non-clustered, that is with Int64 as the HandleID type.
And the range of a Region is `["xxx_r{handleId1}{AnySuffix}", "xxx_r{handleId2}{AnySuffix}")`, where handleId1 = 100, handleId2 = 200.

If `AnySuffix` is empty, then Region contains the left-closed-right-open range of `[handleID1, handleID2)`
If `AnySuffix` is not empty, then Region contains the range `(handleID1, handleID2]`, which is actually the left-closed-right-open range of `[handleID1+1, handleID2+1)`.

Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>
Signed-off-by: JaySon-Huang <tshent@qq.com>

Co-authored-by: JaySon <tshent@qq.com>
Co-authored-by: JaySon-Huang <tshent@qq.com>
JaySon-Huang added a commit that referenced this pull request May 20, 2025
)

cherry-pick of #10151


### What problem does this PR solve?

Issue Number: close #10147

Problem Summary: As the issue describe

### What is changed and how it works?

* Adapt to irregular region range with any suffix, not only "\x00"
* Refine the irregular region range parsing warning that we can get more
information for better diagnosing
* Turn `RowKeyValue(bool is_common_handle_, HandleValuePtr value_)` to
`RowKeyValue::fromHandleWithSuffix(bool is_common_handle_,
HandleValuePtr value_)`
* Move the parse warning from parsing `RowKeyValue` to parsing
`RowKeyRange::fromRegionRange`

```commit-message
Storage: Adapt to ingesting snapshot with irregular region range
If the key format is "t${tableID}_r${handleID1}{AnySuffix}", and `AnySuffix` is not empty, `"t${tableID}_r${handleID1}{AnySuffix}" > "t${tableID}_r${handleID1}"` according to the the comparison semantics of "Key" on TiDB/TiKV.

For example,
If a table is non-clustered, that is with Int64 as the HandleID type.
And the range of a Region is `["xxx_r{handleId1}{AnySuffix}", "xxx_r{handleId2}{AnySuffix}")`, where handleId1 = 100, handleId2 = 200.

If `AnySuffix` is empty, then Region contains the left-closed-right-open range of `[handleID1, handleID2)`
If `AnySuffix` is not empty, then Region contains the range `(handleID1, handleID2]`, which is actually the left-closed-right-open range of `[handleID1+1, handleID2+1)`.
```

### Check List

Tests <!-- At least one of them must be included. -->

- [x] Unit test
- [ ] Integration test
- [x] Manual test (add detailed scripts or steps below)
```
create table if not exists t (a bigint primary key, b int);
insert into t values (1,2),(2,3),(3,4),(4,5),(20, 50), (21, 51),(200002,2),(200003,3),(200004,4), (9223372036854775807, 101),(9223372036854775806, 100),(-9223372036854775808, -101),(-9223372036854775807, -100);

// The keys is related to the table_id. For these verification cases, the table_id=471

tiup ctl pd -u {pd_ip}:{pd_port} -i
# split with table_id=471, row_id=20, suffix='\x00'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F728000000000FF0000140000000000FB
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=20, suffix='\x01'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F728000000000FF0000140100000000FB
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=20, suffix='\x0130FF'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F728000000000FF0000140130FF0000FD
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=-9223372036854775808, suffix=''
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F720000000000FF0000000000000000FA
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=-9223372036854775808, suffix='\x01'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F720000000000FF0000000100000000FB
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=9223372036854775807, suffix=''
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F72FFFFFFFFFFFFFFFFFF0000000000FA
# then add tiflash replica
alter table set tiflash replica 1;

# split with table_id=471, row_id=9223372036854775807, suffix='\x01'
operator add split-region 2077 --policy usekey --keys 7480000000000001FFD75F72FFFFFFFFFFFFFFFFFF0100000000FB
# then add tiflash replica
alter table set tiflash replica 1;
```
- [ ] 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

<!-- bugfix or new feature needs a release note -->

```release-note
Fix the issue that TiFlash might panic when handling snapshot with irregular region range
```

Signed-off-by: JaySon-Huang <tshent@qq.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

approved lgtm needs-cherry-pick-release-7.5 Should cherry pick this PR to release-7.5 branch. needs-cherry-pick-release-8.1 Should cherry pick this PR to release-8.1 branch. needs-cherry-pick-release-8.5 Should cherry pick this PR to release-8.5 branch. release-note Denotes a PR that will be considered when it comes time to generate release notes. size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

TiFlash panic with "Detected illegal region boundary"

4 participants