* feat(log-store): add the object store WAL batch, catalog and I/O
Add the three modules between the object format and the store of the
object store WAL:
- batch: the open batch that accumulates admitted entries into the next
object and assigns entry ids at admission. Ids are object-sequence-major,
`(object_seq << 20) | position`, with positions starting at one per
region per object, so a batch that is rolled back and admitted again
under the same sequence hands out the same ids. The time the first entry
was admitted is kept for the store's age-based sealing; an empty
admission does not start it.
- catalog: the in-memory index over object footers, by sequence and per
region. Insertion is atomic and rejects an empty footer, duplicate
region segments, invalid entry ranges, an already indexed sequence and
entry ranges that are not strictly increasing across objects. The next
object sequence continues after the largest indexed one and is raised
above the largest entry id of any region, and is rejected once it no
longer fits an entry id.
- io: object store access under `<prefix>/objects/`: a conditional create
whose retry with identical content is a no-op and whose conflicting
content is rejected, whole and range reads, and a listing of well-formed
keys in sequence order. A prefix with a `.` or `..` component is
rejected. When the store reports that the object exists but the read
that compares its content fails, the read failure is returned with its
retry hint instead of the create failure.
The modules have no callers until the store lands, so they are declared
with `#[allow(dead_code)]`. Only `entry_id` is exported.
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* test(log-store): drop comments that restate the batch test assertions
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* refactor(log-store): use pub(crate) in the object store WAL batch, catalog and I/O
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* fix(log-store): keep entry_id crate-private and narrow the create collision check
Re-export `entry_id` as `pub(crate)`: nothing outside `log-store` uses it
yet, and exporting it would freeze the raw `(object_seq, position)`
encoding before the store owns id allocation.
Treat only `ConditionNotMatch` as the sign that a conditional create
collided with an existing object. A store may report `AlreadyExists` for
an unrelated path, for example when a parent of the object is a file, and
that failure must come back as the write failure rather than as the error
of the read that would compare content.
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* refactor(log-store): rename the catalog's out_of_order helper to out_of_order_reason
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
---------
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* feat(log-store): add the object store WAL object format
Add the byte format of a single object store WAL object: a header with
the GTWALOBJ magic, format version 1, the object sequence and the writer
instance id; one segment per region ordered by region id with entries
ordered by entry id; a footer that records each segment's region id,
entry id range, entry count, byte range and CRC32; and a fixed trailer
with the GTWALTRL magic, the footer location, the footer CRC32 and the
whole-object CRC32.
The module encodes objects deterministically and decodes the header,
trailer, footer and segments separately, with structural checks on
footer ranges and segment tiling. It has no callers yet; the store that
writes and reads objects follows in later changes.
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* test(log-store): pin the object store WAL format with a byte fixture
Add a fixed version 1 object with two regions and five entries as a hex
literal. The test decodes it and checks the exact header, trailer,
footer entries and records, and checks that encoding the same records,
in either input order, reproduces the fixture byte for byte.
Round-trip tests alone pass when a refactor changes field order,
endianness or checksum coverage in both the encoder and the decoder.
The fixture bytes were derived from the documented layout rather than
from the encoder, so such a change now fails.
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* fix(log-store): reject an empty footer when decoding a WAL object
The encoder never writes an object without records, but decode_footer
accepted a footer that declares zero segments, and
verify_segment_ranges accepts an empty footer too. Only the test-only
decode_object rejected it, so a checksum-valid empty object would pass
the header, trailer and footer checks that recovery runs.
Reject a zero entry count in decode_footer and drop the now unreachable
check in decode_object. Add a test that builds a checksum-valid object
with an empty footer and checks that decode_footer and decode_object
reject it.
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* refactor(log-store): use pub(crate) for the WAL object format API
Other log-store modules use pub(crate) for items shared across module
boundaries. Switch the format module from pub(super) to pub(crate) to
follow that convention. No behavior change.
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
---------
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* fix: flush soft-dropped regions on close
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* feat(mito2): handle flush-on-close race with concurrent in-flight flush
When a region close with `flush_on_close: true` races with an
already-running flush, pass the actual close request (including the
flush_on_close flag) to the DDL handler instead of a default request
so the pending flush is correctly awaited.
Files: `src/mito2/src/worker/handle_close.rs`
Also adds a test verifying that closing with flush-on-close while a
flush is in progress still persists all written data correctly.
Files: `src/mito2/src/engine/close_test.rs`
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* feat: support full WAL retirement
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* fix: complete close request migration
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* fix: finish close request callsites
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* refactor: guard Kafka provider setup behind index collector check
Move Kafka provider initialization and `get_or_insert` inside the
existing `if let Some(collector)` block so these operations are
skipped when no global index collector is configured.
Affected file:
- `src/log-store/src/kafka/log_store.rs`
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* chore: avoid to_vec
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* refactor: replace imperative close-region loop with functional combinators
Transform the region close dispatch in `DropTableExecutor` from mutable
`Vec` and `push` loops to iterator chains with `join_all`, improving
idiomatic Rust style and readability.
- `src/common/meta/src/ddl/drop_table/executor.rs` — rewired datanode
region-close logic to use `peers.map()` and nested `join_all`, moving
`node_manager.datanode()` inside the closure to align with the new
structure
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* fix: decouple Kafka client from WAL checkpoint
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* fix: merge Kafka WAL index checkpoints
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* fix: delegate Kafka WAL retirement to metasrv
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* chore: rebase main and resolve conflicts
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* fix: license header
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* chore: bump proto to commits on main
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* refactor: remove Kafka obsolete-all index changes
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
---------
Signed-off-by: Lei, HUANG <ratuthomm@gmail.com>
* fix(mito2): handle corner case in catchup where compacted entry id exceeds region last entry id
Signed-off-by: WenyXu <wenymedia@gmail.com>
* chore: apply suggestions from CR
Signed-off-by: WenyXu <wenymedia@gmail.com>
---------
Signed-off-by: WenyXu <wenymedia@gmail.com>
* feat/bulk-wal:
### Refactor: Simplify Data Handling in LogStore Implementations
- **`kafka/log_store.rs`, `raft_engine/log_store.rs`, `wal.rs`, `raw_entry_reader.rs`, `logstore.rs`:**
- Refactored `entry` and `build_entry` functions to accept `Vec<u8>` directly instead of `&mut Vec<u8>`.
- Removed usage of `std::mem::take` for data handling, simplifying the code and improving readability.
- Updated test cases to align with the new function signatures.
* feat/bulk-wal:
### Add Support for Bulk WAL Entries and Flight Data Encoding
- **Add `raw_data` field to `BulkPart` and related structs**: Updated `BulkPart` and related structures in `src/mito2/src/memtable/bulk/part.rs`, `src/mito2/src/memtable/simple_bulk_memtable.rs`, `src/mito2/src/memtable/time_partition.rs`, `src/mito2/src/region_write_ctx.rs`,
`src/mito2/src/worker/handle_bulk_insert.rs`, and `src/store-api/src/region_request.rs` to include a new `raw_data` field for handling Arrow IPC data.
- **Implement Flight Data Encoding**: Added a new module `flight` in `src/common/test-util/src/flight.rs` to encode record batches to Flight data format.
- **Update `greptime-proto` dependency**: Changed the revision of the `greptime-proto` dependency in `Cargo.lock` and `Cargo.toml`.
- **Enhance WAL Writer and Tests**: Modified `src/mito2/src/wal.rs` and related test files to support bulk WAL entries and added tests for encoding and handling bulk data.
* feat/bulk-wal:
- **Update `greptime-proto` Dependency**: Updated the `greptime-proto` dependency to a new revision in `Cargo.lock` and `Cargo.toml`.
- **Add `common-grpc` Dependency**: Added `common-grpc` as a dependency in `Cargo.lock` and `src/mito2/Cargo.toml`.
- **Refactor `BulkPart` Structure**: Removed `num_rows` field and added `num_rows()` method in `src/mito2/src/memtable/bulk/part.rs`. Updated related usages in `src/mito2/src/memtable/simple_bulk_memtable.rs`, `src/mito2/src/memtable/time_partition.rs`, `src/mito2/src/memtable/time_series.rs`,
`src/mito2/src/region_write_ctx.rs`, and `src/mito2/src/worker/handle_bulk_insert.rs`.
- **Implement `TryFrom` and `From` for `BulkWalEntry`**: Added implementations for converting between `BulkPart` and `BulkWalEntry` in `src/mito2/src/memtable/bulk/part.rs`.
- **Handle Bulk Entries in Region Opener**: Added logic to process bulk entries in `src/mito2/src/region/opener.rs`.
- **Fix `BulkInsertRequest` Handling**: Corrected `region_id` handling in `src/operator/src/bulk_insert.rs` and `src/store-api/src/region_request.rs`.
- **Add Error Variant for `ConvertBulkWalEntry`**: Added a new error variant in `src/mito2/src/error.rs` for handling bulk WAL entry conversion errors.
* fix: ci
* feat/bulk-wal:
Add bulk write operation in `opener.rs`
- Enhanced the region write context by adding a call to `write_bulk()` after `write_memtable()` in `opener.rs`.
- This change aims to improve the efficiency of writing operations by enabling bulk writes.
* feat/bulk-wal:
Enhance error handling and metrics in `bulk_insert.rs`
- Updated `Inserter` to improve error handling by capturing the result of `datanode.handle(request)` and incrementing the `DIST_INGEST_ROW_COUNT` metric with the number of affected rows.
* feat/bulk-wal:
### Remove Encode Error Handling for WAL Entries
- **`error.rs`**: Removed the `EncodeWal` error variant and its associated handling.
- **`wal.rs`**: Eliminated the `entry_encode_buf` buffer and its usage for encoding WAL entries. Replaced with direct encoding to a vector using `encode_to_vec()`.
* feat: improve topic management and add stale records cleanup
* fix: fix unit tests
* chore: apply suggestions from CR
* chore: apply suggestions from CR
chore/move-wal-sync-to-bg:
### Refactor Log Store Task Management
- **Error Handling Enhancements**: Updated error handling for task management in `error.rs` by renaming `StartGcTask` and `StopGcTask` to `StartWalTask` and `StopWalTask`, respectively, and added a `name` field for more descriptive error messages.
- **Task Management Improvements**: Introduced `SyncWalTaskFunction` in `log_store.rs` to handle periodic synchronization of WAL tasks, replacing the previous atomic-based sync logic.
- **Backend Adjustments**: Modified `backend.rs` to use the new `StartWalTaskSnafu` for starting tasks, ensuring consistency with the updated error handling approach.
* feat: set max log files to 720 by default, info log only
* expose max_log_files in tomls
* include dir info when panicing, limit max_log_files of err_log to 30, and that of slow_queries to opt.max_log_files
* fix clippy
* update config.md
* update expected config str
* limit err_log max files size to `max_log_files` too, include err info when panicing, put `max_l_f` in right position
* fix typos
* chore: config
Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>
---------
Co-authored-by: dennis zhuang <killme2008@gmail.com>
Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>
* chore: cherrypick 52e8eebb2dbbbe81179583c05094004a5eedd7fd
* refactor/tables: Change variable from immutable to mutable in KvBackendCatalogManager's method
* refactor/tables: Replace unbounded channel with bounded and use semaphore for concurrency control in KvBackendCatalogManager
* refactor/tables: Add common-runtime dependency and update KvBackendCatalogManager to use common_runtime::spawn_global
* refactor/tables: Await on sending error through channel in KvBackendCatalogManager
* Refactor RaftEngineLogStore to use references for config
- Updated `RaftEngineLogStore::try_new` to accept a reference to `RaftEngineConfig` instead of taking ownership.
- Replaced direct usage of `config` with individual fields (`sync_write`, `sync_period`, `read_batch_size`).
- Adjusted test cases to pass references to `RaftEngineConfig`.
* Add parallelism configuration for WAL recovery
- Introduced `recovery_parallelism` setting in `datanode.example.toml` and `standalone.example.toml` for configuring parallelism during WAL recovery.
- Updated `Cargo.lock` and `Cargo.toml` to include `num_cpus` dependency.
- Modified `RaftEngineConfig` to include `recovery_parallelism` with a default value set to the number of CP
* feat/wal-recovery-parallelism:
Add `wal.recovery_parallelism` configuration option
- Introduced `wal.recovery_parallelism` to config.md for specifying parallelism during WAL recovery.
- Updated `RaftEngineLogStore` to include `recovery_threads` from the new configuration.
* fix: ut
* feat(log_store): use new `Consumer`
* feat: add `from_peer_id`
* feat: read WAL entries respect index
* test: add test for `build_region_wal_index_iterator`
* fix: keep the handle
* fix: incorrect last index
* fix: replay last entry id may be greater than expected
* chore: remove unused code
* chore: apply suggestions from CR
* chore: rename `datanode_id` to `location_id`
* chore: rename `from_peer_id` to `location_id`
* chore: rename `from_peer_id` to `location_id`
* chore: apply suggestions from CR
jeremyhi
XuanwoandWenyXu
shuiyisongandfys
Lei, HUANG
Ning Sun
LFCandNing Sun
Ruihang Xia
Yuhan Wang
JohnsonLeeandYingwen