fix: remove redundancy length field in wal record

.asf.yaml

@@ -39,7 +39,7 @@ github:
   protected_branches:
     main:
       required_pull_request_reviews:
-        dismiss_stale_reviews: true
+        dismiss_stale_reviews: false
         required_approving_review_count: 1
   protected_tags: []
 

.github/workflows/ci.yml

@@ -39,6 +39,10 @@ on:
       - 'Cargo.lock'
       - '.github/workflows/ci.yml'
 
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
 # Common environment variables
 env:
   RUSTFLAGS: "-C debuginfo=1"

.github/workflows/tsbs.yml

@@ -46,7 +46,7 @@ jobs:
       - name: Setup Build Environment
         run: |
           sudo apt update
-          sudo apt install --yes protobuf-compiler
+          sudo apt install --yes protobuf-compiler liblzma-dev
       - name: Build server
         run: |
           make build

Cargo.toml

@@ -84,6 +84,8 @@ members = [
     "src/wal"
 ]
 
+default-members = ["src/horaedb"]
+
 [workspace.dependencies]
 alloc_tracker = { path = "src/components/alloc_tracker" }
 arrow = { version = "49.0.0", features = ["prettyprint"] }

docs/example-cluster-0.toml

@@ -37,7 +37,7 @@ type = "Local"
 data_dir = "/tmp/horaedb0"
 
 [analytic.wal]
-type = "RocksDB"
+type = "Local"
 data_dir = "/tmp/horaedb0"
 
 [cluster_deployment]

docs/example-cluster-1.toml

@@ -38,7 +38,7 @@ type = "Local"
 data_dir = "/tmp/horaedb1"
 
 [analytic.wal]
-type = "RocksDB"
+type = "Local"
 data_dir = "/tmp/horaedb1"
 
 [cluster_deployment]

docs/example-standalone-static-routing.toml

@@ -36,7 +36,7 @@ max_replay_tables_per_batch = 1024
 write_group_command_channel_cap = 1024
 
 [analytic.wal]
-type = "RocksDB"
+type = "Local"
 data_dir = "/tmp/horaedb1"
 
 [analytic.storage]
@@ -91,4 +91,3 @@ shards = [ 1 ]
 [limiter]
 write_block_list = ['mytable1']
 read_block_list = ['mytable1']
-

docs/minimal.toml

@@ -32,7 +32,7 @@ type = "Local"
 data_dir = "/tmp/horaedb"
 
 [analytic.wal]
-type = "RocksDB"
+type = "Local"
 data_dir = "/tmp/horaedb"
 
 [analytic]

integration_tests/Makefile

@@ -54,7 +54,7 @@ build-meta:
 	./build_meta.sh
 
 build-horaedb:
-	cd .. && cargo build --bin horaedb-server --features wal-table-kv,wal-message-queue,wal-rocksdb,wal-local-storage
+	cd .. && make build-debug
 
 build-test:
 	cargo build

integration_tests/cases/env/cluster/ddl/partition_table.result

@@ -100,21 +100,23 @@ UInt64(16367588166920223437),Timestamp(1651737067000),String("horaedb9"),Int32(0
 -- SQLNESS REPLACE compute=\d+.?\d*(µ|m|n) compute=xx
 -- SQLNESS REPLACE time=\d+.?\d*(µ|m|n) time=xx
 -- SQLNESS REPLACE metrics=\[.*?s\] metrics=xx
+-- SQLNESS REPLACE scan_memtable_\d+ scan_memtable_n
 EXPLAIN ANALYZE SELECT * from partition_table_t where name = "ceresdb0";
 
 plan_type,plan,
-String("Plan with Metrics"),String("ResolvedPartitionedScan: pushdown_continue:false, partition_count:1, metrics=xx\n  ScanTable: table=__partition_table_t_1, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[name = Utf8(\"ceresdb0\")], time_range:TimeRange { inclusive_start: Timestamp(-9223372036854775808), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=0\n        num_ssts=0\n        scan_count=1\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=0\n        total_rows_fetch_from_one=0\n        scan_memtable_1, fetched_columns:[tsid,t,name,id,value]:\n=0]\n=0]\n"),
+String("Plan with Metrics"),String("ResolvedPartitionedScan: pushdown_continue:false, partition_count:1, metrics=xx\n  ScanTable: table=__partition_table_t_1, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[name = Utf8(\"ceresdb0\")], time_range:TimeRange { inclusive_start: Timestamp(-9223372036854775808), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=0\n        num_ssts=0\n        scan_count=1\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=0\n        total_rows_fetch_from_one=0\n        scan_memtable_n, fetched_columns:[tsid,t,name,id,value]:\n=0]\n=0]\n"),
 
 
 -- SQLNESS REPLACE duration=\d+.?\d*(µ|m|n) duration=xx
 -- SQLNESS REPLACE compute=\d+.?\d*(µ|m|n) compute=xx
 -- SQLNESS REPLACE __partition_table_t_\d __partition_table_t_x
 -- SQLNESS REPLACE time=\d+.?\d*(µ|m|n) time=xx
 -- SQLNESS REPLACE metrics=\[.*?s\] metrics=xx
+-- SQLNESS REPLACE scan_memtable_\d+ scan_memtable_n
 EXPLAIN ANALYZE SELECT * from partition_table_t where name in ("ceresdb0", "ceresdb1", "ceresdb2", "ceresdb3", "ceresdb4");
 
 plan_type,plan,
-String("Plan with Metrics"),String("ResolvedPartitionedScan: pushdown_continue:false, partition_count:3, metrics=xx\n  ScanTable: table=__partition_table_t_x, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=xx\n  ScanTable: table=__partition_table_t_x, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=xx\n  ScanTable: table=__partition_table_t_x, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[name IN ([Utf8(\"ceresdb0\"), Utf8(\"ceresdb1\"), Utf8(\"ceresdb2\"), Utf8(\"ceresdb3\"), Utf8(\"ceresdb4\")])], time_range:TimeRange { inclusive_start: Timestamp(-9223372036854775808), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=0\n        num_ssts=0\n        scan_count=1\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=0\n        total_rows_fetch_from_one=0\n        scan_memtable_1, fetched_columns:[tsid,t,name,id,value]:\n=0]\n=0]\n"),
+String("Plan with Metrics"),String("ResolvedPartitionedScan: pushdown_continue:false, partition_count:3, metrics=xx\n  ScanTable: table=__partition_table_t_x, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=xx\n  ScanTable: table=__partition_table_t_x, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=xx\n  ScanTable: table=__partition_table_t_x, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[name IN ([Utf8(\"ceresdb0\"), Utf8(\"ceresdb1\"), Utf8(\"ceresdb2\"), Utf8(\"ceresdb3\"), Utf8(\"ceresdb4\")])], time_range:TimeRange { inclusive_start: Timestamp(-9223372036854775808), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=0\n        num_ssts=0\n        scan_count=1\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=0\n        total_rows_fetch_from_one=0\n        scan_memtable_n, fetched_columns:[tsid,t,name,id,value]:\n=0]\n=0]\n"),
 
 
 ALTER TABLE partition_table_t ADD COLUMN (b string);

integration_tests/cases/env/cluster/ddl/partition_table.sql

@@ -58,13 +58,15 @@ SELECT * from partition_table_t where name in ("horaedb5", "horaedb6", "horaedb7
 -- SQLNESS REPLACE compute=\d+.?\d*(µ|m|n) compute=xx
 -- SQLNESS REPLACE time=\d+.?\d*(µ|m|n) time=xx
 -- SQLNESS REPLACE metrics=\[.*?s\] metrics=xx
+-- SQLNESS REPLACE scan_memtable_\d+ scan_memtable_n
 EXPLAIN ANALYZE SELECT * from partition_table_t where name = "ceresdb0";
 
 -- SQLNESS REPLACE duration=\d+.?\d*(µ|m|n) duration=xx
 -- SQLNESS REPLACE compute=\d+.?\d*(µ|m|n) compute=xx
 -- SQLNESS REPLACE __partition_table_t_\d __partition_table_t_x
 -- SQLNESS REPLACE time=\d+.?\d*(µ|m|n) time=xx
 -- SQLNESS REPLACE metrics=\[.*?s\] metrics=xx
+-- SQLNESS REPLACE scan_memtable_\d+ scan_memtable_n
 EXPLAIN ANALYZE SELECT * from partition_table_t where name in ("ceresdb0", "ceresdb1", "ceresdb2", "ceresdb3", "ceresdb4");
 
 ALTER TABLE partition_table_t ADD COLUMN (b string);

integration_tests/cases/env/local/ddl/query-plan.result

@@ -50,7 +50,7 @@ explain analyze select t from `03_dml_select_real_time_range`
 where t > 1695348001000;
 
 plan_type,plan,
-String("Plan with Metrics"),String("ScanTable: table=03_dml_select_real_time_range, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t > TimestampMillisecond(1695348001000, None)], time_range:TimeRange { inclusive_start: Timestamp(1695348001001), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=1\n        num_ssts=0\n        scan_count=2\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=1\n        total_rows_fetch_from_one=1\n        scan_memtable_1, fetched_columns:[tsid,t]:\n=0]\n"),
+String("Plan with Metrics"),String("ScanTable: table=03_dml_select_real_time_range, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t > TimestampMillisecond(1695348001000, None)], time_range:TimeRange { inclusive_start: Timestamp(1695348001001), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=1\n        num_ssts=0\n        scan_count=2\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=1\n        total_rows_fetch_from_one=1\n        scan_memtable_164, fetched_columns:[tsid,t]:\n=0]\n"),
 
 
 -- This query should have higher priority
@@ -60,7 +60,7 @@ explain analyze select t from `03_dml_select_real_time_range`
 where t >= 1695348001000 and t < 1695348002000;
 
 plan_type,plan,
-String("Plan with Metrics"),String("ScanTable: table=03_dml_select_real_time_range, parallelism=8, priority=High, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t >= TimestampMillisecond(1695348001000, None), t < TimestampMillisecond(1695348002000, None)], time_range:TimeRange { inclusive_start: Timestamp(1695348001000), exclusive_end: Timestamp(1695348002000) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=1\n        num_ssts=0\n        scan_count=2\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=1\n        total_rows_fetch_from_one=1\n        scan_memtable_1, fetched_columns:[tsid,t]:\n=0]\n"),
+String("Plan with Metrics"),String("ScanTable: table=03_dml_select_real_time_range, parallelism=8, priority=High, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t >= TimestampMillisecond(1695348001000, None), t < TimestampMillisecond(1695348002000, None)], time_range:TimeRange { inclusive_start: Timestamp(1695348001000), exclusive_end: Timestamp(1695348002000) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=1\n        num_ssts=0\n        scan_count=2\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=1\n        total_rows_fetch_from_one=1\n        scan_memtable_164, fetched_columns:[tsid,t]:\n=0]\n"),
 
 
 -- This query should have higher priority
@@ -70,7 +70,7 @@ explain analyze select name from `03_dml_select_real_time_range`
 where t >= 1695348001000 and t < 1695348002000;
 
 plan_type,plan,
-String("Plan with Metrics"),String("ProjectionExec: expr=[name@0 as name], metrics=xx\n  ScanTable: table=03_dml_select_real_time_range, parallelism=8, priority=High, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t >= TimestampMillisecond(1695348001000, None), t < TimestampMillisecond(1695348002000, None)], time_range:TimeRange { inclusive_start: Timestamp(1695348001000), exclusive_end: Timestamp(1695348002000) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=1\n        num_ssts=0\n        scan_count=2\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=1\n        total_rows_fetch_from_one=1\n        scan_memtable_1, fetched_columns:[tsid,t,name]:\n=0]\n"),
+String("Plan with Metrics"),String("ProjectionExec: expr=[name@0 as name], metrics=xx\n  ScanTable: table=03_dml_select_real_time_range, parallelism=8, priority=High, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t >= TimestampMillisecond(1695348001000, None), t < TimestampMillisecond(1695348002000, None)], time_range:TimeRange { inclusive_start: Timestamp(1695348001000), exclusive_end: Timestamp(1695348002000) } }\nscan_table:\n    do_merge_sort=true\n    iter_num=1\n    merge_iter_0:\n        init_duration=xxs\n        num_memtables=1\n        num_ssts=0\n        scan_count=2\n        scan_duration=xxs\n        times_fetch_row_from_multiple=0\n        times_fetch_rows_from_one=1\n        total_rows_fetch_from_one=1\n        scan_memtable_164, fetched_columns:[tsid,t,name]:\n=0]\n"),
 
 
 -- This query should not include memtable
@@ -135,7 +135,7 @@ explain analyze select t from `03_append_mode_table`
 where t >= 1695348001000 and name = 'ceresdb';
 
 plan_type,plan,
-String("Plan with Metrics"),String("ProjectionExec: expr=[t@0 as t], metrics=xx\n  ScanTable: table=03_append_mode_table, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t >= TimestampMillisecond(1695348001000, None), name = Utf8(\"ceresdb\")], time_range:TimeRange { inclusive_start: Timestamp(1695348001000), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=false\n    chain_iter_0:\n        num_memtables=1\n        num_ssts=0\n        scan_duration=xxs\n        since_create=xxs\n        since_init=xxs\n        total_batch_fetched=1\n        total_rows_fetched=2\n        scan_memtable_1, fetched_columns:[t,name]:\n=0]\n"),
+String("Plan with Metrics"),String("ProjectionExec: expr=[t@0 as t], metrics=xx\n  ScanTable: table=03_append_mode_table, parallelism=8, priority=Low, partition_count=UnknownPartitioning(8), metrics=[\nPredicate { exprs:[t >= TimestampMillisecond(1695348001000, None), name = Utf8(\"ceresdb\")], time_range:TimeRange { inclusive_start: Timestamp(1695348001000), exclusive_end: Timestamp(9223372036854775807) } }\nscan_table:\n    do_merge_sort=false\n    chain_iter_0:\n        num_memtables=1\n        num_ssts=0\n        scan_duration=xxs\n        since_create=xxs\n        since_init=xxs\n        total_batch_fetched=1\n        total_rows_fetched=2\n        scan_memtable_166, fetched_columns:[t,name]:\n=0]\n"),
 
 
 -- Should just fetch projected columns from SST

integration_tests/config/horaedb-cluster-0.toml

@@ -37,7 +37,7 @@ type = "Local"
 data_dir = "/tmp/horaedb0"
 
 [analytic.wal]
-type = "RocksDB"
+type = "Local"
 data_dir = "/tmp/horaedb0"
 
 [cluster_deployment]

integration_tests/config/horaedb-cluster-1.toml

@@ -38,7 +38,7 @@ type = "Local"
 data_dir = "/tmp/horaedb1"
 
 [analytic.wal]
-type = "RocksDB"
+type = "Local"
 data_dir = "/tmp/horaedb1"
 
 [cluster_deployment]

integration_tests/config/shard-based-recovery.toml

@@ -34,5 +34,5 @@ type = "Local"
 data_dir = "/tmp/horaedb"
 
 [analytic.wal]
-type = "RocksDB"
+type = "Local"
 data_dir = "/tmp/horaedb"

src/horaedb/Cargo.toml

@@ -31,7 +31,7 @@ workspace = true
 workspace = true
 
 [features]
-default = ["wal-rocksdb", "wal-table-kv", "wal-message-queue", "wal-local-storage"]
+default = ["wal-table-kv", "wal-message-queue", "wal-local-storage"]
 wal-table-kv = ["wal/wal-table-kv", "analytic_engine/wal-table-kv"]
 wal-message-queue = ["wal/wal-message-queue", "analytic_engine/wal-message-queue"]
 wal-rocksdb = ["wal/wal-rocksdb", "analytic_engine/wal-rocksdb"]

src/wal/Cargo.toml

@@ -32,7 +32,7 @@ workspace = true
 
 [dependencies.rocksdb]
 git = "https://github.com/tikv/rust-rocksdb.git"
-rev = "f04f4dd8eacc30e67c24bc2529a6d9c6edb85f8f"
+rev = "85e79e52c6ad80b8c547fcb90b3cade64f141fac"
 features = ["portable"]
 optional = true