Feature(2.3): add compaction task for delta files

[zipper-meng]

Rationale for this change

Related #1244.

Conclusion

Others

1. DataBlock - Add method split_at(self, index)-> (DataBlock, DataBlock) and intersection(self, time_range) -> Option<DataBlock>.
2. ColumnFile, LevelInfo - Implement std::fmt::Display, also add ColumnFiles<F: AsRef<ColumnFile>>(&[F]) and LevelInfos(&[LevelInfo]) to decrease code when we need to log them.
3. TimeRange, add function compact(time_ranges: &mut Vec<TimeRange>) to compact time ranges.
4. TimeRanges, add method push(&mut self, time_range: TimeRange), extend_from_slice(&mut self, time_ranges: &[TimeRange]) to modify a TimeRanges.

Compaction

• Changes measurement of compaction from "db", "ts_family", "level" to "db", "ts_family", "in_level", "out_level".
• Add file delta_compact.rs for delta compaction:
  1. Pick delta files and the out_level, get the out_level_max_ts of the out_level.
  2. Now we have many delta files, for each field_id in these files, find grouped blocks which should be merged together: Vec<CompactingBlockMetaGroup>.
  3. For each CompactingBlockMetaGroup, get merge-split blocks(only data in time_range 0..=out_level_max_ts is needed) and write them into files in the out_level, store these files in VersionEdit.
  4. Write a temporary tsm-tombstone file for the delta-file, named xxxxxxx.tombstone.compact.tmp, includes all it's field_ids and time_range 0..=out_level_max_ts, because data in the time_range is already merged into the out_level. Tombstones will be compacted by compact(time_ranges: &mut Vec<TimeRange>) before write.
  5. Write the VersionEdit into summary, and rename these xxxxxxx.tombstone.compact.tmp files to xxxxxxx.tombstone.

Task definition

Some changes in CompactTask:

• Normal(from old CompactTask::Vnode) - Compact the files in the in_level into the out_level.
• Cold(from old CompactTask::ColdVnode) - Flush caches and then compact the files in the in_level into the out_level.
• Delta - Compact the files in level-0 into the out_level.

Some changes in CompactReq:

• Add field lv0_files: Option<Vec<Arc<ColumnFile>>> to store files in delta compaction.

Pick

• Remove trait Picker: Send + Sync + Debug, add two function instead of it:
  • pick_level_compaction(compact_task: CompactTask, version: Arc<Version>) -> Option<CompactReq>
  • async pick_delta_compaction(compact_task: CompactTask, version: Arc<Version>) -> Option<CompactReq>
• Add function pick_delta_compaction only for picking delta files with a tsm file to partly merges data in level-0 files into the destination level, the merged data will leave a tombstone file for the source level-0 files.
• Improve unit test code.

Scheduler

• Now schedule_compaction() not only check if a tseries family is cold but also check num of level-0 files with config compact_trigger_file_num or constant DEFAULT_COMPACT_TRIGGER_DETLA_FILE_NUM (This change makes unit tests in summary.rs won't stop, so I fixed it by some changes in unit test cases in summary.rs).
• schedule_compaction() holds a shared reference of Arc<TseriesFamily, now we must manually stop it by TseriesFamily::close(). Implementation of Drop for Database will call that method of all it's tseries families.
• Compact job structure compaction::job::CompactProcessor is changed to hold the map of (TsFamilyId, IsDeltaCompaction) to ShouldFlushBeforeCompaction.

Tombstone

Add tombstone type ALL, now a tombstone can not only specifies an excluded time range for a field but can also specify the time range for all fields.

pub enum TombstoneField {
     One(FieldId), // 0x00, u64
     All,          // 0x01
 }

---

As we know the tombstone file is record file, and each record is like:

+--------------+--------------+-----------+------------+--------------+---------------+
| 0: 4 bytes   | 4: 1 byte    | 5: 1 byte | 6: 4 bytes | 10: 4 bytes  | 14: data_size |
+--------------+--------------+-----------+------------+--------------+---------------+
| magic_number | data_version | data_type | data_size  | crc32_number |     data      |
+--------------+--------------+-----------+------------+--------------+---------------+

We can control the decoding of tombstones by data_type in record header.

Tombstone record v1

The v1 tombstone record, I think it's sparse.

+------------+---------------+---------------+
| 0: 8 bytes | 8: 8 bytes    | 16: 8 bytes   |
+------------+---------------+---------------+
|  field_id  | min_timestamp | max_timestamp |
+------------+---------------+---------------+

Tombstone record v2

Each record may contains many fields and time ranges.

# field_typ = FIELD_TYPE_ONE(0x00)
+-----------------+------------+-----------------+---------------+---------------+----
| 0: 1 byte       | 1: 8 bytes | 9: 4 bytes      | 8: 8 bytes    | 16: 8 bytes   | ...
+-----------------+------------+-----------------+---------------+---------------+----
| field_typ(0x00) |  field_id  | time_ranges_num | min_timestamp | max_timestamp | ...
+-----------------+------------+-----------------+---------------+---------------+----

# field_typ = FIELD_TYPE_ALL(0x01)
+-----------------+-----------------+---------------+---------------+----
| 0: 1 byte       | 9: 4 bytes      | 8: 8 bytes    | 16: 8 bytes   | ...
+-----------------+-----------------+---------------+---------------+----
| field_typ(0x01) | time_ranges_num | min_timestamp | max_timestamp | ...
+-----------------+-----------------+---------------+---------------+----

TODOs

There may be some logs for debug may be deleted.

Are there any user-facing changes?

No.

[roseboy-liu]

1. Generating a large number of delta files, verifying the consistency of the merged data, and completing the merging of the delta files.
2. Version 2.3.2 is compatible with tomb files and summary files. There is a compatibility upgrade for the tomb file format, which needs to be verified.