feat(standalone): support to dump/restore metadata

* feat: enable submitting wal prune procedure periodically

* chore: fix and add options

* test: add unit test

* test: fix unit test

* test: enable active_wal_pruning in test

* test: update default config

* chore: update config name

* refactor: use semaphore to control the number of prune process

* refactor: use split client for wal prune manager and topic creator

* chore: add configs

* chore: apply review comments

* fix: use tracker properly

* fix: use guard to track semaphore

* test: update unit tests

* chore: update config name

* chore: use prunable_entry_id

* refactor: semaphore to only limit the process of submitting

* chore: remove legacy sort

* chore: better configs

* fix: update config.md

* chore: respect fmt

* test: update unit tests

* chore: use interval_at

* fix: fix unit test

* test: fix unit test

* test: fix unit test

* chore: apply review comments

* docs: update config docs

.github/actions/build-windows-artifacts/action.yml

@@ -47,7 +47,6 @@ runs:
       shell: pwsh
       run: make test sqlness-test
       env:
-        RUSTUP_WINDOWS_PATH_ADD_BIN: 1 # Workaround for https://github.com/nextest-rs/nextest/issues/1493
         RUST_BACKTRACE: 1
         SQLNESS_OPTS: "--preserve-state"
 

.github/actions/setup-greptimedb-cluster/with-remote-wal.yaml

@@ -2,13 +2,13 @@ meta:
   configData: |-
     [runtime]
     global_rt_size = 4
-    
+
     [wal]
     provider = "kafka"
     broker_endpoints = ["kafka.kafka-cluster.svc.cluster.local:9092"]
     num_topics = 3
+    auto_prune_topic_records = true
 
-        
     [datanode]
     [datanode.client]
     timeout = "120s"

.github/scripts/create-version.sh

@@ -25,7 +25,7 @@ function create_version() {
   fi
 
   # Reuse $NEXT_RELEASE_VERSION to identify whether it's a nightly build.
-  # It will be like 'nigtly-20230808-7d0d8dc6'.
+  # It will be like 'nightly-20230808-7d0d8dc6'.
   if [ "$NEXT_RELEASE_VERSION" = nightly ]; then
     echo "$NIGHTLY_RELEASE_PREFIX-$(date "+%Y%m%d")-$(git rev-parse --short HEAD)"
     exit 0
@@ -60,9 +60,9 @@ function create_version() {
 }
 
 # You can run as following examples:
-#  GITHUB_EVENT_NAME=push NEXT_RELEASE_VERSION=v0.4.0 NIGHTLY_RELEASE_PREFIX=nigtly GITHUB_REF_NAME=v0.3.0 ./create-version.sh
-#  GITHUB_EVENT_NAME=workflow_dispatch NEXT_RELEASE_VERSION=v0.4.0 NIGHTLY_RELEASE_PREFIX=nigtly ./create-version.sh
-#  GITHUB_EVENT_NAME=schedule NEXT_RELEASE_VERSION=v0.4.0 NIGHTLY_RELEASE_PREFIX=nigtly ./create-version.sh
-#  GITHUB_EVENT_NAME=schedule NEXT_RELEASE_VERSION=nightly NIGHTLY_RELEASE_PREFIX=nigtly ./create-version.sh
-#  GITHUB_EVENT_NAME=workflow_dispatch COMMIT_SHA=f0e7216c4bb6acce9b29a21ec2d683be2e3f984a NEXT_RELEASE_VERSION=dev NIGHTLY_RELEASE_PREFIX=nigtly ./create-version.sh
+#  GITHUB_EVENT_NAME=push NEXT_RELEASE_VERSION=v0.4.0 NIGHTLY_RELEASE_PREFIX=nightly GITHUB_REF_NAME=v0.3.0 ./create-version.sh
+#  GITHUB_EVENT_NAME=workflow_dispatch NEXT_RELEASE_VERSION=v0.4.0 NIGHTLY_RELEASE_PREFIX=nightly ./create-version.sh
+#  GITHUB_EVENT_NAME=schedule NEXT_RELEASE_VERSION=v0.4.0 NIGHTLY_RELEASE_PREFIX=nightly ./create-version.sh
+#  GITHUB_EVENT_NAME=schedule NEXT_RELEASE_VERSION=nightly NIGHTLY_RELEASE_PREFIX=nightly ./create-version.sh
+#  GITHUB_EVENT_NAME=workflow_dispatch COMMIT_SHA=f0e7216c4bb6acce9b29a21ec2d683be2e3f984a NEXT_RELEASE_VERSION=dev NIGHTLY_RELEASE_PREFIX=nightly ./create-version.sh
 create_version

.github/workflows/nightly-ci.yml

@@ -107,7 +107,6 @@ jobs:
           CARGO_BUILD_RUSTFLAGS: "-C linker=lld-link"
           RUST_BACKTRACE: 1
           CARGO_INCREMENTAL: 0
-          RUSTUP_WINDOWS_PATH_ADD_BIN: 1 # Workaround for https://github.com/nextest-rs/nextest/issues/1493
           GT_S3_BUCKET: ${{ vars.AWS_CI_TEST_BUCKET }}
           GT_S3_ACCESS_KEY_ID: ${{ secrets.AWS_CI_TEST_ACCESS_KEY_ID }}
           GT_S3_ACCESS_KEY: ${{ secrets.AWS_CI_TEST_SECRET_ACCESS_KEY }}

.github/workflows/release.yml

@@ -317,7 +317,7 @@ jobs:
           image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
           image-registry-password: ${{ secrets.DOCKERHUB_TOKEN }}
           version: ${{ needs.allocate-runners.outputs.version }}
-          push-latest-tag: true
+          push-latest-tag: ${{ github.ref_type == 'tag' && !contains(github.ref_name, 'nightly') && github.event_name != 'schedule' }}
 
       - name: Set build image result
         id: set-build-image-result
@@ -364,7 +364,7 @@ jobs:
           dev-mode: false
           upload-to-s3: true
           update-version-info: true
-          push-latest-tag: true
+          push-latest-tag: ${{ github.ref_type == 'tag' && !contains(github.ref_name, 'nightly') && github.event_name != 'schedule' }}
 
   publish-github-release:
     name: Create GitHub release and upload artifacts

Cargo.toml

@@ -29,6 +29,7 @@ members = [
     "src/common/query",
     "src/common/recordbatch",
     "src/common/runtime",
+    "src/common/session",
     "src/common/substrait",
     "src/common/telemetry",
     "src/common/test-util",
@@ -76,7 +77,6 @@ clippy.print_stdout = "warn"
 clippy.print_stderr = "warn"
 clippy.dbg_macro = "warn"
 clippy.implicit_clone = "warn"
-clippy.readonly_write_lock = "allow"
 rust.unknown_lints = "deny"
 rust.unexpected_cfgs = { level = "warn", check-cfg = ['cfg(tokio_unstable)'] }
 
@@ -88,20 +88,20 @@ rust.unexpected_cfgs = { level = "warn", check-cfg = ['cfg(tokio_unstable)'] }
 #
 # See for more detaiils: https://github.com/rust-lang/cargo/issues/11329
 ahash = { version = "0.8", features = ["compile-time-rng"] }
-aquamarine = "0.3"
-arrow = { version = "53.0.0", features = ["prettyprint"] }
-arrow-array = { version = "53.0.0", default-features = false, features = ["chrono-tz"] }
-arrow-flight = "53.0"
-arrow-ipc = { version = "53.0.0", default-features = false, features = ["lz4", "zstd"] }
-arrow-schema = { version = "53.0", features = ["serde"] }
+aquamarine = "0.6"
+arrow = { version = "54.2", features = ["prettyprint"] }
+arrow-array = { version = "54.2", default-features = false, features = ["chrono-tz"] }
+arrow-flight = "54.2"
+arrow-ipc = { version = "54.2", default-features = false, features = ["lz4", "zstd"] }
+arrow-schema = { version = "54.2", features = ["serde"] }
 async-stream = "0.3"
 async-trait = "0.1"
 # Remember to update axum-extra, axum-macros when updating axum
 axum = "0.8"
 axum-extra = "0.10"
-axum-macros = "0.4"
+axum-macros = "0.5"
 backon = "1"
-base64 = "0.21"
+base64 = "0.22"
 bigdecimal = "0.4.2"
 bitflags = "2.4.1"
 bytemuck = "1.12"
@@ -111,42 +111,43 @@ chrono-tz = "0.10.1"
 clap = { version = "4.4", features = ["derive"] }
 config = "0.13.0"
 crossbeam-utils = "0.8"
-dashmap = "5.4"
-datafusion = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-common = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-expr = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-functions = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-optimizer = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-physical-expr = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-physical-plan = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-sql = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-datafusion-substrait = { git = "https://github.com/apache/datafusion.git", rev = "2464703c84c400a09cc59277018813f0e797bb4e" }
-deadpool = "0.10"
-deadpool-postgres = "0.12"
-derive_builder = "0.12"
+dashmap = "6.1"
+datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-functions = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-physical-plan = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-sql = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+datafusion-substrait = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "5bbedc6704162afb03478f56ffb629405a4e1220" }
+deadpool = "0.12"
+deadpool-postgres = "0.14"
+derive_builder = "0.20"
 dotenv = "0.15"
 etcd-client = "0.14"
 fst = "0.4.7"
 futures = "0.3"
 futures-util = "0.3"
-greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "a7274ddce299f33d23dbe8af5bbe6219f07c559a" }
+greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "b6d9cffd43c4e6358805a798f17e03e232994b82" }
 hex = "0.4"
 http = "1"
 humantime = "2.1"
 humantime-serde = "1.1"
 hyper = "1.1"
 hyper-util = "0.1"
-itertools = "0.10"
+itertools = "0.14"
 jsonb = { git = "https://github.com/databendlabs/jsonb.git", rev = "8c8d2fc294a39f3ff08909d60f718639cfba3875", default-features = false }
 lazy_static = "1.4"
 local-ip-address = "0.6"
 loki-proto = { git = "https://github.com/GreptimeTeam/loki-proto.git", rev = "1434ecf23a2654025d86188fb5205e7a74b225d3" }
 meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "5618e779cf2bb4755b499c630fba4c35e91898cb" }
-mockall = "0.11.4"
+mockall = "0.13"
 moka = "0.12"
 nalgebra = "0.33"
-notify = "6.1"
+notify = "8.0"
 num_cpus = "1.16"
+object_store_opendal = "0.50"
 once_cell = "1.18"
 opentelemetry-proto = { version = "0.27", features = [
     "gen-tonic",
@@ -156,15 +157,15 @@ opentelemetry-proto = { version = "0.27", features = [
     "logs",
 ] }
 parking_lot = "0.12"
-parquet = { version = "53.0.0", default-features = false, features = ["arrow", "async", "object_store"] }
+parquet = { version = "54.2", default-features = false, features = ["arrow", "async", "object_store"] }
 paste = "1.0"
 pin-project = "1.0"
 prometheus = { version = "0.13.3", features = ["process"] }
-promql-parser = { version = "0.5", features = ["ser"] }
+promql-parser = { version = "0.5.1", features = ["ser"] }
 prost = "0.13"
 raft-engine = { version = "0.4.1", default-features = false }
-rand = "0.8"
-ratelimit = "0.9"
+rand = "0.9"
+ratelimit = "0.10"
 regex = "1.8"
 regex-automata = "0.4"
 reqwest = { version = "0.12", default-features = false, features = [
@@ -176,35 +177,36 @@ reqwest = { version = "0.12", default-features = false, features = [
 rskafka = { git = "https://github.com/influxdata/rskafka.git", rev = "75535b5ad9bae4a5dbb582c82e44dfd81ec10105", features = [
     "transport-tls",
 ] }
-rstest = "0.21"
+rstest = "0.25"
 rstest_reuse = "0.7"
 rust_decimal = "1.33"
 rustc-hash = "2.0"
-rustls = { version = "0.23.20", default-features = false } # override by patch, see [patch.crates-io]
+# It is worth noting that we should try to avoid using aws-lc-rs until it can be compiled on various platforms.
+rustls = { version = "0.23.25", default-features = false }
 serde = { version = "1.0", features = ["derive"] }
 serde_json = { version = "1.0", features = ["float_roundtrip"] }
 serde_with = "3"
-shadow-rs = "0.38"
+shadow-rs = "1.1"
+simd-json = "0.15"
 similar-asserts = "1.6.0"
 smallvec = { version = "1", features = ["serde"] }
 snafu = "0.8"
+sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "e98e6b322426a9d397a71efef17075966223c089", features = [
+    "visitor",
+    "serde",
+] } # branch = "v0.54.x"
 sqlx = { version = "0.8", features = [
     "runtime-tokio-rustls",
     "mysql",
     "postgres",
     "chrono",
 ] }
-sysinfo = "0.30"
-# on branch v0.52.x
-sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "71dd86058d2af97b9925093d40c4e03360403170", features = [
-    "visitor",
-    "serde",
-] } # on branch v0.44.x
-strum = { version = "0.25", features = ["derive"] }
+strum = { version = "0.27", features = ["derive"] }
+sysinfo = "0.33"
 tempfile = "3"
 tokio = { version = "1.40", features = ["full"] }
 tokio-postgres = "0.7"
-tokio-rustls = { version = "0.26.0", default-features = false } # override by patch, see [patch.crates-io]
+tokio-rustls = { version = "0.26.2", default-features = false }
 tokio-stream = "0.1"
 tokio-util = { version = "0.7", features = ["io-util", "compat"] }
 toml = "0.8.8"
@@ -247,6 +249,7 @@ common-procedure-test = { path = "src/common/procedure-test" }
 common-query = { path = "src/common/query" }
 common-recordbatch = { path = "src/common/recordbatch" }
 common-runtime = { path = "src/common/runtime" }
+common-session = { path = "src/common/session" }
 common-telemetry = { path = "src/common/telemetry" }
 common-test-util = { path = "src/common/test-util" }
 common-time = { path = "src/common/time" }
@@ -279,15 +282,6 @@ store-api = { path = "src/store-api" }
 substrait = { path = "src/common/substrait" }
 table = { path = "src/table" }
 
-[patch.crates-io]
-# change all rustls dependencies to use our fork to default to `ring` to make it "just work"
-hyper-rustls = { git = "https://github.com/GreptimeTeam/hyper-rustls", rev = "a951e03" } # version = "0.27.5" with ring patch
-rustls = { git = "https://github.com/GreptimeTeam/rustls", rev = "34fd0c6" }             # version = "0.23.20" with ring patch
-tokio-rustls = { git = "https://github.com/GreptimeTeam/tokio-rustls", rev = "4604ca6" } # version = "0.26.0" with ring patch
-# This is commented, since we are not using aws-lc-sys, if we need to use it, we need to uncomment this line or use a release after this commit, or it wouldn't compile with gcc < 8.1
-# see https://github.com/aws/aws-lc-rs/pull/526
-# aws-lc-sys = { git ="https://github.com/aws/aws-lc-rs", rev = "556558441e3494af4b156ae95ebc07ebc2fd38aa" }
-
 [workspace.dependencies.meter-macros]
 git = "https://github.com/GreptimeTeam/greptime-meter.git"
 rev = "5618e779cf2bb4755b499c630fba4c35e91898cb"

Makefile

@@ -32,6 +32,10 @@ ifneq ($(strip $(BUILD_JOBS)),)
 	NEXTEST_OPTS += --build-jobs=${BUILD_JOBS}
 endif
 
+ifneq ($(strip $(BUILD_JOBS)),)
+	SQLNESS_OPTS += --jobs ${BUILD_JOBS}
+endif
+
 ifneq ($(strip $(CARGO_PROFILE)),)
 	CARGO_BUILD_OPTS += --profile ${CARGO_PROFILE}
 endif
@@ -193,6 +197,7 @@ fix-clippy: ## Fix clippy violations.
 fmt-check: ## Check code format.
 	cargo fmt --all -- --check
 	python3 scripts/check-snafu.py
+	python3 scripts/check-super-imports.py
 
 .PHONY: start-etcd
 start-etcd: ## Start single node etcd for testing purpose.

README.md

@@ -6,7 +6,7 @@
   </picture>
 </p>
 
-<h2 align="center">Unified & Cost-Effective Observerability Database for Metrics, Logs, and Events</h2>
+<h2 align="center">Unified & Cost-Effective Observability  Database for Metrics, Logs, and Events</h2>
 
 <div align="center">
 <h3 align="center">
@@ -62,35 +62,35 @@
 
 ## Introduction
 
-**GreptimeDB** is an open-source unified & cost-effective observerability database for **Metrics**, **Logs**, and **Events** (also **Traces** in plan). You can gain real-time insights from Edge to Cloud at Any Scale.
+**GreptimeDB** is an open-source, cloud-native, unified & cost-effective observability database for **Metrics**, **Logs**, and **Traces**. You can gain real-time insights from Edge to Cloud at Any Scale.
 
 ## News
 
-**[GreptimeDB archives 1 billion cold run #1 in JSONBench!](https://greptime.com/blogs/2025-03-18-jsonbench-greptimedb-performance)**
+**[GreptimeDB tops JSONBench's billion-record cold run test!](https://greptime.com/blogs/2025-03-18-jsonbench-greptimedb-performance)**
 
 ## Why GreptimeDB
 
-Our core developers have been building observerability data platforms for years. Based on our best practices, GreptimeDB was born to give you:
+Our core developers have been building observability data platforms for years. Based on our best practices, GreptimeDB was born to give you:
 
-* **Unified Processing of Metrics, Logs, and Events**
+* **Unified Processing of Observability Data**
 
-  GreptimeDB unifies observerability data processing by treating all data - whether metrics, logs, or events - as timestamped events with context. Users can analyze this data using either [SQL](https://docs.greptime.com/user-guide/query-data/sql) or [PromQL](https://docs.greptime.com/user-guide/query-data/promql) and leverage stream processing ([Flow](https://docs.greptime.com/user-guide/flow-computation/overview)) to enable continuous aggregation. [Read more](https://docs.greptime.com/user-guide/concepts/data-model).
+  A unified database that treats metrics, logs, and traces as timestamped wide events with context, supporting [SQL](https://docs.greptime.com/user-guide/query-data/sql)/[PromQL](https://docs.greptime.com/user-guide/query-data/promql) queries and [stream processing](https://docs.greptime.com/user-guide/flow-computation/overview) to simplify complex data stacks.
+
+* **High Performance and Cost-effective**
+
+   Written in Rust, combines a distributed query engine with [rich indexing](https://docs.greptime.com/user-guide/manage-data/data-index) (inverted, fulltext, skip data, and vector) and optimized columnar storage to deliver sub-second responses on petabyte-scale data and high-cost efficiency.
 
 * **Cloud-native Distributed Database**
 
   Built for [Kubernetes](https://docs.greptime.com/user-guide/deployments/deploy-on-kubernetes/greptimedb-operator-management). GreptimeDB achieves seamless scalability with its [cloud-native architecture](https://docs.greptime.com/user-guide/concepts/architecture) of separated compute and storage, built on object storage (AWS S3, Azure Blob Storage, etc.) while enabling cross-cloud deployment through a unified data access layer.
 
-* **Performance and Cost-effective**
+* **Developer-Friendly**
 
-  Written in pure Rust for superior performance and reliability. GreptimeDB features a distributed query engine with intelligent indexing to handle high cardinality data efficiently. Its optimized columnar storage achieves 50x cost efficiency on cloud object storage through advanced compression. [Benchmark reports](https://www.greptime.com/blogs/2024-09-09-report-summary).
+  Access standardized SQL/PromQL interfaces through built-in web dashboard, REST API, and MySQL/PostgreSQL protocols. Supports widely adopted data ingestion [protocols](https://docs.greptime.com/user-guide/protocols/overview) for seamless migration and integration.
 
-* **Cloud-Edge Collaboration**
+* **Flexible Deployment Options**
 
-  GreptimeDB seamlessly operates across cloud and edge (ARM/Android/Linux), providing consistent APIs and control plane for unified data management and efficient synchronization. [Learn how to run on Android](https://docs.greptime.com/user-guide/deployments/run-on-android/).
-
-* **Multi-protocol Ingestion, SQL & PromQL Ready**
-
-  Widely adopted database protocols and APIs, including MySQL, PostgreSQL, InfluxDB, OpenTelemetry, Loki and Prometheus, etc.  Effortless Adoption & Seamless Migration. [Supported Protocols Overview](https://docs.greptime.com/user-guide/protocols/overview).
+  Deploy GreptimeDB anywhere from ARM-based edge devices to cloud environments with unified APIs and bandwidth-efficient data synchronization. Query edge and cloud data seamlessly through identical APIs. [Learn how to run on Android](https://docs.greptime.com/user-guide/deployments/run-on-android/).
 
 For more detailed info please read  [Why GreptimeDB](https://docs.greptime.com/user-guide/concepts/why-greptimedb).
 
@@ -233,3 +233,5 @@ Special thanks to all the contributors who have propelled GreptimeDB forward. Fo
 - GreptimeDB's query engine is powered by [Apache Arrow DataFusion™](https://arrow.apache.org/datafusion/).
 - [Apache OpenDAL™](https://opendal.apache.org) gives GreptimeDB a very general and elegant data access abstraction layer.
 - GreptimeDB's meta service is based on [etcd](https://etcd.io/).
+
+<img alt="Known Users" src="https://greptime.com/logo/img/users.png"/>

config/config.md

@@ -12,7 +12,6 @@
 
 | Key | Type | Default | Descriptions |
 | --- | -----| ------- | ----------- |
-| `mode` | String | `standalone` | The running mode of the datanode. It can be `standalone` or `distributed`. |
 | `default_timezone` | String | Unset | The default timezone of the server. |
 | `init_regions_in_background` | Bool | `false` | Initialize all regions in the background during the startup.<br/>By default, it provides services after all regions have been initialized. |
 | `init_regions_parallelism` | Integer | `16` | Parallelism of initializing regions. |
@@ -86,10 +85,6 @@
 | `wal.create_topic_timeout` | String | `30s` | Above which a topic creation operation will be cancelled.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.max_batch_bytes` | String | `1MB` | The max size of a single producer batch.<br/>Warning: Kafka has a default limit of 1MB per message in a topic.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.consumer_wait_timeout` | String | `100ms` | The consumer wait timeout.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_init` | String | `500ms` | The initial backoff delay.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_max` | String | `10s` | The maximum backoff delay.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_base` | Integer | `2` | The exponential backoff rate, i.e. next backoff = base * current backoff.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_deadline` | String | `5mins` | The deadline of retries.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.overwrite_entry_start_id` | Bool | `false` | Ignore missing entries during read WAL.<br/>**It's only used when the provider is `kafka`**.<br/><br/>This option ensures that when Kafka messages are deleted, the system<br/>can still successfully replay memtable data without throwing an<br/>out-of-range error.<br/>However, enabling this option might lead to unexpected data loss,<br/>as the system will skip over missing entries instead of treating<br/>them as critical errors. |
 | `metadata_store` | -- | -- | Metadata storage options. |
 | `metadata_store.file_size` | String | `64MB` | The size of the metadata store log file. |
@@ -98,8 +93,11 @@
 | `procedure` | -- | -- | Procedure storage options. |
 | `procedure.max_retry_times` | Integer | `3` | Procedure max retry time. |
 | `procedure.retry_delay` | String | `500ms` | Initial retry delay of procedures, increases exponentially |
+| `procedure.max_running_procedures` | Integer | `128` | Max running procedures.<br/>The maximum number of procedures that can be running at the same time.<br/>If the number of running procedures exceeds this limit, the procedure will be rejected. |
 | `flow` | -- | -- | flow engine options. |
 | `flow.num_workers` | Integer | `0` | The number of flow worker in flownode.<br/>Not setting(or set to 0) this value will use the number of CPU cores divided by 2. |
+| `query` | -- | -- | The query engine options. |
+| `query.parallelism` | Integer | `0` | Parallelism of the query engine.<br/>Default to 0, which means the number of CPU cores. |
 | `storage` | -- | -- | The data storage options. |
 | `storage.data_home` | String | `./greptimedb_data/` | The working home directory. |
 | `storage.type` | String | `File` | The storage type used to store the data.<br/>- `File`: the data is stored in the local file system.<br/>- `S3`: the data is stored in the S3 object storage.<br/>- `Gcs`: the data is stored in the Google Cloud Storage.<br/>- `Azblob`: the data is stored in the Azure Blob Storage.<br/>- `Oss`: the data is stored in the Aliyun OSS. |
@@ -274,6 +272,8 @@
 | `meta_client.metadata_cache_max_capacity` | Integer | `100000` | The configuration about the cache of the metadata. |
 | `meta_client.metadata_cache_ttl` | String | `10m` | TTL of the metadata cache. |
 | `meta_client.metadata_cache_tti` | String | `5m` | -- |
+| `query` | -- | -- | The query engine options. |
+| `query.parallelism` | Integer | `0` | Parallelism of the query engine.<br/>Default to 0, which means the number of CPU cores. |
 | `datanode` | -- | -- | Datanode options. |
 | `datanode.client` | -- | -- | Datanode client options. |
 | `datanode.client.connect_timeout` | String | `10s` | -- |
@@ -328,6 +328,7 @@
 | `procedure.max_retry_times` | Integer | `12` | Procedure max retry time. |
 | `procedure.retry_delay` | String | `500ms` | Initial retry delay of procedures, increases exponentially |
 | `procedure.max_metadata_value_size` | String | `1500KiB` | Auto split large value<br/>GreptimeDB procedure uses etcd as the default metadata storage backend.<br/>The etcd the maximum size of any request is 1.5 MiB<br/>1500KiB = 1536KiB (1.5MiB) - 36KiB (reserved size of key)<br/>Comments out the `max_metadata_value_size`, for don't split large value (no limit). |
+| `procedure.max_running_procedures` | Integer | `128` | Max running procedures.<br/>The maximum number of procedures that can be running at the same time.<br/>If the number of running procedures exceeds this limit, the procedure will be rejected. |
 | `failure_detector` | -- | -- | -- |
 | `failure_detector.threshold` | Float | `8.0` | The threshold value used by the failure detector to determine failure conditions. |
 | `failure_detector.min_std_deviation` | String | `100ms` | The minimum standard deviation of the heartbeat intervals, used to calculate acceptable variations. |
@@ -342,15 +343,14 @@
 | `wal.provider` | String | `raft_engine` | -- |
 | `wal.broker_endpoints` | Array | -- | The broker endpoints of the Kafka cluster. |
 | `wal.auto_create_topics` | Bool | `true` | Automatically create topics for WAL.<br/>Set to `true` to automatically create topics for WAL.<br/>Otherwise, use topics named `topic_name_prefix_[0..num_topics)` |
+| `wal.auto_prune_interval` | String | `0s` | Interval of automatically WAL pruning.<br/>Set to `0s` to disable automatically WAL pruning which delete unused remote WAL entries periodically. |
+| `wal.trigger_flush_threshold` | Integer | `0` | The threshold to trigger a flush operation of a region in automatically WAL pruning.<br/>Metasrv will send a flush request to flush the region when:<br/>`trigger_flush_threshold` + `prunable_entry_id` < `max_prunable_entry_id`<br/>where:<br/>- `prunable_entry_id` is the maximum entry id that can be pruned of the region.<br/>- `max_prunable_entry_id` is the maximum prunable entry id among all regions in the same topic.<br/>Set to `0` to disable the flush operation. |
+| `wal.auto_prune_parallelism` | Integer | `10` | Concurrent task limit for automatically WAL pruning. |
 | `wal.num_topics` | Integer | `64` | Number of topics. |
 | `wal.selector_type` | String | `round_robin` | Topic selector type.<br/>Available selector types:<br/>- `round_robin` (default) |
 | `wal.topic_name_prefix` | String | `greptimedb_wal_topic` | A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`.<br/>Only accepts strings that match the following regular expression pattern:<br/>[a-zA-Z_:-][a-zA-Z0-9_:\-\.@#]*<br/>i.g., greptimedb_wal_topic_0, greptimedb_wal_topic_1. |
 | `wal.replication_factor` | Integer | `1` | Expected number of replicas of each partition. |
 | `wal.create_topic_timeout` | String | `30s` | Above which a topic creation operation will be cancelled. |
-| `wal.backoff_init` | String | `500ms` | The initial backoff for kafka clients. |
-| `wal.backoff_max` | String | `10s` | The maximum backoff for kafka clients. |
-| `wal.backoff_base` | Integer | `2` | Exponential backoff rate, i.e. next backoff = base * current backoff. |
-| `wal.backoff_deadline` | String | `5mins` | Stop reconnecting if the total wait time reaches the deadline. If this config is missing, the reconnecting won't terminate. |
 | `logging` | -- | -- | The logging options. |
 | `logging.dir` | String | `./greptimedb_data/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
 | `logging.level` | String | Unset | The log level. Can be `info`/`debug`/`warn`/`error`. |
@@ -381,7 +381,6 @@
 
 | Key | Type | Default | Descriptions |
 | --- | -----| ------- | ----------- |
-| `mode` | String | `standalone` | The running mode of the datanode. It can be `standalone` or `distributed`. |
 | `node_id` | Integer | Unset | The datanode identifier and should be unique in the cluster. |
 | `require_lease_before_startup` | Bool | `false` | Start services after regions have obtained leases.<br/>It will block the datanode start if it can't receive leases in the heartbeat from metasrv. |
 | `init_regions_in_background` | Bool | `false` | Initialize all regions in the background during the startup.<br/>By default, it provides services after all regions have been initialized. |
@@ -434,13 +433,11 @@
 | `wal.broker_endpoints` | Array | -- | The Kafka broker endpoints.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.max_batch_bytes` | String | `1MB` | The max size of a single producer batch.<br/>Warning: Kafka has a default limit of 1MB per message in a topic.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.consumer_wait_timeout` | String | `100ms` | The consumer wait timeout.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_init` | String | `500ms` | The initial backoff delay.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_max` | String | `10s` | The maximum backoff delay.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_base` | Integer | `2` | The exponential backoff rate, i.e. next backoff = base * current backoff.<br/>**It's only used when the provider is `kafka`**. |
-| `wal.backoff_deadline` | String | `5mins` | The deadline of retries.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.create_index` | Bool | `true` | Whether to enable WAL index creation.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.dump_index_interval` | String | `60s` | The interval for dumping WAL indexes.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.overwrite_entry_start_id` | Bool | `false` | Ignore missing entries during read WAL.<br/>**It's only used when the provider is `kafka`**.<br/><br/>This option ensures that when Kafka messages are deleted, the system<br/>can still successfully replay memtable data without throwing an<br/>out-of-range error.<br/>However, enabling this option might lead to unexpected data loss,<br/>as the system will skip over missing entries instead of treating<br/>them as critical errors. |
+| `query` | -- | -- | The query engine options. |
+| `query.parallelism` | Integer | `0` | Parallelism of the query engine.<br/>Default to 0, which means the number of CPU cores. |
 | `storage` | -- | -- | The data storage options. |
 | `storage.data_home` | String | `./greptimedb_data/` | The working home directory. |
 | `storage.type` | String | `File` | The storage type used to store the data.<br/>- `File`: the data is stored in the local file system.<br/>- `S3`: the data is stored in the S3 object storage.<br/>- `Gcs`: the data is stored in the Google Cloud Storage.<br/>- `Azblob`: the data is stored in the Azure Blob Storage.<br/>- `Oss`: the data is stored in the Aliyun OSS. |
@@ -551,7 +548,6 @@
 
 | Key | Type | Default | Descriptions |
 | --- | -----| ------- | ----------- |
-| `mode` | String | `distributed` | The running mode of the flownode. It can be `standalone` or `distributed`. |
 | `node_id` | Integer | Unset | The flownode identifier and should be unique in the cluster. |
 | `flow` | -- | -- | flow engine options. |
 | `flow.num_workers` | Integer | `0` | The number of flow worker in flownode.<br/>Not setting(or set to 0) this value will use the number of CPU cores divided by 2. |

config/datanode.example.toml

@@ -1,6 +1,3 @@
-## The running mode of the datanode. It can be `standalone` or `distributed`.
-mode = "standalone"
-
 ## The datanode identifier and should be unique in the cluster.
 ## @toml2docs:none-default
 node_id = 42
@@ -169,22 +166,6 @@ max_batch_bytes = "1MB"
 ## **It's only used when the provider is `kafka`**.
 consumer_wait_timeout = "100ms"
 
-## The initial backoff delay.
-## **It's only used when the provider is `kafka`**.
-backoff_init = "500ms"
-
-## The maximum backoff delay.
-## **It's only used when the provider is `kafka`**.
-backoff_max = "10s"
-
-## The exponential backoff rate, i.e. next backoff = base * current backoff.
-## **It's only used when the provider is `kafka`**.
-backoff_base = 2
-
-## The deadline of retries.
-## **It's only used when the provider is `kafka`**.
-backoff_deadline = "5mins"
-
 ## Whether to enable WAL index creation.
 ## **It's only used when the provider is `kafka`**.
 create_index = true
@@ -262,6 +243,12 @@ overwrite_entry_start_id = false
 # credential = "base64-credential"
 # endpoint = "https://storage.googleapis.com"
 
+## The query engine options.
+[query]
+## Parallelism of the query engine.
+## Default to 0, which means the number of CPU cores.
+parallelism = 0
+
 ## The data storage options.
 [storage]
 ## The working home directory.

config/flownode.example.toml

@@ -1,6 +1,3 @@
-## The running mode of the flownode. It can be `standalone` or `distributed`.
-mode = "distributed"
-
 ## The flownode identifier and should be unique in the cluster.
 ## @toml2docs:none-default
 node_id = 14

config/frontend.example.toml

@@ -179,6 +179,12 @@ metadata_cache_ttl = "10m"
 # TTI of the metadata cache.
 metadata_cache_tti = "5m"
 
+## The query engine options.
+[query]
+## Parallelism of the query engine.
+## Default to 0, which means the number of CPU cores.
+parallelism = 0
+
 ## Datanode options.
 [datanode]
 ## Datanode client options.

config/metasrv.example.toml

@@ -79,6 +79,11 @@ retry_delay = "500ms"
 ## Comments out the `max_metadata_value_size`, for don't split large value (no limit).
 max_metadata_value_size = "1500KiB"
 
+## Max running procedures.
+## The maximum number of procedures that can be running at the same time.
+## If the number of running procedures exceeds this limit, the procedure will be rejected.
+max_running_procedures = 128
+
 # Failure detectors options.
 [failure_detector]
 
@@ -125,6 +130,22 @@ broker_endpoints = ["127.0.0.1:9092"]
 ## Otherwise, use topics named `topic_name_prefix_[0..num_topics)`
 auto_create_topics = true
 
+## Interval of automatically WAL pruning.
+## Set to `0s` to disable automatically WAL pruning which delete unused remote WAL entries periodically.
+auto_prune_interval = "0s"
+
+## The threshold to trigger a flush operation of a region in automatically WAL pruning.
+## Metasrv will send a flush request to flush the region when:
+## `trigger_flush_threshold` + `prunable_entry_id` < `max_prunable_entry_id`
+## where:
+## - `prunable_entry_id` is the maximum entry id that can be pruned of the region.
+## - `max_prunable_entry_id` is the maximum prunable entry id among all regions in the same topic.
+## Set to `0` to disable the flush operation.
+trigger_flush_threshold = 0
+
+## Concurrent task limit for automatically WAL pruning.
+auto_prune_parallelism = 10
+
 ## Number of topics.
 num_topics = 64
 
@@ -144,17 +165,6 @@ replication_factor = 1
 
 ## Above which a topic creation operation will be cancelled.
 create_topic_timeout = "30s"
-## The initial backoff for kafka clients.
-backoff_init = "500ms"
-
-## The maximum backoff for kafka clients.
-backoff_max = "10s"
-
-## Exponential backoff rate, i.e. next backoff = base * current backoff.
-backoff_base = 2
-
-## Stop reconnecting if the total wait time reaches the deadline. If this config is missing, the reconnecting won't terminate.
-backoff_deadline = "5mins"
 
 # The Kafka SASL configuration.
 # **It's only used when the provider is `kafka`**.

config/standalone.example.toml

@@ -1,6 +1,3 @@
-## The running mode of the datanode. It can be `standalone` or `distributed`.
-mode = "standalone"
-
 ## The default timezone of the server.
 ## @toml2docs:none-default
 default_timezone = "UTC"
@@ -242,22 +239,6 @@ max_batch_bytes = "1MB"
 ## **It's only used when the provider is `kafka`**.
 consumer_wait_timeout = "100ms"
 
-## The initial backoff delay.
-## **It's only used when the provider is `kafka`**.
-backoff_init = "500ms"
-
-## The maximum backoff delay.
-## **It's only used when the provider is `kafka`**.
-backoff_max = "10s"
-
-## The exponential backoff rate, i.e. next backoff = base * current backoff.
-## **It's only used when the provider is `kafka`**.
-backoff_base = 2
-
-## The deadline of retries.
-## **It's only used when the provider is `kafka`**.
-backoff_deadline = "5mins"
-
 ## Ignore missing entries during read WAL.
 ## **It's only used when the provider is `kafka`**.
 ##
@@ -302,6 +283,10 @@ purge_interval = "1m"
 max_retry_times = 3
 ## Initial retry delay of procedures, increases exponentially
 retry_delay = "500ms"
+## Max running procedures.
+## The maximum number of procedures that can be running at the same time.
+## If the number of running procedures exceeds this limit, the procedure will be rejected.
+max_running_procedures = 128
 
 ## flow engine options.
 [flow]
@@ -349,6 +334,12 @@ retry_delay = "500ms"
 # credential = "base64-credential"
 # endpoint = "https://storage.googleapis.com"
 
+## The query engine options.
+[query]
+## Parallelism of the query engine.
+## Default to 0, which means the number of CPU cores.
+parallelism = 0
+
 ## The data storage options.
 [storage]
 ## The working home directory.

docs/how-to/how-to-profile-memory.md

@@ -1,6 +1,6 @@
 # Profile memory usage of GreptimeDB
 
-This crate provides an easy approach to dump memory profiling info.
+This crate provides an easy approach to dump memory profiling info. A set of ready to use scripts is provided in [docs/how-to/memory-profile-scripts](docs/how-to/memory-profile-scripts).
 
 ## Prerequisites
 ### jemalloc

docs/how-to/how-to-write-fuzz-tests.md

@@ -3,7 +3,7 @@
 This document introduces how to write fuzz tests in GreptimeDB.
 
 ## What is a fuzz test
-Fuzz test is tool that leverage deterministic random generation to assist in finding bugs. The goal of fuzz tests is to identify inputs generated by the fuzzer that cause system panics, crashes, or unexpected behaviors to occur. And we are using the [cargo-fuzz](https://github.com/rust-fuzz/cargo-fuzz) to run our fuzz test targets. 
+Fuzz test is tool that leverage deterministic random generation to assist in finding bugs. The goal of fuzz tests is to identify inputs generated by the fuzzer that cause system panics, crashes, or unexpected behaviors to occur. And we are using the [cargo-fuzz](https://github.com/rust-fuzz/cargo-fuzz) to run our fuzz test targets.
 
 ## Why we need them
 - Find bugs by leveraging random generation
@@ -13,7 +13,7 @@ Fuzz test is tool that leverage deterministic random generation to assist in fin
 All fuzz test-related resources are located in the `/tests-fuzz` directory.
 There are two types of resources: (1) fundamental components and (2) test targets.
 
-### Fundamental components 
+### Fundamental components
 They are located in the `/tests-fuzz/src` directory. The fundamental components define how to generate SQLs (including dialects for different protocols) and validate execution results (e.g., column attribute validation), etc.
 
 ### Test targets
@@ -21,25 +21,25 @@ They are located in the `/tests-fuzz/targets` directory, with each file represen
 
 Figure 1 illustrates the fundamental components of the fuzz test provide the ability to generate random SQLs. It utilizes a Random Number Generator (Rng) to generate the Intermediate Representation (IR), then employs a DialectTranslator to produce specified dialects for different protocols. Finally, the fuzz tests send the generated SQL via the specified protocol and verify that the execution results meet expectations.
 ```
-                            Rng                                 
-                             |                                  
-                             |                                  
-                             v                                  
-                       ExprGenerator                            
-                             |                                  
-                             |                                  
-                             v                                  
-               Intermediate representation (IR)                 
-                             |                                  
-                             |                                  
-      +----------------------+----------------------+           
-      |                      |                      |           
-      v                      v                      v           
+                            Rng
+                             |
+                             |
+                             v
+                       ExprGenerator
+                             |
+                             |
+                             v
+               Intermediate representation (IR)
+                             |
+                             |
+      +----------------------+----------------------+
+      |                      |                      |
+      v                      v                      v
 MySQLTranslator    PostgreSQLTranslator   OtherDialectTranslator
-      |                      |                      |           
-      |                      |                      |           
-      v                      v                      v           
-SQL(MySQL Dialect)         .....                  .....         
+      |                      |                      |
+      |                      |                      |
+      v                      v                      v
+SQL(MySQL Dialect)         .....                  .....
       |
       |
       v
@@ -133,4 +133,4 @@ fuzz_target!(|input: FuzzInput| {
     cargo fuzz run <fuzz-target> --fuzz-dir tests-fuzz
 ```
 
-For more details, please refer to this [document](/tests-fuzz/README.md).
+For more details, please refer to this [document](/tests-fuzz/README.md).

docs/how-to/memory-profile-scripts/scripts/README.md

@@ -0,0 +1,52 @@
+# Memory Analysis Process
+This section will guide you through the process of analyzing memory usage for greptimedb.
+
+1. Get the `jeprof` tool script, see the next section("Getting the `jeprof` tool") for details.
+
+2. After starting `greptimedb`(with env var `MALLOC_CONF=prof:true`), execute the `dump.sh` script with the PID of the `greptimedb` process as an argument. This continuously monitors memory usage and captures profiles when exceeding thresholds (e.g. +20MB within 10 minutes). Outputs `greptime-{timestamp}.gprof` files.
+
+3. With 2-3 gprof files, run `gen_flamegraph.sh` in the same environment to generate flame graphs showing memory allocation call stacks.
+
+4.  **NOTE:** The `gen_flamegraph.sh` script requires `jeprof` and optionally `flamegraph.pl` to be in the current directory. If needed to gen flamegraph now, run the `get_flamegraph_tool.sh` script, which downloads the flame graph generation tool `flamegraph.pl` to the current directory.
+    The usage of `gen_flamegraph.sh` is:
+
+    `Usage: ./gen_flamegraph.sh <binary_path> <gprof_directory>`
+    where `<binary_path>` is the path to the greptimedb binary, `<gprof_directory>` is the directory containing the gprof files(the directory `dump.sh` is dumping profiles to).
+    Example call: `./gen_flamegraph.sh ./greptime .`
+
+    Generating the flame graph might take a few minutes. The generated flame graphs are located in the `<gprof_directory>/flamegraphs` directory. Or if no `flamegraph.pl` is found, it will only contain `.collapse` files which is also fine.
+5.  You can send the generated flame graphs(the entire folder of `<gprof_directory>/flamegraphs`) to developers for further analysis.
+
+
+## Getting the `jeprof` tool
+there are three ways to get `jeprof`, list in here from simple to complex, using any one of those methods is ok, as long as it's the same environment as the `greptimedb` will be running on:
+1. If you are compiling greptimedb from source, then `jeprof` is already produced during compilation. After running `cargo build`, execute `find_compiled_jeprof.sh`. This will copy `jeprof` to the current directory.
+2. Or, if you have the Rust toolchain installed locally, simply follow these commands:
+```bash
+cargo new get_jeprof
+cd get_jeprof
+```
+Then add this line to `Cargo.toml`:
+```toml
+[dependencies]
+tikv-jemalloc-ctl = { version = "0.6", features = ["use_std", "stats"] }
+```
+then run:
+```bash
+cargo build
+```
+after that the `jeprof` tool is produced. Now run `find_compiled_jeprof.sh` in current directory, it will copy the `jeprof` tool to the current directory.
+
+3. compile jemalloc from source
+you can first clone this repo, and checkout to this commit:
+```bash
+git clone https://github.com/tikv/jemalloc.git
+cd jemalloc
+git checkout e13ca993e8ccb9ba9847cc330696e02839f328f7
+```
+then run:
+```bash
+./configure
+make
+```
+and `jeprof` is in `.bin/` directory. Copy it to the current directory.

docs/how-to/memory-profile-scripts/scripts/dump.sh

@@ -0,0 +1,78 @@
+#!/bin/bash
+
+# Monitors greptime process memory usage every 10 minutes
+# Triggers memory profile capture via `curl -X POST localhost:4000/debug/prof/mem > greptime-{timestamp}.gprof`
+# when memory increases by more than 20MB since last check
+# Generated profiles can be analyzed using flame graphs as described in `how-to-profile-memory.md`
+# (jeprof is compiled with the database - see documentation)
+# Alternative: Share binaries + profiles for analysis (Docker images preferred)
+
+# Threshold in Kilobytes (20 MB)
+threshold_kb=$((20 * 1024))
+sleep_interval=$((10 * 60))
+
+# Variable to store the last measured memory usage in KB
+last_mem_kb=0
+
+echo "Starting memory monitoring for 'greptime' process..."
+
+while true; do
+
+    # Check if PID is provided as an argument
+    if [ -z "$1" ]; then
+        echo "$(date): PID must be provided as a command-line argument."
+        exit 1
+    fi
+
+    pid="$1"
+
+    # Validate that the PID is a number
+    if ! [[ "$pid" =~ ^[0-9]+$ ]]; then
+        echo "$(date): Invalid PID: '$pid'. PID must be a number."
+        exit 1
+    fi
+
+    # Get the current Resident Set Size (RSS) in Kilobytes
+    current_mem_kb=$(ps -o rss= -p "$pid")
+
+    # Check if ps command was successful and returned a number
+    if ! [[ "$current_mem_kb" =~ ^[0-9]+$ ]]; then
+        echo "$(date): Failed to get memory usage for PID $pid. Skipping check."
+        # Keep last_mem_kb to avoid false positives if the process briefly becomes unreadable.
+        continue
+    fi
+
+    echo "$(date): Current memory usage for PID $pid: ${current_mem_kb} KB"
+
+    # Compare with the last measurement
+    # if it's the first run, also do a baseline dump just to make sure we can dump
+    
+    diff_kb=$((current_mem_kb - last_mem_kb))
+    echo "$(date): Memory usage change since last check: ${diff_kb} KB"
+
+    if [ "$diff_kb" -gt "$threshold_kb" ]; then
+        echo "$(date): Memory increase (${diff_kb} KB) exceeded threshold (${threshold_kb} KB). Dumping profile..."
+        timestamp=$(date +%Y%m%d%H%M%S)
+        profile_file="greptime-${timestamp}.gprof"
+        # Execute curl and capture output to file
+        if curl -sf -X POST localhost:4000/debug/prof/mem > "$profile_file"; then
+            echo "$(date): Memory profile saved to $profile_file"
+        else
+            echo "$(date): Failed to dump memory profile (curl exit code: $?)."
+            # Remove the potentially empty/failed profile file
+            rm -f "$profile_file"
+        fi
+    else
+            echo "$(date): Memory increase (${diff_kb} KB) is within the threshold (${threshold_kb} KB)."
+    fi
+    
+
+    # Update the last memory usage
+    last_mem_kb=$current_mem_kb
+    
+    # Wait for 5 minutes
+    echo "$(date): Sleeping for $sleep_interval seconds..."
+    sleep $sleep_interval
+done
+
+echo "Memory monitoring script stopped." # This line might not be reached in normal operation

docs/how-to/memory-profile-scripts/scripts/find_compiled_jeprof.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+# Locates compiled jeprof binary (memory analysis tool) after cargo build
+# Copies it to current directory from target/ build directories
+
+JPROF_PATH=$(find . -name 'jeprof' -print -quit)
+if [ -n "$JPROF_PATH" ]; then
+  echo "Found jeprof at $JPROF_PATH"
+  cp "$JPROF_PATH" .
+  chmod +x jeprof
+  echo "Copied jeprof to current directory and made it executable."
+else
+  echo "jeprof not found"
+  exit 1
+fi

docs/how-to/memory-profile-scripts/scripts/gen_flamegraph.sh

@@ -0,0 +1,89 @@
+#!/bin/bash
+
+# Generate flame graphs from a series of `.gprof` files
+# First argument: Path to the binary executable
+# Second argument: Path to directory containing gprof files
+# Requires `jeprof` and `flamegraph.pl` in current directory
+# What this script essentially does is:
+# ./jeprof <binary> <gprof> --collapse | ./flamegraph.pl > <output>
+# For differential analysis between consecutive profiles:
+# ./jeprof <binary> --base <gprof1> <gprof2> --collapse | ./flamegraph.pl > <output_diff>
+
+set -e # Exit immediately if a command exits with a non-zero status.
+
+# Check for required tools
+if [ ! -f "./jeprof" ]; then
+    echo "Error: jeprof not found in the current directory."
+    exit 1
+fi
+
+if [ ! -f "./flamegraph.pl" ]; then
+    echo "Error: flamegraph.pl not found in the current directory."
+    exit 1
+fi
+
+# Check arguments
+if [ "$#" -ne 2 ]; then
+    echo "Usage: $0 <binary_path> <gprof_directory>"
+    exit 1
+fi
+
+BINARY_PATH=$1
+GPROF_DIR=$2
+OUTPUT_DIR="${GPROF_DIR}/flamegraphs" # Store outputs in a subdirectory
+
+if [ ! -f "$BINARY_PATH" ]; then
+    echo "Error: Binary file not found at $BINARY_PATH"
+    exit 1
+fi
+
+if [ ! -d "$GPROF_DIR" ]; then
+    echo "Error: gprof directory not found at $GPROF_DIR"
+    exit 1
+fi
+
+mkdir -p "$OUTPUT_DIR"
+echo "Generating flamegraphs in $OUTPUT_DIR"
+
+# Find and sort gprof files
+# Use find + sort -V for natural sort of version numbers if present in filenames
+# Use null-terminated strings for safety with find/xargs/sort
+mapfile -d $'\0' gprof_files < <(find "$GPROF_DIR" -maxdepth 1 -name '*.gprof' -print0 | sort -zV)
+
+if [ ${#gprof_files[@]} -eq 0 ]; then
+    echo "No .gprof files found in $GPROF_DIR"
+    exit 0
+fi
+
+prev_gprof=""
+
+# Generate flamegraphs
+for gprof_file in "${gprof_files[@]}"; do
+    # Skip empty entries if any
+    if [ -z "$gprof_file" ]; then
+        continue
+    fi
+
+    filename=$(basename "$gprof_file" .gprof)
+    output_collapse="${OUTPUT_DIR}/${filename}.collapse"
+    output_svg="${OUTPUT_DIR}/${filename}.svg"
+    echo "Generating collapse file for $gprof_file -> $output_collapse"
+    ./jeprof "$BINARY_PATH" "$gprof_file" --collapse > "$output_collapse"
+    echo "Generating flamegraph for $gprof_file -> $output_svg"
+    ./flamegraph.pl "$output_collapse" > "$output_svg" || true
+
+    # Generate diff flamegraph if not the first file
+    if [ -n "$prev_gprof" ]; then
+        prev_filename=$(basename "$prev_gprof" .gprof)
+        diff_output_collapse="${OUTPUT_DIR}/${prev_filename}_vs_${filename}_diff.collapse"
+        diff_output_svg="${OUTPUT_DIR}/${prev_filename}_vs_${filename}_diff.svg"
+        echo "Generating diff collapse file for $prev_gprof vs $gprof_file -> $diff_output_collapse"
+        ./jeprof "$BINARY_PATH" --base "$prev_gprof" "$gprof_file" --collapse > "$diff_output_collapse"
+        echo "Generating diff flamegraph for $prev_gprof vs $gprof_file -> $diff_output_svg"
+        ./flamegraph.pl "$diff_output_collapse" > "$diff_output_svg" || true
+    fi
+
+    prev_gprof="$gprof_file"
+done
+
+echo "Flamegraph generation complete."

docs/how-to/memory-profile-scripts/scripts/gen_from_collapse.sh

@@ -0,0 +1,44 @@
+#!/bin/bash
+
+# Generate flame graphs from .collapse files
+# Argument: Path to directory containing collapse files
+# Requires `flamegraph.pl` in current directory
+
+# Check if flamegraph.pl exists
+if [ ! -f "./flamegraph.pl" ]; then
+    echo "Error: flamegraph.pl not found in the current directory."
+    exit 1
+fi
+
+# Check if directory argument is provided
+if [ -z "$1" ]; then
+    echo "Usage: $0 <collapse_directory>"
+    exit 1
+fi
+
+COLLAPSE_DIR=$1
+
+# Check if the provided argument is a directory
+if [ ! -d "$COLLAPSE_DIR" ]; then
+    echo "Error: '$COLLAPSE_DIR' is not a valid directory."
+    exit 1
+fi
+
+echo "Generating flame graphs from collapse files in '$COLLAPSE_DIR'..."
+
+# Find and process each .collapse file
+find "$COLLAPSE_DIR" -maxdepth 1 -name "*.collapse" -print0 | while IFS= read -r -d $'\0' collapse_file; do
+    if [ -f "$collapse_file" ]; then
+        # Construct the output SVG filename
+        svg_file="${collapse_file%.collapse}.svg"
+        echo "Generating $svg_file from $collapse_file..."
+        ./flamegraph.pl "$collapse_file" > "$svg_file"
+        if [ $? -ne 0 ]; then
+            echo "Error generating flame graph for $collapse_file"
+        else
+            echo "Successfully generated $svg_file"
+        fi
+    fi
+done
+
+echo "Flame graph generation complete."

docs/how-to/memory-profile-scripts/scripts/get_flamegraph_tool.sh

@@ -0,0 +1,6 @@
+#!/bin/bash
+
+# Download flamegraph.pl to current directory - this is the flame graph generation tool script
+
+curl https://raw.githubusercontent.com/brendangregg/FlameGraph/master/flamegraph.pl > ./flamegraph.pl
+chmod +x ./flamegraph.pl

docs/rfcs/2025-02-06-remote-wal-purge.md

@@ -0,0 +1,77 @@
+---
+Feature Name: Remote WAL Purge
+Tracking Issue: https://github.com/GreptimeTeam/greptimedb/issues/5474
+Date: 2025-02-06
+Author: "Yuhan Wang <profsyb@gmail.com>"
+---
+
+# Summary
+
+This RFC proposes a method for purging remote WAL in the database.
+
+# Motivation
+
+Currently only local wal entries are purged when flushing, while remote wal does nothing.
+
+# Details
+
+```mermaid
+sequenceDiagram
+    Region0->>Kafka: Last entry id of the topic in use
+    Region0->>WALPruner: Heartbeat with last entry id
+    WALPruner->>+WALPruner: Time Loop
+    WALPruner->>+ProcedureManager: Submit purge procedure
+    ProcedureManager->>Region0: Flush request
+    ProcedureManager->>Kafka: Prune WAL entries
+    Region0->>Region0: Flush
+```
+
+## Steps
+
+### Before purge
+
+Before purging remote WAL, metasrv needs to know:
+
+1. `last_entry_id` of each region.
+2. `kafka_topic_last_entry_id` which is the last entry id of the topic in use. Can be lazily updated and needed when region has empty memtable.
+3. Kafka topics that each region uses.
+
+The states are maintained through:
+1. Heartbeat: Datanode sends `last_entry_id` to metasrv in heartbeat. As for regions with empty memtable, `last_entry_id` should equals to `kafka_topic_last_entry_id`.
+2. Metasrv maintains a topic-region map to know which region uses which topic.
+
+`kafka_topic_last_entry_id` will be maintained by the region itself. Region will update the value after `k` heartbeats if the memtable is empty.
+
+### Purge procedure
+
+We can better handle locks utilizing current procedure. It's quite similar to the region migration procedure.
+
+After a period of time, metasrv will submit a purge procedure to ProcedureManager. The purge will apply to all topics.
+
+The procedure is divided into following stages:
+
+1. Preparation:
+   - Retrieve `last_entry_id` of each region kvbackend.
+   - Choose regions that have a relatively small `last_entry_id` as candidate regions, which means we need to send a flush request to these regions.
+2. Communication:
+   - Send flush requests to candidate regions.
+3. Purge:
+   - Choose proper entry id to delete for each topic. The entry should be the smallest `last_entry_id - 1` among all regions. 
+   - Delete legacy entries in Kafka.
+   - Store the `last_purged_entry_id` in kvbackend. It should be locked to prevent other regions from replaying the purged entries.
+
+### After purge
+
+After purge, there may be some regions that have `last_entry_id` smaller than the entry we just deleted. It's legal since we only delete the entries that are not needed anymore.
+
+When restarting a region, it should query the `last_purged_entry_id` from metasrv and replay from `min(last_entry_id, last_purged_entry_id)`.
+
+### Error handling
+
+No persisted states are needed since all states are maintained in kvbackend.
+
+Retry when failed to retrieving metadata from kvbackend.
+
+# Alternatives
+
+Purge time can depend on the size of the WAL entries instead of a fixed period of time, which may be more efficient.

scripts/check-super-imports.py

@@ -0,0 +1,74 @@
+# Copyright 2023 Greptime Team
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+import re
+from multiprocessing import Pool
+
+
+def find_rust_files(directory):
+    rust_files = []
+    for root, _, files in os.walk(directory):
+        # Skip files with "test" in the path
+        if "test" in root.lower():
+            continue
+
+        for file in files:
+            # Skip files with "test" in the filename
+            if "test" in file.lower():
+                continue
+
+            if file.endswith(".rs"):
+                rust_files.append(os.path.join(root, file))
+    return rust_files
+
+
+def check_file_for_super_import(file_path):
+    with open(file_path, "r") as file:
+        lines = file.readlines()
+
+    violations = []
+    for line_number, line in enumerate(lines, 1):
+        # Check for "use super::" without leading tab
+        if line.startswith("use super::"):
+            violations.append((line_number, line.strip()))
+
+    if violations:
+        return file_path, violations
+    return None
+
+
+def main():
+    rust_files = find_rust_files(".")
+
+    with Pool() as pool:
+        results = pool.map(check_file_for_super_import, rust_files)
+
+    # Filter out None results
+    violations = [result for result in results if result]
+
+    if violations:
+        print("Found 'use super::' without leading tab in the following files:")
+        counter = 1
+        for file_path, file_violations in violations:
+            for line_number, line in file_violations:
+                print(f"{counter:>5} {file_path}:{line_number} - {line}")
+                counter += 1
+        raise SystemExit(1)
+    else:
+        print("No 'use super::' without leading tab found. All files are compliant.")
+
+
+if __name__ == "__main__":
+    main()

src/api/src/v1/column_def.rs

@@ -15,10 +15,13 @@
 use std::collections::HashMap;
 
 use datatypes::schema::{
-    ColumnDefaultConstraint, ColumnSchema, FulltextAnalyzer, FulltextOptions, SkippingIndexOptions,
-    SkippingIndexType, COMMENT_KEY, FULLTEXT_KEY, INVERTED_INDEX_KEY, SKIPPING_INDEX_KEY,
+    ColumnDefaultConstraint, ColumnSchema, FulltextAnalyzer, FulltextBackend, FulltextOptions,
+    SkippingIndexOptions, SkippingIndexType, COMMENT_KEY, FULLTEXT_KEY, INVERTED_INDEX_KEY,
+    SKIPPING_INDEX_KEY,
+};
+use greptime_proto::v1::{
+    Analyzer, FulltextBackend as PbFulltextBackend, SkippingIndexType as PbSkippingIndexType,
 };
-use greptime_proto::v1::{Analyzer, SkippingIndexType as PbSkippingIndexType};
 use snafu::ResultExt;
 
 use crate::error::{self, Result};
@@ -142,13 +145,21 @@ pub fn options_from_inverted() -> ColumnOptions {
 }
 
 /// Tries to construct a `FulltextAnalyzer` from the given analyzer.
-pub fn as_fulltext_option(analyzer: Analyzer) -> FulltextAnalyzer {
+pub fn as_fulltext_option_analyzer(analyzer: Analyzer) -> FulltextAnalyzer {
     match analyzer {
         Analyzer::English => FulltextAnalyzer::English,
         Analyzer::Chinese => FulltextAnalyzer::Chinese,
     }
 }
 
+/// Tries to construct a `FulltextBackend` from the given backend.
+pub fn as_fulltext_option_backend(backend: PbFulltextBackend) -> FulltextBackend {
+    match backend {
+        PbFulltextBackend::Bloom => FulltextBackend::Bloom,
+        PbFulltextBackend::Tantivy => FulltextBackend::Tantivy,
+    }
+}
+
 /// Tries to construct a `SkippingIndexType` from the given skipping index type.
 pub fn as_skipping_index_type(skipping_index_type: PbSkippingIndexType) -> SkippingIndexType {
     match skipping_index_type {
@@ -160,7 +171,7 @@ pub fn as_skipping_index_type(skipping_index_type: PbSkippingIndexType) -> Skipp
 mod tests {
 
     use datatypes::data_type::ConcreteDataType;
-    use datatypes::schema::FulltextAnalyzer;
+    use datatypes::schema::{FulltextAnalyzer, FulltextBackend};
 
     use super::*;
     use crate::v1::ColumnDataType;
@@ -219,13 +230,14 @@ mod tests {
                 enable: true,
                 analyzer: FulltextAnalyzer::English,
                 case_sensitive: false,
+                backend: FulltextBackend::Bloom,
             })
             .unwrap();
         schema.set_inverted_index(true);
         let options = options_from_column_schema(&schema).unwrap();
         assert_eq!(
             options.options.get(FULLTEXT_GRPC_KEY).unwrap(),
-            "{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false}"
+            "{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false,\"backend\":\"bloom\"}"
         );
         assert_eq!(
             options.options.get(INVERTED_INDEX_GRPC_KEY).unwrap(),
@@ -239,11 +251,12 @@ mod tests {
             enable: true,
             analyzer: FulltextAnalyzer::English,
             case_sensitive: false,
+            backend: FulltextBackend::Bloom,
         };
         let options = options_from_fulltext(&fulltext).unwrap().unwrap();
         assert_eq!(
             options.options.get(FULLTEXT_GRPC_KEY).unwrap(),
-            "{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false}"
+            "{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false,\"backend\":\"bloom\"}"
         );
     }
 

src/catalog/src/system_schema/information_schema.rs

@@ -19,7 +19,7 @@ mod information_memory_table;
 pub mod key_column_usage;
 mod partitions;
 mod procedure_info;
-mod region_peers;
+pub mod region_peers;
 mod region_statistics;
 mod runtime_metrics;
 pub mod schemata;
@@ -49,7 +49,6 @@ pub use table_names::*;
 use views::InformationSchemaViews;
 
 use self::columns::InformationSchemaColumns;
-use super::{SystemSchemaProviderInner, SystemTable, SystemTableRef};
 use crate::error::{Error, Result};
 use crate::system_schema::information_schema::cluster_info::InformationSchemaClusterInfo;
 use crate::system_schema::information_schema::flows::InformationSchemaFlows;
@@ -63,7 +62,9 @@ use crate::system_schema::information_schema::table_constraints::InformationSche
 use crate::system_schema::information_schema::tables::InformationSchemaTables;
 use crate::system_schema::memory_table::MemoryTable;
 pub(crate) use crate::system_schema::predicate::Predicates;
-use crate::system_schema::SystemSchemaProvider;
+use crate::system_schema::{
+    SystemSchemaProvider, SystemSchemaProviderInner, SystemTable, SystemTableRef,
+};
 use crate::CatalogManager;
 
 lazy_static! {

src/catalog/src/system_schema/information_schema/cluster_info.rs

@@ -36,9 +36,8 @@ use datatypes::vectors::{
 use snafu::ResultExt;
 use store_api::storage::{ScanRequest, TableId};
 
-use super::CLUSTER_INFO;
 use crate::error::{CreateRecordBatchSnafu, InternalSnafu, Result};
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, CLUSTER_INFO};
 use crate::system_schema::utils;
 use crate::CatalogManager;
 

src/catalog/src/system_schema/information_schema/columns.rs

@@ -38,11 +38,11 @@ use snafu::{OptionExt, ResultExt};
 use sql::statements;
 use store_api::storage::{ScanRequest, TableId};
 
-use super::{InformationTable, COLUMNS};
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::information_schema::Predicates;
+use crate::system_schema::information_schema::{InformationTable, COLUMNS};
 use crate::CatalogManager;
 
 #[derive(Debug)]
@@ -56,6 +56,8 @@ pub const TABLE_CATALOG: &str = "table_catalog";
 pub const TABLE_SCHEMA: &str = "table_schema";
 pub const TABLE_NAME: &str = "table_name";
 pub const COLUMN_NAME: &str = "column_name";
+pub const REGION_ID: &str = "region_id";
+pub const PEER_ID: &str = "peer_id";
 const ORDINAL_POSITION: &str = "ordinal_position";
 const CHARACTER_MAXIMUM_LENGTH: &str = "character_maximum_length";
 const CHARACTER_OCTET_LENGTH: &str = "character_octet_length";

src/catalog/src/system_schema/information_schema/information_memory_table.rs

@@ -18,7 +18,7 @@ use common_catalog::consts::{METRIC_ENGINE, MITO_ENGINE};
 use datatypes::schema::{Schema, SchemaRef};
 use datatypes::vectors::{Int64Vector, StringVector, VectorRef};
 
-use super::table_names::*;
+use crate::system_schema::information_schema::table_names::*;
 use crate::system_schema::utils::tables::{
     bigint_column, string_column, string_columns, timestamp_micro_column,
 };

src/catalog/src/system_schema/information_schema/key_column_usage.rs

@@ -24,18 +24,17 @@ use datafusion::physical_plan::stream::RecordBatchStreamAdapter as DfRecordBatch
 use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
 use datafusion::physical_plan::SendableRecordBatchStream as DfSendableRecordBatchStream;
 use datatypes::prelude::{ConcreteDataType, MutableVector, ScalarVectorBuilder, VectorRef};
-use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
+use datatypes::schema::{ColumnSchema, FulltextBackend, Schema, SchemaRef};
 use datatypes::value::Value;
 use datatypes::vectors::{ConstantVector, StringVector, StringVectorBuilder, UInt32VectorBuilder};
 use futures_util::TryStreamExt;
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 
-use super::KEY_COLUMN_USAGE;
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, KEY_COLUMN_USAGE};
 use crate::CatalogManager;
 
 pub const CONSTRAINT_SCHEMA: &str = "constraint_schema";
@@ -48,20 +47,38 @@ pub const TABLE_SCHEMA: &str = "table_schema";
 pub const TABLE_NAME: &str = "table_name";
 pub const COLUMN_NAME: &str = "column_name";
 pub const ORDINAL_POSITION: &str = "ordinal_position";
+/// The type of the index.
+pub const GREPTIME_INDEX_TYPE: &str = "greptime_index_type";
 const INIT_CAPACITY: usize = 42;
 
-/// Primary key constraint name
-pub(crate) const PRI_CONSTRAINT_NAME: &str = "PRIMARY";
 /// Time index constraint name
-pub(crate) const TIME_INDEX_CONSTRAINT_NAME: &str = "TIME INDEX";
+pub(crate) const CONSTRAINT_NAME_TIME_INDEX: &str = "TIME INDEX";
+
+/// Primary key constraint name
+pub(crate) const CONSTRAINT_NAME_PRI: &str = "PRIMARY";
+/// Primary key index type
+pub(crate) const INDEX_TYPE_PRI: &str = "greptime-primary-key-v1";
+
 /// Inverted index constraint name
-pub(crate) const INVERTED_INDEX_CONSTRAINT_NAME: &str = "INVERTED INDEX";
+pub(crate) const CONSTRAINT_NAME_INVERTED_INDEX: &str = "INVERTED INDEX";
+/// Inverted index type
+pub(crate) const INDEX_TYPE_INVERTED_INDEX: &str = "greptime-inverted-index-v1";
+
 /// Fulltext index constraint name
-pub(crate) const FULLTEXT_INDEX_CONSTRAINT_NAME: &str = "FULLTEXT INDEX";
+pub(crate) const CONSTRAINT_NAME_FULLTEXT_INDEX: &str = "FULLTEXT INDEX";
+/// Fulltext index v1 type
+pub(crate) const INDEX_TYPE_FULLTEXT_TANTIVY: &str = "greptime-fulltext-index-v1";
+/// Fulltext index bloom type
+pub(crate) const INDEX_TYPE_FULLTEXT_BLOOM: &str = "greptime-fulltext-index-bloom";
+
 /// Skipping index constraint name
-pub(crate) const SKIPPING_INDEX_CONSTRAINT_NAME: &str = "SKIPPING INDEX";
+pub(crate) const CONSTRAINT_NAME_SKIPPING_INDEX: &str = "SKIPPING INDEX";
+/// Skipping index type
+pub(crate) const INDEX_TYPE_SKIPPING_INDEX: &str = "greptime-bloom-filter-v1";
 
 /// The virtual table implementation for `information_schema.KEY_COLUMN_USAGE`.
+///
+/// Provides an extra column `greptime_index_type` for the index type of the key column.
 #[derive(Debug)]
 pub(super) struct InformationSchemaKeyColumnUsage {
     schema: SchemaRef,
@@ -121,6 +138,11 @@ impl InformationSchemaKeyColumnUsage {
                 ConcreteDataType::string_datatype(),
                 true,
             ),
+            ColumnSchema::new(
+                GREPTIME_INDEX_TYPE,
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
         ]))
     }
 
@@ -185,6 +207,7 @@ struct InformationSchemaKeyColumnUsageBuilder {
     column_name: StringVectorBuilder,
     ordinal_position: UInt32VectorBuilder,
     position_in_unique_constraint: UInt32VectorBuilder,
+    greptime_index_type: StringVectorBuilder,
 }
 
 impl InformationSchemaKeyColumnUsageBuilder {
@@ -207,6 +230,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
             column_name: StringVectorBuilder::with_capacity(INIT_CAPACITY),
             ordinal_position: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
             position_in_unique_constraint: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
+            greptime_index_type: StringVectorBuilder::with_capacity(INIT_CAPACITY),
         }
     }
 
@@ -230,34 +254,47 @@ impl InformationSchemaKeyColumnUsageBuilder {
 
                 for (idx, column) in schema.column_schemas().iter().enumerate() {
                     let mut constraints = vec![];
+                    let mut greptime_index_type = vec![];
                     if column.is_time_index() {
                         self.add_key_column_usage(
                             &predicates,
                             &schema_name,
-                            TIME_INDEX_CONSTRAINT_NAME,
+                            CONSTRAINT_NAME_TIME_INDEX,
                             &catalog_name,
                             &schema_name,
                             table_name,
                             &column.name,
                             1, //always 1 for time index
+                            "",
                         );
                     }
                     // TODO(dimbtp): foreign key constraint not supported yet
                     if keys.contains(&idx) {
-                        constraints.push(PRI_CONSTRAINT_NAME);
+                        constraints.push(CONSTRAINT_NAME_PRI);
+                        greptime_index_type.push(INDEX_TYPE_PRI);
                     }
                     if column.is_inverted_indexed() {
-                        constraints.push(INVERTED_INDEX_CONSTRAINT_NAME);
+                        constraints.push(CONSTRAINT_NAME_INVERTED_INDEX);
+                        greptime_index_type.push(INDEX_TYPE_INVERTED_INDEX);
                     }
-                    if column.is_fulltext_indexed() {
-                        constraints.push(FULLTEXT_INDEX_CONSTRAINT_NAME);
+                    if let Ok(Some(options)) = column.fulltext_options() {
+                        if options.enable {
+                            constraints.push(CONSTRAINT_NAME_FULLTEXT_INDEX);
+                            let index_type = match options.backend {
+                                FulltextBackend::Bloom => INDEX_TYPE_FULLTEXT_BLOOM,
+                                FulltextBackend::Tantivy => INDEX_TYPE_FULLTEXT_TANTIVY,
+                            };
+                            greptime_index_type.push(index_type);
+                        }
                     }
                     if column.is_skipping_indexed() {
-                        constraints.push(SKIPPING_INDEX_CONSTRAINT_NAME);
+                        constraints.push(CONSTRAINT_NAME_SKIPPING_INDEX);
+                        greptime_index_type.push(INDEX_TYPE_SKIPPING_INDEX);
                     }
 
                     if !constraints.is_empty() {
                         let aggregated_constraints = constraints.join(", ");
+                        let aggregated_index_types = greptime_index_type.join(", ");
                         self.add_key_column_usage(
                             &predicates,
                             &schema_name,
@@ -267,6 +304,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
                             table_name,
                             &column.name,
                             idx as u32 + 1,
+                            &aggregated_index_types,
                         );
                     }
                 }
@@ -289,6 +327,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
         table_name: &str,
         column_name: &str,
         ordinal_position: u32,
+        index_types: &str,
     ) {
         let row = [
             (CONSTRAINT_SCHEMA, &Value::from(constraint_schema)),
@@ -298,6 +337,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
             (TABLE_NAME, &Value::from(table_name)),
             (COLUMN_NAME, &Value::from(column_name)),
             (ORDINAL_POSITION, &Value::from(ordinal_position)),
+            (GREPTIME_INDEX_TYPE, &Value::from(index_types)),
         ];
 
         if !predicates.eval(&row) {
@@ -314,6 +354,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
         self.column_name.push(Some(column_name));
         self.ordinal_position.push(Some(ordinal_position));
         self.position_in_unique_constraint.push(None);
+        self.greptime_index_type.push(Some(index_types));
     }
 
     fn finish(&mut self) -> Result<RecordBatch> {
@@ -337,6 +378,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
             null_string_vector.clone(),
             null_string_vector.clone(),
             null_string_vector,
+            Arc::new(self.greptime_index_type.finish()),
         ];
         RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)
     }

src/catalog/src/system_schema/information_schema/partitions.rs

@@ -39,13 +39,12 @@ use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 use table::metadata::{TableInfo, TableType};
 
-use super::PARTITIONS;
 use crate::error::{
     CreateRecordBatchSnafu, FindPartitionsSnafu, InternalSnafu, PartitionManagerNotFoundSnafu,
     Result, UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::kvbackend::KvBackendCatalogManager;
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, PARTITIONS};
 use crate::CatalogManager;
 
 const TABLE_CATALOG: &str = "table_catalog";

src/catalog/src/system_schema/information_schema/procedure_info.rs

@@ -33,9 +33,8 @@ use datatypes::vectors::{StringVectorBuilder, TimestampMillisecondVectorBuilder}
 use snafu::ResultExt;
 use store_api::storage::{ScanRequest, TableId};
 
-use super::PROCEDURE_INFO;
 use crate::error::{CreateRecordBatchSnafu, InternalSnafu, Result};
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, PROCEDURE_INFO};
 use crate::system_schema::utils;
 use crate::CatalogManager;
 

src/catalog/src/system_schema/information_schema/region_peers.rs

@@ -21,6 +21,7 @@ use common_error::ext::BoxedError;
 use common_meta::rpc::router::RegionRoute;
 use common_recordbatch::adapter::RecordBatchStreamAdapter;
 use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
+use datafusion::common::HashMap;
 use datafusion::execution::TaskContext;
 use datafusion::physical_plan::stream::RecordBatchStreamAdapter as DfRecordBatchStreamAdapter;
 use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
@@ -34,25 +35,30 @@ use snafu::{OptionExt, ResultExt};
 use store_api::storage::{RegionId, ScanRequest, TableId};
 use table::metadata::TableType;
 
-use super::REGION_PEERS;
 use crate::error::{
     CreateRecordBatchSnafu, FindRegionRoutesSnafu, InternalSnafu, Result,
     UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::kvbackend::KvBackendCatalogManager;
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, REGION_PEERS};
 use crate::CatalogManager;
 
-const REGION_ID: &str = "region_id";
-const PEER_ID: &str = "peer_id";
+pub const TABLE_CATALOG: &str = "table_catalog";
+pub const TABLE_SCHEMA: &str = "table_schema";
+pub const TABLE_NAME: &str = "table_name";
+pub const REGION_ID: &str = "region_id";
+pub const PEER_ID: &str = "peer_id";
 const PEER_ADDR: &str = "peer_addr";
-const IS_LEADER: &str = "is_leader";
+pub const IS_LEADER: &str = "is_leader";
 const STATUS: &str = "status";
 const DOWN_SECONDS: &str = "down_seconds";
 const INIT_CAPACITY: usize = 42;
 
 /// The `REGION_PEERS` table provides information about the region distribution and routes. Including fields:
 ///
+/// - `table_catalog`: the table catalog name
+/// - `table_schema`: the table schema name
+/// - `table_name`: the table name
 /// - `region_id`: the region id
 /// - `peer_id`: the region storage datanode peer id
 /// - `peer_addr`: the region storage datanode gRPC peer address
@@ -77,6 +83,9 @@ impl InformationSchemaRegionPeers {
 
     pub(crate) fn schema() -> SchemaRef {
         Arc::new(Schema::new(vec![
+            ColumnSchema::new(TABLE_CATALOG, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(TABLE_SCHEMA, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(TABLE_NAME, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(REGION_ID, ConcreteDataType::uint64_datatype(), false),
             ColumnSchema::new(PEER_ID, ConcreteDataType::uint64_datatype(), true),
             ColumnSchema::new(PEER_ADDR, ConcreteDataType::string_datatype(), true),
@@ -134,6 +143,9 @@ struct InformationSchemaRegionPeersBuilder {
     catalog_name: String,
     catalog_manager: Weak<dyn CatalogManager>,
 
+    table_catalogs: StringVectorBuilder,
+    table_schemas: StringVectorBuilder,
+    table_names: StringVectorBuilder,
     region_ids: UInt64VectorBuilder,
     peer_ids: UInt64VectorBuilder,
     peer_addrs: StringVectorBuilder,
@@ -152,6 +164,9 @@ impl InformationSchemaRegionPeersBuilder {
             schema,
             catalog_name,
             catalog_manager,
+            table_catalogs: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_schemas: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
             region_ids: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
             peer_ids: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
             peer_addrs: StringVectorBuilder::with_capacity(INIT_CAPACITY),
@@ -177,24 +192,28 @@ impl InformationSchemaRegionPeersBuilder {
         let predicates = Predicates::from_scan_request(&request);
 
         for schema_name in catalog_manager.schema_names(&catalog_name, None).await? {
-            let table_id_stream = catalog_manager
+            let table_stream = catalog_manager
                 .tables(&catalog_name, &schema_name, None)
                 .try_filter_map(|t| async move {
                     let table_info = t.table_info();
                     if table_info.table_type == TableType::Temporary {
                         Ok(None)
                     } else {
-                        Ok(Some(table_info.ident.table_id))
+                        Ok(Some((
+                            table_info.ident.table_id,
+                            table_info.name.to_string(),
+                        )))
                     }
                 });
 
             const BATCH_SIZE: usize = 128;
 
-            // Split table ids into chunks
-            let mut table_id_chunks = pin!(table_id_stream.ready_chunks(BATCH_SIZE));
+            // Split tables into chunks
+            let mut table_chunks = pin!(table_stream.ready_chunks(BATCH_SIZE));
 
-            while let Some(table_ids) = table_id_chunks.next().await {
-                let table_ids = table_ids.into_iter().collect::<Result<Vec<_>>>()?;
+            while let Some(tables) = table_chunks.next().await {
+                let tables = tables.into_iter().collect::<Result<HashMap<_, _>>>()?;
+                let table_ids = tables.keys().cloned().collect::<Vec<_>>();
 
                 let table_routes = if let Some(partition_manager) = &partition_manager {
                     partition_manager
@@ -206,7 +225,16 @@ impl InformationSchemaRegionPeersBuilder {
                 };
 
                 for (table_id, routes) in table_routes {
-                    self.add_region_peers(&predicates, table_id, &routes);
+                    // Safety: table_id is guaranteed to be in the map
+                    let table_name = tables.get(&table_id).unwrap();
+                    self.add_region_peers(
+                        &catalog_name,
+                        &schema_name,
+                        table_name,
+                        &predicates,
+                        table_id,
+                        &routes,
+                    );
                 }
             }
         }
@@ -216,6 +244,9 @@ impl InformationSchemaRegionPeersBuilder {
 
     fn add_region_peers(
         &mut self,
+        table_catalog: &str,
+        table_schema: &str,
+        table_name: &str,
         predicates: &Predicates,
         table_id: TableId,
         routes: &[RegionRoute],
@@ -231,13 +262,20 @@ impl InformationSchemaRegionPeersBuilder {
                 Some("ALIVE".to_string())
             };
 
-            let row = [(REGION_ID, &Value::from(region_id))];
+            let row = [
+                (TABLE_CATALOG, &Value::from(table_catalog)),
+                (TABLE_SCHEMA, &Value::from(table_schema)),
+                (TABLE_NAME, &Value::from(table_name)),
+                (REGION_ID, &Value::from(region_id)),
+            ];
 
             if !predicates.eval(&row) {
                 return;
             }
 
-            // TODO(dennis): adds followers.
+            self.table_catalogs.push(Some(table_catalog));
+            self.table_schemas.push(Some(table_schema));
+            self.table_names.push(Some(table_name));
             self.region_ids.push(Some(region_id));
             self.peer_ids.push(peer_id);
             self.peer_addrs.push(peer_addr.as_deref());
@@ -245,11 +283,26 @@ impl InformationSchemaRegionPeersBuilder {
             self.statuses.push(state.as_deref());
             self.down_seconds
                 .push(route.leader_down_millis().map(|m| m / 1000));
+
+            for follower in &route.follower_peers {
+                self.table_catalogs.push(Some(table_catalog));
+                self.table_schemas.push(Some(table_schema));
+                self.table_names.push(Some(table_name));
+                self.region_ids.push(Some(region_id));
+                self.peer_ids.push(Some(follower.id));
+                self.peer_addrs.push(Some(follower.addr.as_str()));
+                self.is_leaders.push(Some("No"));
+                self.statuses.push(None);
+                self.down_seconds.push(None);
+            }
         }
     }
 
     fn finish(&mut self) -> Result<RecordBatch> {
         let columns: Vec<VectorRef> = vec![
+            Arc::new(self.table_catalogs.finish()),
+            Arc::new(self.table_schemas.finish()),
+            Arc::new(self.table_names.finish()),
             Arc::new(self.region_ids.finish()),
             Arc::new(self.peer_ids.finish()),
             Arc::new(self.peer_addrs.finish()),

src/catalog/src/system_schema/information_schema/region_statistics.rs

@@ -30,9 +30,9 @@ use datatypes::vectors::{StringVectorBuilder, UInt32VectorBuilder, UInt64VectorB
 use snafu::ResultExt;
 use store_api::storage::{ScanRequest, TableId};
 
-use super::{InformationTable, REGION_STATISTICS};
 use crate::error::{CreateRecordBatchSnafu, InternalSnafu, Result};
 use crate::information_schema::Predicates;
+use crate::system_schema::information_schema::{InformationTable, REGION_STATISTICS};
 use crate::system_schema::utils;
 use crate::CatalogManager;
 

src/catalog/src/system_schema/information_schema/runtime_metrics.rs

@@ -35,8 +35,8 @@ use itertools::Itertools;
 use snafu::ResultExt;
 use store_api::storage::{ScanRequest, TableId};
 
-use super::{InformationTable, RUNTIME_METRICS};
 use crate::error::{CreateRecordBatchSnafu, InternalSnafu, Result};
+use crate::system_schema::information_schema::{InformationTable, RUNTIME_METRICS};
 
 #[derive(Debug)]
 pub(super) struct InformationSchemaMetrics {

src/catalog/src/system_schema/information_schema/schemata.rs

@@ -31,12 +31,11 @@ use datatypes::vectors::StringVectorBuilder;
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 
-use super::SCHEMATA;
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, TableMetadataManagerSnafu,
     UpgradeWeakCatalogManagerRefSnafu,
 };
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, SCHEMATA};
 use crate::system_schema::utils;
 use crate::CatalogManager;
 

src/catalog/src/system_schema/information_schema/table_constraints.rs

@@ -32,14 +32,14 @@ use futures::TryStreamExt;
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 
-use super::{InformationTable, TABLE_CONSTRAINTS};
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::information_schema::key_column_usage::{
-    PRI_CONSTRAINT_NAME, TIME_INDEX_CONSTRAINT_NAME,
+    CONSTRAINT_NAME_PRI, CONSTRAINT_NAME_TIME_INDEX,
 };
 use crate::information_schema::Predicates;
+use crate::system_schema::information_schema::{InformationTable, TABLE_CONSTRAINTS};
 use crate::CatalogManager;
 
 /// The `TABLE_CONSTRAINTS` table describes which tables have constraints.
@@ -188,7 +188,7 @@ impl InformationSchemaTableConstraintsBuilder {
                     self.add_table_constraint(
                         &predicates,
                         &schema_name,
-                        TIME_INDEX_CONSTRAINT_NAME,
+                        CONSTRAINT_NAME_TIME_INDEX,
                         &schema_name,
                         &table.table_info().name,
                         TIME_INDEX_CONSTRAINT_TYPE,
@@ -199,7 +199,7 @@ impl InformationSchemaTableConstraintsBuilder {
                     self.add_table_constraint(
                         &predicates,
                         &schema_name,
-                        PRI_CONSTRAINT_NAME,
+                        CONSTRAINT_NAME_PRI,
                         &schema_name,
                         &table.table_info().name,
                         PRI_KEY_CONSTRAINT_TYPE,

src/catalog/src/system_schema/information_schema/tables.rs

@@ -38,11 +38,10 @@ use snafu::{OptionExt, ResultExt};
 use store_api::storage::{RegionId, ScanRequest, TableId};
 use table::metadata::{TableInfo, TableType};
 
-use super::TABLES;
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, TABLES};
 use crate::system_schema::utils;
 use crate::CatalogManager;
 

src/catalog/src/system_schema/information_schema/views.rs

@@ -32,13 +32,12 @@ use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 use table::metadata::TableType;
 
-use super::VIEWS;
 use crate::error::{
     CastManagerSnafu, CreateRecordBatchSnafu, GetViewCacheSnafu, InternalSnafu, Result,
     UpgradeWeakCatalogManagerRefSnafu, ViewInfoNotFoundSnafu,
 };
 use crate::kvbackend::KvBackendCatalogManager;
-use crate::system_schema::information_schema::{InformationTable, Predicates};
+use crate::system_schema::information_schema::{InformationTable, Predicates, VIEWS};
 use crate::CatalogManager;
 const INIT_CAPACITY: usize = 42;
 

src/catalog/src/system_schema/memory_table.rs

@@ -29,8 +29,8 @@ use datatypes::vectors::VectorRef;
 use snafu::ResultExt;
 use store_api::storage::{ScanRequest, TableId};
 
-use super::SystemTable;
 use crate::error::{CreateRecordBatchSnafu, InternalSnafu, Result};
+use crate::system_schema::SystemTable;
 
 /// A memory table with specified schema and columns.
 #[derive(Debug)]

src/catalog/src/system_schema/pg_catalog.rs

@@ -34,9 +34,9 @@ use table::TableRef;
 pub use table_names::*;
 
 use self::pg_namespace::oid_map::{PGNamespaceOidMap, PGNamespaceOidMapRef};
-use super::memory_table::MemoryTable;
-use super::utils::tables::u32_column;
-use super::{SystemSchemaProvider, SystemSchemaProviderInner, SystemTableRef};
+use crate::system_schema::memory_table::MemoryTable;
+use crate::system_schema::utils::tables::u32_column;
+use crate::system_schema::{SystemSchemaProvider, SystemSchemaProviderInner, SystemTableRef};
 use crate::CatalogManager;
 
 lazy_static! {

src/catalog/src/system_schema/pg_catalog/pg_catalog_memory_table.rs

@@ -17,9 +17,9 @@ use std::sync::Arc;
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
 use datatypes::vectors::{Int16Vector, StringVector, UInt32Vector, VectorRef};
 
-use super::oid_column;
-use super::table_names::PG_TYPE;
 use crate::memory_table_cols;
+use crate::system_schema::pg_catalog::oid_column;
+use crate::system_schema::pg_catalog::table_names::PG_TYPE;
 use crate::system_schema::utils::tables::{i16_column, string_column};
 
 fn pg_type_schema_columns() -> (Vec<ColumnSchema>, Vec<VectorRef>) {

src/catalog/src/system_schema/pg_catalog/pg_class.rs

@@ -32,12 +32,12 @@ use snafu::{OptionExt, ResultExt};
 use store_api::storage::ScanRequest;
 use table::metadata::TableType;
 
-use super::pg_namespace::oid_map::PGNamespaceOidMapRef;
-use super::{query_ctx, OID_COLUMN_NAME, PG_CLASS};
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::information_schema::Predicates;
+use crate::system_schema::pg_catalog::pg_namespace::oid_map::PGNamespaceOidMapRef;
+use crate::system_schema::pg_catalog::{query_ctx, OID_COLUMN_NAME, PG_CLASS};
 use crate::system_schema::utils::tables::{string_column, u32_column};
 use crate::system_schema::SystemTable;
 use crate::CatalogManager;

src/catalog/src/system_schema/pg_catalog/pg_database.rs

@@ -29,12 +29,12 @@ use datatypes::vectors::{StringVectorBuilder, UInt32VectorBuilder, VectorRef};
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::ScanRequest;
 
-use super::pg_namespace::oid_map::PGNamespaceOidMapRef;
-use super::{query_ctx, OID_COLUMN_NAME, PG_DATABASE};
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::information_schema::Predicates;
+use crate::system_schema::pg_catalog::pg_namespace::oid_map::PGNamespaceOidMapRef;
+use crate::system_schema::pg_catalog::{query_ctx, OID_COLUMN_NAME, PG_DATABASE};
 use crate::system_schema::utils::tables::{string_column, u32_column};
 use crate::system_schema::SystemTable;
 use crate::CatalogManager;

src/catalog/src/system_schema/pg_catalog/pg_namespace.rs

@@ -35,11 +35,13 @@ use datatypes::vectors::{StringVectorBuilder, UInt32VectorBuilder, VectorRef};
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::ScanRequest;
 
-use super::{query_ctx, PGNamespaceOidMapRef, OID_COLUMN_NAME, PG_NAMESPACE};
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::information_schema::Predicates;
+use crate::system_schema::pg_catalog::{
+    query_ctx, PGNamespaceOidMapRef, OID_COLUMN_NAME, PG_NAMESPACE,
+};
 use crate::system_schema::utils::tables::{string_column, u32_column};
 use crate::system_schema::SystemTable;
 use crate::CatalogManager;

src/catalog/src/system_schema/predicate.rs

@@ -437,10 +437,7 @@ mod tests {
     }
 
     fn column(name: &str) -> Expr {
-        Expr::Column(Column {
-            relation: None,
-            name: name.to_string(),
-        })
+        Expr::Column(Column::from_name(name))
     }
 
     fn string_literal(v: &str) -> Expr {

src/cli/src/bench.rs

@@ -177,7 +177,7 @@ fn create_table_info(table_id: TableId, table_name: TableName) -> RawTableInfo {
 
 fn create_region_routes(regions: Vec<RegionNumber>) -> Vec<RegionRoute> {
     let mut region_routes = Vec::with_capacity(100);
-    let mut rng = rand::thread_rng();
+    let mut rng = rand::rng();
 
     for region_id in regions.into_iter().map(u64::from) {
         region_routes.push(RegionRoute {
@@ -188,7 +188,7 @@ fn create_region_routes(regions: Vec<RegionNumber>) -> Vec<RegionRoute> {
                 attrs: BTreeMap::new(),
             },
             leader_peer: Some(Peer {
-                id: rng.gen_range(0..10),
+                id: rng.random_range(0..10),
                 addr: String::new(),
             }),
             follower_peers: vec![],

src/cli/src/error.rs

@@ -17,7 +17,6 @@ use std::any::Any;
 use common_error::ext::{BoxedError, ErrorExt};
 use common_error::status_code::StatusCode;
 use common_macro::stack_trace_debug;
-use rustyline::error::ReadlineError;
 use snafu::{Location, Snafu};
 
 #[derive(Snafu)]
@@ -105,52 +104,6 @@ pub enum Error {
     #[snafu(display("Invalid REPL command: {reason}"))]
     InvalidReplCommand { reason: String },
 
-    #[snafu(display("Cannot create REPL"))]
-    ReplCreation {
-        #[snafu(source)]
-        error: ReadlineError,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
-    #[snafu(display("Error reading command"))]
-    Readline {
-        #[snafu(source)]
-        error: ReadlineError,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
-    #[snafu(display("Failed to request database, sql: {sql}"))]
-    RequestDatabase {
-        sql: String,
-        #[snafu(source)]
-        source: client::Error,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
-    #[snafu(display("Failed to collect RecordBatches"))]
-    CollectRecordBatches {
-        #[snafu(implicit)]
-        location: Location,
-        source: common_recordbatch::error::Error,
-    },
-
-    #[snafu(display("Failed to pretty print Recordbatches"))]
-    PrettyPrintRecordBatches {
-        #[snafu(implicit)]
-        location: Location,
-        source: common_recordbatch::error::Error,
-    },
-
-    #[snafu(display("Failed to start Meta client"))]
-    StartMetaClient {
-        #[snafu(implicit)]
-        location: Location,
-        source: meta_client::error::Error,
-    },
-
     #[snafu(display("Failed to parse SQL: {}", sql))]
     ParseSql {
         sql: String,
@@ -166,13 +119,6 @@ pub enum Error {
         source: query::error::Error,
     },
 
-    #[snafu(display("Failed to encode logical plan in substrait"))]
-    SubstraitEncodeLogicalPlan {
-        #[snafu(implicit)]
-        location: Location,
-        source: substrait::error::Error,
-    },
-
     #[snafu(display("Failed to load layered config"))]
     LoadLayeredConfig {
         #[snafu(source(from(common_config::error::Error, Box::new)))]
@@ -318,17 +264,10 @@ impl ErrorExt for Error {
             Error::StartProcedureManager { source, .. }
             | Error::StopProcedureManager { source, .. } => source.status_code(),
             Error::StartWalOptionsAllocator { source, .. } => source.status_code(),
-            Error::ReplCreation { .. } | Error::Readline { .. } | Error::HttpQuerySql { .. } => {
-                StatusCode::Internal
-            }
-            Error::RequestDatabase { source, .. } => source.status_code(),
-            Error::CollectRecordBatches { source, .. }
-            | Error::PrettyPrintRecordBatches { source, .. } => source.status_code(),
-            Error::StartMetaClient { source, .. } => source.status_code(),
+            Error::HttpQuerySql { .. } => StatusCode::Internal,
             Error::ParseSql { source, .. } | Error::PlanStatement { source, .. } => {
                 source.status_code()
             }
-            Error::SubstraitEncodeLogicalPlan { source, .. } => source.status_code(),
 
             Error::SerdeJson { .. }
             | Error::FileIo { .. }

src/cli/src/lib.rs

@@ -23,15 +23,12 @@ mod helper;
 // Wait for https://github.com/GreptimeTeam/greptimedb/issues/2373
 mod database;
 mod import;
-#[allow(unused)]
-mod repl;
 
 use async_trait::async_trait;
 use clap::Parser;
 use common_error::ext::BoxedError;
 pub use database::DatabaseClient;
 use error::Result;
-pub use repl::Repl;
 
 pub use crate::bench::BenchTableMetadataCommand;
 pub use crate::export::ExportCommand;

src/cli/src/repl.rs

@@ -1,299 +0,0 @@
-// Copyright 2023 Greptime Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::path::PathBuf;
-use std::sync::Arc;
-use std::time::Instant;
-
-use cache::{
-    build_fundamental_cache_registry, with_default_composite_cache_registry, TABLE_CACHE_NAME,
-    TABLE_ROUTE_CACHE_NAME,
-};
-use catalog::information_extension::DistributedInformationExtension;
-use catalog::kvbackend::{
-    CachedKvBackend, CachedKvBackendBuilder, KvBackendCatalogManager, MetaKvBackend,
-};
-use client::{Client, Database, OutputData, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
-use common_base::Plugins;
-use common_config::Mode;
-use common_error::ext::ErrorExt;
-use common_meta::cache::{CacheRegistryBuilder, LayeredCacheRegistryBuilder};
-use common_meta::kv_backend::KvBackendRef;
-use common_query::Output;
-use common_recordbatch::RecordBatches;
-use common_telemetry::debug;
-use either::Either;
-use meta_client::client::{ClusterKvBackend, MetaClientBuilder};
-use query::datafusion::DatafusionQueryEngine;
-use query::parser::QueryLanguageParser;
-use query::query_engine::{DefaultSerializer, QueryEngineState};
-use query::QueryEngine;
-use rustyline::error::ReadlineError;
-use rustyline::Editor;
-use session::context::QueryContext;
-use snafu::{OptionExt, ResultExt};
-use substrait::{DFLogicalSubstraitConvertor, SubstraitPlan};
-
-use crate::cmd::ReplCommand;
-use crate::error::{
-    CollectRecordBatchesSnafu, ParseSqlSnafu, PlanStatementSnafu, PrettyPrintRecordBatchesSnafu,
-    ReadlineSnafu, ReplCreationSnafu, RequestDatabaseSnafu, Result, StartMetaClientSnafu,
-    SubstraitEncodeLogicalPlanSnafu,
-};
-use crate::helper::RustylineHelper;
-use crate::{error, AttachCommand};
-
-/// Captures the state of the repl, gathers commands and executes them one by one
-pub struct Repl {
-    /// Rustyline editor for interacting with user on command line
-    rl: Editor<RustylineHelper>,
-
-    /// Current prompt
-    prompt: String,
-
-    /// Client for interacting with GreptimeDB
-    database: Database,
-
-    query_engine: Option<DatafusionQueryEngine>,
-}
-
-#[allow(clippy::print_stdout)]
-impl Repl {
-    fn print_help(&self) {
-        println!("{}", ReplCommand::help())
-    }
-
-    pub(crate) async fn try_new(cmd: &AttachCommand) -> Result<Self> {
-        let mut rl = Editor::new().context(ReplCreationSnafu)?;
-
-        if !cmd.disable_helper {
-            rl.set_helper(Some(RustylineHelper::default()));
-
-            let history_file = history_file();
-            if let Err(e) = rl.load_history(&history_file) {
-                debug!(
-                    "failed to load history file on {}, error: {e}",
-                    history_file.display()
-                );
-            }
-        }
-
-        let client = Client::with_urls([&cmd.grpc_addr]);
-        let database = Database::new(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, client);
-
-        let query_engine = if let Some(meta_addr) = &cmd.meta_addr {
-            create_query_engine(meta_addr).await.map(Some)?
-        } else {
-            None
-        };
-
-        Ok(Self {
-            rl,
-            prompt: "> ".to_string(),
-            database,
-            query_engine,
-        })
-    }
-
-    /// Parse the next command
-    fn next_command(&mut self) -> Result<ReplCommand> {
-        match self.rl.readline(&self.prompt) {
-            Ok(ref line) => {
-                let request = line.trim();
-
-                let _ = self.rl.add_history_entry(request.to_string());
-
-                request.try_into()
-            }
-            Err(ReadlineError::Eof) | Err(ReadlineError::Interrupted) => Ok(ReplCommand::Exit),
-            // Some sort of real underlying error
-            Err(e) => Err(e).context(ReadlineSnafu),
-        }
-    }
-
-    /// Read Evaluate Print Loop (interactive command line) for GreptimeDB
-    ///
-    /// Inspired / based on repl.rs from InfluxDB IOX
-    pub(crate) async fn run(&mut self) -> Result<()> {
-        println!("Ready for commands. (Hint: try 'help')");
-
-        loop {
-            match self.next_command()? {
-                ReplCommand::Help => {
-                    self.print_help();
-                }
-                ReplCommand::UseDatabase { db_name } => {
-                    if self.execute_sql(format!("USE {db_name}")).await {
-                        println!("Using {db_name}");
-                        self.database.set_schema(&db_name);
-                        self.prompt = format!("[{db_name}] > ");
-                    }
-                }
-                ReplCommand::Sql { sql } => {
-                    let _ = self.execute_sql(sql).await;
-                }
-                ReplCommand::Exit => {
-                    return Ok(());
-                }
-            }
-        }
-    }
-
-    async fn execute_sql(&self, sql: String) -> bool {
-        self.do_execute_sql(sql)
-            .await
-            .map_err(|e| {
-                let status_code = e.status_code();
-                let root_cause = e.output_msg();
-                println!("Error: {}({status_code}), {root_cause}", status_code as u32)
-            })
-            .is_ok()
-    }
-
-    async fn do_execute_sql(&self, sql: String) -> Result<()> {
-        let start = Instant::now();
-
-        let output = if let Some(query_engine) = &self.query_engine {
-            let query_ctx = Arc::new(QueryContext::with(
-                self.database.catalog(),
-                self.database.schema(),
-            ));
-
-            let stmt = QueryLanguageParser::parse_sql(&sql, &query_ctx)
-                .with_context(|_| ParseSqlSnafu { sql: sql.clone() })?;
-
-            let plan = query_engine
-                .planner()
-                .plan(&stmt, query_ctx.clone())
-                .await
-                .context(PlanStatementSnafu)?;
-
-            let plan = query_engine
-                .optimize(&query_engine.engine_context(query_ctx), &plan)
-                .context(PlanStatementSnafu)?;
-
-            let plan = DFLogicalSubstraitConvertor {}
-                .encode(&plan, DefaultSerializer)
-                .context(SubstraitEncodeLogicalPlanSnafu)?;
-
-            self.database.logical_plan(plan.to_vec()).await
-        } else {
-            self.database.sql(&sql).await
-        }
-        .context(RequestDatabaseSnafu { sql: &sql })?;
-
-        let either = match output.data {
-            OutputData::Stream(s) => {
-                let x = RecordBatches::try_collect(s)
-                    .await
-                    .context(CollectRecordBatchesSnafu)?;
-                Either::Left(x)
-            }
-            OutputData::RecordBatches(x) => Either::Left(x),
-            OutputData::AffectedRows(rows) => Either::Right(rows),
-        };
-
-        let end = Instant::now();
-
-        match either {
-            Either::Left(recordbatches) => {
-                let total_rows: usize = recordbatches.iter().map(|x| x.num_rows()).sum();
-                if total_rows > 0 {
-                    println!(
-                        "{}",
-                        recordbatches
-                            .pretty_print()
-                            .context(PrettyPrintRecordBatchesSnafu)?
-                    );
-                }
-                println!("Total Rows: {total_rows}")
-            }
-            Either::Right(rows) => println!("Affected Rows: {rows}"),
-        };
-
-        println!("Cost {} ms", (end - start).as_millis());
-        Ok(())
-    }
-}
-
-impl Drop for Repl {
-    fn drop(&mut self) {
-        if self.rl.helper().is_some() {
-            let history_file = history_file();
-            if let Err(e) = self.rl.save_history(&history_file) {
-                debug!(
-                    "failed to save history file on {}, error: {e}",
-                    history_file.display()
-                );
-            }
-        }
-    }
-}
-
-/// Return the location of the history file (defaults to $HOME/".greptimedb_cli_history")
-fn history_file() -> PathBuf {
-    let mut buf = match std::env::var("HOME") {
-        Ok(home) => PathBuf::from(home),
-        Err(_) => PathBuf::new(),
-    };
-    buf.push(".greptimedb_cli_history");
-    buf
-}
-
-async fn create_query_engine(meta_addr: &str) -> Result<DatafusionQueryEngine> {
-    let mut meta_client = MetaClientBuilder::default().enable_store().build();
-    meta_client
-        .start([meta_addr])
-        .await
-        .context(StartMetaClientSnafu)?;
-    let meta_client = Arc::new(meta_client);
-
-    let cached_meta_backend = Arc::new(
-        CachedKvBackendBuilder::new(Arc::new(MetaKvBackend::new(meta_client.clone()))).build(),
-    );
-    let layered_cache_builder = LayeredCacheRegistryBuilder::default().add_cache_registry(
-        CacheRegistryBuilder::default()
-            .add_cache(cached_meta_backend.clone())
-            .build(),
-    );
-    let fundamental_cache_registry =
-        build_fundamental_cache_registry(Arc::new(MetaKvBackend::new(meta_client.clone())));
-    let layered_cache_registry = Arc::new(
-        with_default_composite_cache_registry(
-            layered_cache_builder.add_cache_registry(fundamental_cache_registry),
-        )
-        .context(error::BuildCacheRegistrySnafu)?
-        .build(),
-    );
-
-    let information_extension = Arc::new(DistributedInformationExtension::new(meta_client.clone()));
-    let catalog_manager = KvBackendCatalogManager::new(
-        information_extension,
-        cached_meta_backend.clone(),
-        layered_cache_registry,
-        None,
-    );
-    let plugins: Plugins = Default::default();
-    let state = Arc::new(QueryEngineState::new(
-        catalog_manager,
-        None,
-        None,
-        None,
-        None,
-        false,
-        plugins.clone(),
-    ));
-
-    Ok(DatafusionQueryEngine::new(state, plugins))
-}

src/client/Cargo.toml

@@ -16,6 +16,7 @@ arc-swap = "1.6"
 arrow-flight.workspace = true
 async-stream.workspace = true
 async-trait.workspace = true
+base64.workspace = true
 common-catalog.workspace = true
 common-error.workspace = true
 common-grpc.workspace = true
@@ -25,6 +26,7 @@ common-query.workspace = true
 common-recordbatch.workspace = true
 common-telemetry.workspace = true
 enum_dispatch = "0.3"
+futures.workspace = true
 futures-util.workspace = true
 lazy_static.workspace = true
 moka = { workspace = true, features = ["future"] }

src/client/src/database.rs

@@ -12,36 +12,49 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::pin::Pin;
+use std::str::FromStr;
+
 use api::v1::auth_header::AuthScheme;
 use api::v1::ddl_request::Expr as DdlExpr;
 use api::v1::greptime_database_client::GreptimeDatabaseClient;
 use api::v1::greptime_request::Request;
 use api::v1::query_request::Query;
 use api::v1::{
-    AlterTableExpr, AuthHeader, CreateTableExpr, DdlRequest, GreptimeRequest, InsertRequests,
-    QueryRequest, RequestHeader,
+    AlterTableExpr, AuthHeader, Basic, CreateTableExpr, DdlRequest, GreptimeRequest,
+    InsertRequests, QueryRequest, RequestHeader,
 };
-use arrow_flight::Ticket;
+use arrow_flight::{FlightData, Ticket};
 use async_stream::stream;
+use base64::prelude::BASE64_STANDARD;
+use base64::Engine;
+use common_catalog::build_db_string;
+use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
 use common_error::ext::{BoxedError, ErrorExt};
+use common_grpc::flight::do_put::DoPutResponse;
 use common_grpc::flight::{FlightDecoder, FlightMessage};
 use common_query::Output;
 use common_recordbatch::error::ExternalSnafu;
 use common_recordbatch::RecordBatchStreamWrapper;
 use common_telemetry::error;
 use common_telemetry::tracing_context::W3cTrace;
-use futures_util::StreamExt;
+use futures::future;
+use futures_util::{Stream, StreamExt, TryStreamExt};
 use prost::Message;
 use snafu::{ensure, ResultExt};
-use tonic::metadata::AsciiMetadataKey;
+use tonic::metadata::{AsciiMetadataKey, MetadataValue};
 use tonic::transport::Channel;
 
 use crate::error::{
     ConvertFlightDataSnafu, Error, FlightGetSnafu, IllegalFlightMessagesSnafu, InvalidAsciiSnafu,
-    ServerSnafu,
+    InvalidTonicMetadataValueSnafu, ServerSnafu,
 };
 use crate::{from_grpc_response, Client, Result};
 
+type FlightDataStream = Pin<Box<dyn Stream<Item = FlightData> + Send>>;
+
+type DoPutResponseStream = Pin<Box<dyn Stream<Item = Result<DoPutResponse>>>>;
+
 #[derive(Clone, Debug, Default)]
 pub struct Database {
     // The "catalog" and "schema" to be used in processing the requests at the server side.
@@ -108,16 +121,24 @@ impl Database {
         self.catalog = catalog.into();
     }
 
-    pub fn catalog(&self) -> &String {
-        &self.catalog
+    fn catalog_or_default(&self) -> &str {
+        if self.catalog.is_empty() {
+            DEFAULT_CATALOG_NAME
+        } else {
+            &self.catalog
+        }
     }
 
     pub fn set_schema(&mut self, schema: impl Into<String>) {
         self.schema = schema.into();
     }
 
-    pub fn schema(&self) -> &String {
-        &self.schema
+    fn schema_or_default(&self) -> &str {
+        if self.schema.is_empty() {
+            DEFAULT_SCHEMA_NAME
+        } else {
+            &self.schema
+        }
     }
 
     pub fn set_timezone(&mut self, timezone: impl Into<String>) {
@@ -164,7 +185,7 @@ impl Database {
         from_grpc_response(response)
     }
 
-    async fn handle(&self, request: Request) -> Result<u32> {
+    pub async fn handle(&self, request: Request) -> Result<u32> {
         let mut client = make_database_client(&self.client)?.inner;
         let request = self.to_rpc_request(request);
         let response = client.handle(request).await?.into_inner();
@@ -310,6 +331,41 @@ impl Database {
             }
         }
     }
+
+    /// Ingest a stream of [RecordBatch]es that belong to a table, using Arrow Flight's "`DoPut`"
+    /// method. The return value is also a stream, produces [DoPutResponse]s.
+    pub async fn do_put(&self, stream: FlightDataStream) -> Result<DoPutResponseStream> {
+        let mut request = tonic::Request::new(stream);
+
+        if let Some(AuthHeader {
+            auth_scheme: Some(AuthScheme::Basic(Basic { username, password })),
+        }) = &self.ctx.auth_header
+        {
+            let encoded = BASE64_STANDARD.encode(format!("{username}:{password}"));
+            let value =
+                MetadataValue::from_str(&encoded).context(InvalidTonicMetadataValueSnafu)?;
+            request.metadata_mut().insert("x-greptime-auth", value);
+        }
+
+        let db_to_put = if !self.dbname.is_empty() {
+            &self.dbname
+        } else {
+            &build_db_string(self.catalog_or_default(), self.schema_or_default())
+        };
+        request.metadata_mut().insert(
+            "x-greptime-db-name",
+            MetadataValue::from_str(db_to_put).context(InvalidTonicMetadataValueSnafu)?,
+        );
+
+        let mut client = self.client.make_flight_client()?;
+        let response = client.mut_inner().do_put(request).await?;
+        let response = response
+            .into_inner()
+            .map_err(Into::into)
+            .and_then(|x| future::ready(DoPutResponse::try_from(x).context(ConvertFlightDataSnafu)))
+            .boxed();
+        Ok(response)
+    }
 }
 
 #[derive(Default, Debug, Clone)]

src/client/src/error.rs

@@ -15,10 +15,11 @@
 use std::any::Any;
 
 use common_error::ext::{BoxedError, ErrorExt};
-use common_error::status_code::StatusCode;
+use common_error::status_code::{convert_tonic_code_to_status_code, StatusCode};
 use common_error::{GREPTIME_DB_HEADER_ERROR_CODE, GREPTIME_DB_HEADER_ERROR_MSG};
 use common_macro::stack_trace_debug;
 use snafu::{location, Location, Snafu};
+use tonic::metadata::errors::InvalidMetadataValue;
 use tonic::{Code, Status};
 
 #[derive(Snafu)]
@@ -115,6 +116,14 @@ pub enum Error {
         #[snafu(implicit)]
         location: Location,
     },
+
+    #[snafu(display("Invalid Tonic metadata value"))]
+    InvalidTonicMetadataValue {
+        #[snafu(source)]
+        error: InvalidMetadataValue,
+        #[snafu(implicit)]
+        location: Location,
+    },
 }
 
 pub type Result<T> = std::result::Result<T, Error>;
@@ -135,7 +144,9 @@ impl ErrorExt for Error {
             | Error::CreateTlsChannel { source, .. } => source.status_code(),
             Error::IllegalGrpcClientState { .. } => StatusCode::Unexpected,
 
-            Error::InvalidAscii { .. } => StatusCode::InvalidArguments,
+            Error::InvalidAscii { .. } | Error::InvalidTonicMetadataValue { .. } => {
+                StatusCode::InvalidArguments
+            }
         }
     }
 
@@ -152,15 +163,15 @@ impl From<Status> for Error {
                 .and_then(|v| String::from_utf8(v.as_bytes().to_vec()).ok())
         }
 
-        let code = get_metadata_value(&e, GREPTIME_DB_HEADER_ERROR_CODE)
-            .and_then(|s| {
-                if let Ok(code) = s.parse::<u32>() {
-                    StatusCode::from_u32(code)
-                } else {
-                    None
-                }
-            })
-            .unwrap_or(StatusCode::Unknown);
+        let code = get_metadata_value(&e, GREPTIME_DB_HEADER_ERROR_CODE).and_then(|s| {
+            if let Ok(code) = s.parse::<u32>() {
+                StatusCode::from_u32(code)
+            } else {
+                None
+            }
+        });
+        let tonic_code = e.code();
+        let code = code.unwrap_or_else(|| convert_tonic_code_to_status_code(tonic_code));
 
         let msg = get_metadata_value(&e, GREPTIME_DB_HEADER_ERROR_MSG)
             .unwrap_or_else(|| e.message().to_string());
@@ -187,9 +198,6 @@ impl Error {
             } | Self::RegionServer {
                 code: Code::Unavailable,
                 ..
-            } | Self::RegionServer {
-                code: Code::Unknown,
-                ..
             }
         )
     }

src/client/src/lib.rs

@@ -16,8 +16,7 @@
 
 mod client;
 pub mod client_manager;
-#[cfg(feature = "testing")]
-mod database;
+pub mod database;
 pub mod error;
 pub mod flow;
 pub mod load_balance;
@@ -34,7 +33,6 @@ pub use common_recordbatch::{RecordBatches, SendableRecordBatchStream};
 use snafu::OptionExt;
 
 pub use self::client::Client;
-#[cfg(feature = "testing")]
 pub use self::database::Database;
 pub use self::error::{Error, Result};
 use crate::error::{IllegalDatabaseResponseSnafu, ServerSnafu};

src/client/src/load_balance.rs

@@ -13,7 +13,7 @@
 // limitations under the License.
 
 use enum_dispatch::enum_dispatch;
-use rand::seq::SliceRandom;
+use rand::seq::IndexedRandom;
 
 #[enum_dispatch]
 pub trait LoadBalance {
@@ -37,7 +37,7 @@ pub struct Random;
 
 impl LoadBalance for Random {
     fn get_peer<'a>(&self, peers: &'a [String]) -> Option<&'a String> {
-        peers.choose(&mut rand::thread_rng())
+        peers.choose(&mut rand::rng())
     }
 }
 

src/client/src/region.rs

@@ -201,12 +201,11 @@ impl RegionRequester {
             .await
             .map_err(|e| {
                 let code = e.code();
-                let err: error::Error = e.into();
                 // Uses `Error::RegionServer` instead of `Error::Server`
                 error::Error::RegionServer {
                     addr,
                     code,
-                    source: BoxedError::new(err),
+                    source: BoxedError::new(error::Error::from(e)),
                     location: location!(),
                 }
             })?

src/cmd/Cargo.toml

@@ -68,7 +68,6 @@ query.workspace = true
 rand.workspace = true
 regex.workspace = true
 reqwest.workspace = true
-rustyline = "10.1"
 serde.workspace = true
 serde_json.workspace = true
 servers.workspace = true

src/cmd/src/datanode.rs

@@ -30,7 +30,7 @@ use datanode::datanode::{Datanode, DatanodeBuilder};
 use datanode::service::DatanodeServiceBuilder;
 use meta_client::{MetaClientOptions, MetaClientType};
 use servers::Mode;
-use snafu::{OptionExt, ResultExt};
+use snafu::{ensure, OptionExt, ResultExt};
 use tracing_appender::non_blocking::WorkerGuard;
 
 use crate::error::{
@@ -223,15 +223,14 @@ impl StartCommand {
                 .get_or_insert_with(MetaClientOptions::default)
                 .metasrv_addrs
                 .clone_from(metasrv_addrs);
-            opts.mode = Mode::Distributed;
         }
 
-        if let (Mode::Distributed, None) = (&opts.mode, &opts.node_id) {
-            return MissingConfigSnafu {
-                msg: "Missing node id option",
+        ensure!(
+            opts.node_id.is_some(),
+            MissingConfigSnafu {
+                msg: "Missing node id option"
             }
-            .fail();
-        }
+        );
 
         if let Some(data_home) = &self.data_home {
             opts.storage.data_home.clone_from(data_home);
@@ -295,10 +294,13 @@ impl StartCommand {
             msg: "'meta_client_options'",
         })?;
 
-        let meta_client =
-            meta_client::create_meta_client(MetaClientType::Datanode { member_id }, meta_config)
-                .await
-                .context(MetaClientInitSnafu)?;
+        let meta_client = meta_client::create_meta_client(
+            MetaClientType::Datanode { member_id },
+            meta_config,
+            None,
+        )
+        .await
+        .context(MetaClientInitSnafu)?;
 
         let meta_backend = Arc::new(MetaKvBackend {
             client: meta_client.clone(),
@@ -311,7 +313,7 @@ impl StartCommand {
                 .build(),
         );
 
-        let mut datanode = DatanodeBuilder::new(opts.clone(), plugins)
+        let mut datanode = DatanodeBuilder::new(opts.clone(), plugins, Mode::Distributed)
             .with_meta_client(meta_client)
             .with_kv_backend(meta_backend)
             .with_cache_registry(layered_cache_registry)
@@ -333,6 +335,7 @@ impl StartCommand {
 
 #[cfg(test)]
 mod tests {
+    use std::assert_matches::assert_matches;
     use std::io::Write;
     use std::time::Duration;
 
@@ -340,7 +343,6 @@ mod tests {
     use common_test_util::temp_dir::create_named_temp_file;
     use datanode::config::{FileConfig, GcsConfig, ObjectStoreConfig, S3Config};
     use servers::heartbeat_options::HeartbeatOptions;
-    use servers::Mode;
 
     use super::*;
     use crate::options::GlobalOptions;
@@ -491,22 +493,6 @@ mod tests {
 
     #[test]
     fn test_try_from_cmd() {
-        let opt = StartCommand::default()
-            .load_options(&GlobalOptions::default())
-            .unwrap()
-            .component;
-        assert_eq!(Mode::Standalone, opt.mode);
-
-        let opt = (StartCommand {
-            node_id: Some(42),
-            metasrv_addrs: Some(vec!["127.0.0.1:3002".to_string()]),
-            ..Default::default()
-        })
-        .load_options(&GlobalOptions::default())
-        .unwrap()
-        .component;
-        assert_eq!(Mode::Distributed, opt.mode);
-
         assert!((StartCommand {
             metasrv_addrs: Some(vec!["127.0.0.1:3002".to_string()]),
             ..Default::default()
@@ -525,7 +511,19 @@ mod tests {
 
     #[test]
     fn test_load_log_options_from_cli() {
-        let cmd = StartCommand::default();
+        let mut cmd = StartCommand::default();
+
+        let result = cmd.load_options(&GlobalOptions {
+            log_dir: Some("./greptimedb_data/test/logs".to_string()),
+            log_level: Some("debug".to_string()),
+
+            #[cfg(feature = "tokio-console")]
+            tokio_console_addr: None,
+        });
+        // Missing node_id.
+        assert_matches!(result, Err(crate::error::Error::MissingConfig { .. }));
+
+        cmd.node_id = Some(42);
 
         let options = cmd
             .load_options(&GlobalOptions {

src/cmd/src/error.rs

@@ -17,7 +17,6 @@ use std::any::Any;
 use common_error::ext::{BoxedError, ErrorExt};
 use common_error::status_code::StatusCode;
 use common_macro::stack_trace_debug;
-use rustyline::error::ReadlineError;
 use snafu::{Location, Snafu};
 
 #[derive(Snafu)]
@@ -79,6 +78,13 @@ pub enum Error {
         source: datanode::error::Error,
     },
 
+    #[snafu(display("Failed to build object storage manager"))]
+    BuildObjectStorageManager {
+        #[snafu(implicit)]
+        location: Location,
+        source: datanode::error::Error,
+    },
+
     #[snafu(display("Failed to shutdown datanode"))]
     ShutdownDatanode {
         #[snafu(implicit)]
@@ -100,6 +106,13 @@ pub enum Error {
         source: flow::Error,
     },
 
+    #[snafu(display("Servers error"))]
+    Servers {
+        #[snafu(implicit)]
+        location: Location,
+        source: servers::error::Error,
+    },
+
     #[snafu(display("Failed to start frontend"))]
     StartFrontend {
         #[snafu(implicit)]
@@ -174,52 +187,6 @@ pub enum Error {
     #[snafu(display("Invalid REPL command: {reason}"))]
     InvalidReplCommand { reason: String },
 
-    #[snafu(display("Cannot create REPL"))]
-    ReplCreation {
-        #[snafu(source)]
-        error: ReadlineError,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
-    #[snafu(display("Error reading command"))]
-    Readline {
-        #[snafu(source)]
-        error: ReadlineError,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
-    #[snafu(display("Failed to request database, sql: {sql}"))]
-    RequestDatabase {
-        sql: String,
-        #[snafu(source)]
-        source: client::Error,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
-    #[snafu(display("Failed to collect RecordBatches"))]
-    CollectRecordBatches {
-        #[snafu(implicit)]
-        location: Location,
-        source: common_recordbatch::error::Error,
-    },
-
-    #[snafu(display("Failed to pretty print Recordbatches"))]
-    PrettyPrintRecordBatches {
-        #[snafu(implicit)]
-        location: Location,
-        source: common_recordbatch::error::Error,
-    },
-
-    #[snafu(display("Failed to start Meta client"))]
-    StartMetaClient {
-        #[snafu(implicit)]
-        location: Location,
-        source: meta_client::error::Error,
-    },
-
     #[snafu(display("Failed to parse SQL: {}", sql))]
     ParseSql {
         sql: String,
@@ -235,13 +202,6 @@ pub enum Error {
         source: query::error::Error,
     },
 
-    #[snafu(display("Failed to encode logical plan in substrait"))]
-    SubstraitEncodeLogicalPlan {
-        #[snafu(implicit)]
-        location: Location,
-        source: substrait::error::Error,
-    },
-
     #[snafu(display("Failed to load layered config"))]
     LoadLayeredConfig {
         #[snafu(source(from(common_config::error::Error, Box::new)))]
@@ -365,6 +325,7 @@ impl ErrorExt for Error {
             Error::ShutdownFrontend { source, .. } => source.status_code(),
             Error::StartMetaServer { source, .. } => source.status_code(),
             Error::ShutdownMetaServer { source, .. } => source.status_code(),
+            Error::Servers { source, .. } => source.status_code(),
             Error::BuildMetaServer { source, .. } => source.status_code(),
             Error::UnsupportedSelectorType { source, .. } => source.status_code(),
             Error::BuildCli { source, .. } => source.status_code(),
@@ -374,6 +335,8 @@ impl ErrorExt for Error {
                 source.status_code()
             }
 
+            Error::BuildObjectStorageManager { source, .. } => source.status_code(),
+
             Error::MissingConfig { .. }
             | Error::LoadLayeredConfig { .. }
             | Error::IllegalConfig { .. }
@@ -387,17 +350,10 @@ impl ErrorExt for Error {
             | Error::StopProcedureManager { source, .. } => source.status_code(),
             Error::BuildWalOptionsAllocator { source, .. }
             | Error::StartWalOptionsAllocator { source, .. } => source.status_code(),
-            Error::ReplCreation { .. } | Error::Readline { .. } | Error::HttpQuerySql { .. } => {
-                StatusCode::Internal
-            }
-            Error::RequestDatabase { source, .. } => source.status_code(),
-            Error::CollectRecordBatches { source, .. }
-            | Error::PrettyPrintRecordBatches { source, .. } => source.status_code(),
-            Error::StartMetaClient { source, .. } => source.status_code(),
+            Error::HttpQuerySql { .. } => StatusCode::Internal,
             Error::ParseSql { source, .. } | Error::PlanStatement { source, .. } => {
                 source.status_code()
             }
-            Error::SubstraitEncodeLogicalPlan { source, .. } => source.status_code(),
 
             Error::SerdeJson { .. }
             | Error::FileIo { .. }

src/cmd/src/flownode.rs

@@ -32,10 +32,11 @@ use common_meta::key::TableMetadataManager;
 use common_telemetry::info;
 use common_telemetry::logging::TracingOptions;
 use common_version::{short_version, version};
-use flow::{FlownodeBuilder, FlownodeInstance, FrontendInvoker};
+use flow::{
+    FlownodeBuilder, FlownodeInstance, FlownodeServiceBuilder, FrontendClient, FrontendInvoker,
+};
 use meta_client::{MetaClientOptions, MetaClientType};
-use servers::Mode;
-use snafu::{OptionExt, ResultExt};
+use snafu::{ensure, OptionExt, ResultExt};
 use tracing_appender::non_blocking::WorkerGuard;
 
 use crate::error::{
@@ -203,7 +204,6 @@ impl StartCommand {
                 .get_or_insert_with(MetaClientOptions::default)
                 .metasrv_addrs
                 .clone_from(metasrv_addrs);
-            opts.mode = Mode::Distributed;
         }
 
         if let Some(http_addr) = &self.http_addr {
@@ -214,12 +214,12 @@ impl StartCommand {
             opts.http.timeout = Duration::from_secs(http_timeout);
         }
 
-        if let (Mode::Distributed, None) = (&opts.mode, &opts.node_id) {
-            return MissingConfigSnafu {
-                msg: "Missing node id option",
+        ensure!(
+            opts.node_id.is_some(),
+            MissingConfigSnafu {
+                msg: "Missing node id option"
             }
-            .fail();
-        }
+        );
 
         Ok(())
     }
@@ -249,10 +249,13 @@ impl StartCommand {
             msg: "'meta_client_options'",
         })?;
 
-        let meta_client =
-            meta_client::create_meta_client(MetaClientType::Flownode { member_id }, meta_config)
-                .await
-                .context(MetaClientInitSnafu)?;
+        let meta_client = meta_client::create_meta_client(
+            MetaClientType::Flownode { member_id },
+            meta_config,
+            None,
+        )
+        .await
+        .context(MetaClientInitSnafu)?;
 
         let cache_max_capacity = meta_config.metadata_cache_max_capacity;
         let cache_ttl = meta_config.metadata_cache_ttl;
@@ -312,16 +315,26 @@ impl StartCommand {
         );
 
         let flow_metadata_manager = Arc::new(FlowMetadataManager::new(cached_meta_backend.clone()));
+        let frontend_client = FrontendClient::from_meta_client(meta_client.clone());
         let flownode_builder = FlownodeBuilder::new(
-            opts,
+            opts.clone(),
             Plugins::new(),
             table_metadata_manager,
             catalog_manager.clone(),
             flow_metadata_manager,
+            Arc::new(frontend_client),
         )
         .with_heartbeat_task(heartbeat_task);
 
-        let flownode = flownode_builder.build().await.context(StartFlownodeSnafu)?;
+        let mut flownode = flownode_builder.build().await.context(StartFlownodeSnafu)?;
+        let services = FlownodeServiceBuilder::new(&opts)
+            .with_grpc_server(flownode.flownode_server().clone())
+            .enable_http_service()
+            .build()
+            .await
+            .context(StartFlownodeSnafu)?;
+        flownode.setup_services(services);
+        let flownode = flownode;
 
         // flownode's frontend to datanode need not timeout.
         // Some queries are expected to take long time.

src/cmd/src/frontend.rs

@@ -32,28 +32,25 @@ use common_telemetry::info;
 use common_telemetry::logging::TracingOptions;
 use common_time::timezone::set_default_timezone;
 use common_version::{short_version, version};
+use frontend::frontend::Frontend;
 use frontend::heartbeat::HeartbeatTask;
 use frontend::instance::builder::FrontendBuilder;
-use frontend::instance::{FrontendInstance, Instance as FeInstance};
 use frontend::server::Services;
 use meta_client::{MetaClientOptions, MetaClientType};
 use query::stats::StatementStatistics;
+use servers::export_metrics::ExportMetricsTask;
 use servers::tls::{TlsMode, TlsOption};
 use snafu::{OptionExt, ResultExt};
 use tracing_appender::non_blocking::WorkerGuard;
 
-use crate::error::{
-    self, InitTimezoneSnafu, LoadLayeredConfigSnafu, MetaClientInitSnafu, MissingConfigSnafu,
-    Result, StartFrontendSnafu,
-};
+use crate::error::{self, Result};
 use crate::options::{GlobalOptions, GreptimeOptions};
 use crate::{log_versions, App};
 
 type FrontendOptions = GreptimeOptions<frontend::frontend::FrontendOptions>;
 
 pub struct Instance {
-    frontend: FeInstance,
-
+    frontend: Frontend,
     // Keep the logging guard to prevent the worker from being dropped.
     _guard: Vec<WorkerGuard>,
 }
@@ -61,20 +58,17 @@ pub struct Instance {
 pub const APP_NAME: &str = "greptime-frontend";
 
 impl Instance {
-    pub fn new(frontend: FeInstance, guard: Vec<WorkerGuard>) -> Self {
-        Self {
-            frontend,
-            _guard: guard,
-        }
+    pub fn new(frontend: Frontend, _guard: Vec<WorkerGuard>) -> Self {
+        Self { frontend, _guard }
     }
 
-    pub fn mut_inner(&mut self) -> &mut FeInstance {
-        &mut self.frontend
-    }
-
-    pub fn inner(&self) -> &FeInstance {
+    pub fn inner(&self) -> &Frontend {
         &self.frontend
     }
+
+    pub fn mut_inner(&mut self) -> &mut Frontend {
+        &mut self.frontend
+    }
 }
 
 #[async_trait]
@@ -84,11 +78,15 @@ impl App for Instance {
     }
 
     async fn start(&mut self) -> Result<()> {
-        plugins::start_frontend_plugins(self.frontend.plugins().clone())
+        let plugins = self.frontend.instance.plugins().clone();
+        plugins::start_frontend_plugins(plugins)
             .await
-            .context(StartFrontendSnafu)?;
+            .context(error::StartFrontendSnafu)?;
 
-        self.frontend.start().await.context(StartFrontendSnafu)
+        self.frontend
+            .start()
+            .await
+            .context(error::StartFrontendSnafu)
     }
 
     async fn stop(&self) -> Result<()> {
@@ -178,7 +176,7 @@ impl StartCommand {
             self.config_file.as_deref(),
             self.env_prefix.as_ref(),
         )
-        .context(LoadLayeredConfigSnafu)?;
+        .context(error::LoadLayeredConfigSnafu)?;
 
         self.merge_with_cli_options(global_options, &mut opts)?;
 
@@ -283,22 +281,28 @@ impl StartCommand {
         let mut plugins = Plugins::new();
         plugins::setup_frontend_plugins(&mut plugins, &plugin_opts, &opts)
             .await
-            .context(StartFrontendSnafu)?;
+            .context(error::StartFrontendSnafu)?;
 
-        set_default_timezone(opts.default_timezone.as_deref()).context(InitTimezoneSnafu)?;
+        set_default_timezone(opts.default_timezone.as_deref()).context(error::InitTimezoneSnafu)?;
 
-        let meta_client_options = opts.meta_client.as_ref().context(MissingConfigSnafu {
-            msg: "'meta_client'",
-        })?;
+        let meta_client_options = opts
+            .meta_client
+            .as_ref()
+            .context(error::MissingConfigSnafu {
+                msg: "'meta_client'",
+            })?;
 
         let cache_max_capacity = meta_client_options.metadata_cache_max_capacity;
         let cache_ttl = meta_client_options.metadata_cache_ttl;
         let cache_tti = meta_client_options.metadata_cache_tti;
 
-        let meta_client =
-            meta_client::create_meta_client(MetaClientType::Frontend, meta_client_options)
-                .await
-                .context(MetaClientInitSnafu)?;
+        let meta_client = meta_client::create_meta_client(
+            MetaClientType::Frontend,
+            meta_client_options,
+            Some(&plugins),
+        )
+        .await
+        .context(error::MetaClientInitSnafu)?;
 
         // TODO(discord9): add helper function to ease the creation of cache registry&such
         let cached_meta_backend =
@@ -345,6 +349,7 @@ impl StartCommand {
             opts.heartbeat.clone(),
             Arc::new(executor),
         );
+        let heartbeat_task = Some(heartbeat_task);
 
         // frontend to datanode need not timeout.
         // Some queries are expected to take long time.
@@ -356,7 +361,7 @@ impl StartCommand {
         };
         let client = NodeClients::new(channel_config);
 
-        let mut instance = FrontendBuilder::new(
+        let instance = FrontendBuilder::new(
             opts.clone(),
             cached_meta_backend.clone(),
             layered_cache_registry.clone(),
@@ -367,20 +372,27 @@ impl StartCommand {
         )
         .with_plugin(plugins.clone())
         .with_local_cache_invalidator(layered_cache_registry)
-        .with_heartbeat_task(heartbeat_task)
         .try_build()
         .await
-        .context(StartFrontendSnafu)?;
+        .context(error::StartFrontendSnafu)?;
+        let instance = Arc::new(instance);
 
-        let servers = Services::new(opts, Arc::new(instance.clone()), plugins)
+        let export_metrics_task = ExportMetricsTask::try_new(&opts.export_metrics, Some(&plugins))
+            .context(error::ServersSnafu)?;
+
+        let servers = Services::new(opts, instance.clone(), plugins)
             .build()
             .await
-            .context(StartFrontendSnafu)?;
-        instance
-            .build_servers(servers)
-            .context(StartFrontendSnafu)?;
+            .context(error::StartFrontendSnafu)?;
 
-        Ok(Instance::new(instance, guard))
+        let frontend = Frontend {
+            instance,
+            servers,
+            heartbeat_task,
+            export_metrics_task,
+        };
+
+        Ok(Instance::new(frontend, guard))
     }
 }
 

src/cmd/src/standalone.rs

@@ -42,7 +42,9 @@ use common_meta::kv_backend::KvBackendRef;
 use common_meta::node_manager::NodeManagerRef;
 use common_meta::peer::Peer;
 use common_meta::region_keeper::MemoryRegionKeeper;
+use common_meta::region_registry::LeaderRegionRegistry;
 use common_meta::sequence::SequenceBuilder;
+use common_meta::snapshot::MetadataSnapshotManager;
 use common_meta::wal_options_allocator::{build_wal_options_allocator, WalOptionsAllocatorRef};
 use common_procedure::{ProcedureInfo, ProcedureManagerRef};
 use common_telemetry::info;
@@ -54,10 +56,13 @@ use datanode::config::{DatanodeOptions, ProcedureConfig, RegionEngineConfig, Sto
 use datanode::datanode::{Datanode, DatanodeBuilder};
 use datanode::region_server::RegionServer;
 use file_engine::config::EngineConfig as FileEngineConfig;
-use flow::{FlowConfig, FlowWorkerManager, FlownodeBuilder, FlownodeOptions, FrontendInvoker};
-use frontend::frontend::FrontendOptions;
+use flow::{
+    FlowConfig, FlowWorkerManager, FlownodeBuilder, FlownodeInstance, FlownodeOptions,
+    FrontendClient, FrontendInvoker,
+};
+use frontend::frontend::{Frontend, FrontendOptions};
 use frontend::instance::builder::FrontendBuilder;
-use frontend::instance::{FrontendInstance, Instance as FeInstance, StandaloneDatanodeManager};
+use frontend::instance::{Instance as FeInstance, StandaloneDatanodeManager};
 use frontend::server::Services;
 use frontend::service_config::{
     InfluxdbOptions, JaegerOptions, MysqlOptions, OpentsdbOptions, PostgresOptions,
@@ -67,24 +72,18 @@ use meta_srv::metasrv::{FLOW_ID_SEQ, TABLE_ID_SEQ};
 use mito2::config::MitoConfig;
 use query::stats::StatementStatistics;
 use serde::{Deserialize, Serialize};
-use servers::export_metrics::ExportMetricsOption;
+use servers::export_metrics::{ExportMetricsOption, ExportMetricsTask};
 use servers::grpc::GrpcOptions;
 use servers::http::HttpOptions;
 use servers::tls::{TlsMode, TlsOption};
 use servers::Mode;
 use snafu::ResultExt;
-use tokio::sync::{broadcast, RwLock};
+use tokio::sync::RwLock;
 use tracing_appender::non_blocking::WorkerGuard;
 
-use crate::error::{
-    BuildCacheRegistrySnafu, BuildWalOptionsAllocatorSnafu, CreateDirSnafu, IllegalConfigSnafu,
-    InitDdlManagerSnafu, InitMetadataSnafu, InitTimezoneSnafu, LoadLayeredConfigSnafu, OtherSnafu,
-    Result, ShutdownDatanodeSnafu, ShutdownFlownodeSnafu, ShutdownFrontendSnafu,
-    StartDatanodeSnafu, StartFlownodeSnafu, StartFrontendSnafu, StartProcedureManagerSnafu,
-    StartWalOptionsAllocatorSnafu, StopProcedureManagerSnafu,
-};
+use crate::error::{Result, StartFlownodeSnafu};
 use crate::options::{GlobalOptions, GreptimeOptions};
-use crate::{log_versions, App};
+use crate::{error, log_versions, App};
 
 pub const APP_NAME: &str = "greptime-standalone";
 
@@ -132,7 +131,6 @@ impl SubCommand {
 #[derive(Clone, Debug, Serialize, Deserialize, PartialEq)]
 #[serde(default)]
 pub struct StandaloneOptions {
-    pub mode: Mode,
     pub enable_telemetry: bool,
     pub default_timezone: Option<String>,
     pub http: HttpOptions,
@@ -162,7 +160,6 @@ pub struct StandaloneOptions {
 impl Default for StandaloneOptions {
     fn default() -> Self {
         Self {
-            mode: Mode::Standalone,
             enable_telemetry: true,
             default_timezone: None,
             http: HttpOptions::default(),
@@ -243,7 +240,6 @@ impl StandaloneOptions {
             grpc: cloned_opts.grpc,
             init_regions_in_background: cloned_opts.init_regions_in_background,
             init_regions_parallelism: cloned_opts.init_regions_parallelism,
-            mode: Mode::Standalone,
             ..Default::default()
         }
     }
@@ -251,13 +247,10 @@ impl StandaloneOptions {
 
 pub struct Instance {
     datanode: Datanode,
-    frontend: FeInstance,
-    // TODO(discord9): wrapped it in flownode instance instead
-    flow_worker_manager: Arc<FlowWorkerManager>,
-    flow_shutdown: broadcast::Sender<()>,
+    frontend: Frontend,
+    flownode: FlownodeInstance,
     procedure_manager: ProcedureManagerRef,
     wal_options_allocator: WalOptionsAllocatorRef,
-
     // Keep the logging guard to prevent the worker from being dropped.
     _guard: Vec<WorkerGuard>,
 }
@@ -281,21 +274,24 @@ impl App for Instance {
         self.procedure_manager
             .start()
             .await
-            .context(StartProcedureManagerSnafu)?;
+            .context(error::StartProcedureManagerSnafu)?;
 
         self.wal_options_allocator
             .start()
             .await
-            .context(StartWalOptionsAllocatorSnafu)?;
+            .context(error::StartWalOptionsAllocatorSnafu)?;
 
-        plugins::start_frontend_plugins(self.frontend.plugins().clone())
+        plugins::start_frontend_plugins(self.frontend.instance.plugins().clone())
             .await
-            .context(StartFrontendSnafu)?;
+            .context(error::StartFrontendSnafu)?;
+
+        self.frontend
+            .start()
+            .await
+            .context(error::StartFrontendSnafu)?;
+
+        self.flownode.start().await.context(StartFlownodeSnafu)?;
 
-        self.frontend.start().await.context(StartFrontendSnafu)?;
-        self.flow_worker_manager
-            .clone()
-            .run_background(Some(self.flow_shutdown.subscribe()));
         Ok(())
     }
 
@@ -303,26 +299,23 @@ impl App for Instance {
         self.frontend
             .shutdown()
             .await
-            .context(ShutdownFrontendSnafu)?;
+            .context(error::ShutdownFrontendSnafu)?;
 
         self.procedure_manager
             .stop()
             .await
-            .context(StopProcedureManagerSnafu)?;
+            .context(error::StopProcedureManagerSnafu)?;
 
         self.datanode
             .shutdown()
             .await
-            .context(ShutdownDatanodeSnafu)?;
-        self.flow_shutdown
-            .send(())
-            .map_err(|_e| {
-                flow::error::InternalSnafu {
-                    reason: "Failed to send shutdown signal to flow worker manager, all receiver end already closed".to_string(),
-                }
-                .build()
-            })
-            .context(ShutdownFlownodeSnafu)?;
+            .context(error::ShutdownDatanodeSnafu)?;
+
+        self.flownode
+            .shutdown()
+            .await
+            .context(error::ShutdownFlownodeSnafu)?;
+
         info!("Datanode instance stopped.");
 
         Ok(())
@@ -368,7 +361,7 @@ impl StartCommand {
             self.config_file.as_deref(),
             self.env_prefix.as_ref(),
         )
-        .context(LoadLayeredConfigSnafu)?;
+        .context(error::LoadLayeredConfigSnafu)?;
 
         self.merge_with_cli_options(global_options, &mut opts.component)?;
 
@@ -381,9 +374,6 @@ impl StartCommand {
         global_options: &GlobalOptions,
         opts: &mut StandaloneOptions,
     ) -> Result<()> {
-        // Should always be standalone mode.
-        opts.mode = Mode::Standalone;
-
         if let Some(dir) = &global_options.log_dir {
             opts.logging.dir.clone_from(dir);
         }
@@ -415,7 +405,7 @@ impl StartCommand {
             // frontend grpc addr conflict with datanode default grpc addr
             let datanode_grpc_addr = DatanodeOptions::default().grpc.bind_addr;
             if addr.eq(&datanode_grpc_addr) {
-                return IllegalConfigSnafu {
+                return error::IllegalConfigSnafu {
                     msg: format!(
                         "gRPC listen address conflicts with datanode reserved gRPC addr: {datanode_grpc_addr}",
                     ),
@@ -474,18 +464,19 @@ impl StartCommand {
 
         plugins::setup_frontend_plugins(&mut plugins, &plugin_opts, &fe_opts)
             .await
-            .context(StartFrontendSnafu)?;
+            .context(error::StartFrontendSnafu)?;
 
         plugins::setup_datanode_plugins(&mut plugins, &plugin_opts, &dn_opts)
             .await
-            .context(StartDatanodeSnafu)?;
+            .context(error::StartDatanodeSnafu)?;
 
-        set_default_timezone(fe_opts.default_timezone.as_deref()).context(InitTimezoneSnafu)?;
+        set_default_timezone(fe_opts.default_timezone.as_deref())
+            .context(error::InitTimezoneSnafu)?;
 
         let data_home = &dn_opts.storage.data_home;
         // Ensure the data_home directory exists.
         fs::create_dir_all(path::Path::new(data_home))
-            .context(CreateDirSnafu { dir: data_home })?;
+            .context(error::CreateDirSnafu { dir: data_home })?;
 
         let metadata_dir = metadata_store_dir(data_home);
         let (kv_backend, procedure_manager) = FeInstance::try_build_standalone_components(
@@ -494,7 +485,7 @@ impl StartCommand {
             opts.procedure,
         )
         .await
-        .context(StartFrontendSnafu)?;
+        .context(error::StartFrontendSnafu)?;
 
         // Builds cache registry
         let layered_cache_builder = LayeredCacheRegistryBuilder::default();
@@ -503,16 +494,20 @@ impl StartCommand {
             with_default_composite_cache_registry(
                 layered_cache_builder.add_cache_registry(fundamental_cache_registry),
             )
-            .context(BuildCacheRegistrySnafu)?
+            .context(error::BuildCacheRegistrySnafu)?
             .build(),
         );
 
-        let datanode = DatanodeBuilder::new(dn_opts, plugins.clone())
+        let object_store_manager = DatanodeBuilder::build_object_store_manager(&dn_opts.storage)
+            .await
+            .context(error::BuildObjectStorageManagerSnafu)?;
+
+        let datanode = DatanodeBuilder::new(dn_opts, plugins.clone(), Mode::Standalone)
             .with_kv_backend(kv_backend.clone())
             .with_cache_registry(layered_cache_registry.clone())
             .build()
             .await
-            .context(StartDatanodeSnafu)?;
+            .context(error::StartDatanodeSnafu)?;
 
         let information_extension = Arc::new(StandaloneInformationExtension::new(
             datanode.region_server(),
@@ -533,20 +528,24 @@ impl StartCommand {
             flow: opts.flow.clone(),
             ..Default::default()
         };
+
+        // TODO(discord9): for standalone not use grpc, but just somehow get a handler to frontend grpc client without
+        // actually make a connection
+        let fe_server_addr = fe_opts.grpc.bind_addr.clone();
+        let frontend_client = FrontendClient::from_static_grpc_addr(fe_server_addr);
         let flow_builder = FlownodeBuilder::new(
             flownode_options,
             plugins.clone(),
             table_metadata_manager.clone(),
             catalog_manager.clone(),
             flow_metadata_manager.clone(),
+            Arc::new(frontend_client),
         );
-        let flownode = Arc::new(
-            flow_builder
-                .build()
-                .await
-                .map_err(BoxedError::new)
-                .context(OtherSnafu)?,
-        );
+        let flownode = flow_builder
+            .build()
+            .await
+            .map_err(BoxedError::new)
+            .context(error::OtherSnafu)?;
 
         // set the ref to query for the local flow state
         {
@@ -576,7 +575,7 @@ impl StartCommand {
         let kafka_options = opts.wal.clone().into();
         let wal_options_allocator = build_wal_options_allocator(&kafka_options, kv_backend.clone())
             .await
-            .context(BuildWalOptionsAllocatorSnafu)?;
+            .context(error::BuildWalOptionsAllocatorSnafu)?;
         let wal_options_allocator = Arc::new(wal_options_allocator);
         let table_meta_allocator = Arc::new(TableMetadataAllocator::new(
             table_id_sequence,
@@ -597,8 +596,13 @@ impl StartCommand {
         )
         .await?;
 
-        let mut frontend = FrontendBuilder::new(
-            fe_opts,
+        let metadata_snapshot_manager = MetadataSnapshotManager::new(
+            kv_backend.clone(),
+            object_store_manager.default_object_store().clone(),
+        );
+
+        let fe_instance = FrontendBuilder::new(
+            fe_opts.clone(),
             kv_backend.clone(),
             layered_cache_registry.clone(),
             catalog_manager.clone(),
@@ -607,9 +611,11 @@ impl StartCommand {
             StatementStatistics::new(opts.logging.slow_query.clone()),
         )
         .with_plugin(plugins.clone())
+        .with_metadata_snapshot_manager(metadata_snapshot_manager)
         .try_build()
         .await
-        .context(StartFrontendSnafu)?;
+        .context(error::StartFrontendSnafu)?;
+        let fe_instance = Arc::new(fe_instance);
 
         let flow_worker_manager = flownode.flow_worker_manager();
         // flow server need to be able to use frontend to write insert requests back
@@ -622,24 +628,28 @@ impl StartCommand {
             node_manager,
         )
         .await
-        .context(StartFlownodeSnafu)?;
+        .context(error::StartFlownodeSnafu)?;
         flow_worker_manager.set_frontend_invoker(invoker).await;
 
-        let (tx, _rx) = broadcast::channel(1);
+        let export_metrics_task = ExportMetricsTask::try_new(&opts.export_metrics, Some(&plugins))
+            .context(error::ServersSnafu)?;
 
-        let servers = Services::new(opts, Arc::new(frontend.clone()), plugins)
+        let servers = Services::new(opts, fe_instance.clone(), plugins)
             .build()
             .await
-            .context(StartFrontendSnafu)?;
-        frontend
-            .build_servers(servers)
-            .context(StartFrontendSnafu)?;
+            .context(error::StartFrontendSnafu)?;
+
+        let frontend = Frontend {
+            instance: fe_instance,
+            servers,
+            heartbeat_task: None,
+            export_metrics_task,
+        };
 
         Ok(Instance {
             datanode,
             frontend,
-            flow_worker_manager,
-            flow_shutdown: tx,
+            flownode,
             procedure_manager,
             wal_options_allocator,
             _guard: guard,
@@ -661,6 +671,7 @@ impl StartCommand {
                     node_manager,
                     cache_invalidator,
                     memory_region_keeper: Arc::new(MemoryRegionKeeper::default()),
+                    leader_region_registry: Arc::new(LeaderRegionRegistry::default()),
                     table_metadata_manager,
                     table_metadata_allocator,
                     flow_metadata_manager,
@@ -670,7 +681,7 @@ impl StartCommand {
                 procedure_manager,
                 true,
             )
-            .context(InitDdlManagerSnafu)?,
+            .context(error::InitDdlManagerSnafu)?,
         );
 
         Ok(procedure_executor)
@@ -684,7 +695,7 @@ impl StartCommand {
         table_metadata_manager
             .init()
             .await
-            .context(InitMetadataSnafu)?;
+            .context(error::InitMetadataSnafu)?;
 
         Ok(table_metadata_manager)
     }
@@ -778,6 +789,9 @@ impl InformationExtension for StandaloneInformationExtension {
                     manifest_size: region_stat.manifest_size,
                     sst_size: region_stat.sst_size,
                     index_size: region_stat.index_size,
+                    region_manifest: region_stat.manifest.into(),
+                    data_topic_latest_entry_id: region_stat.data_topic_latest_entry_id,
+                    metadata_topic_latest_entry_id: region_stat.metadata_topic_latest_entry_id,
                 }
             })
             .collect::<Vec<_>>();
@@ -1054,7 +1068,6 @@ mod tests {
         let options =
             StandaloneOptions::load_layered_options(None, "GREPTIMEDB_STANDALONE").unwrap();
         let default_options = StandaloneOptions::default();
-        assert_eq!(options.mode, default_options.mode);
         assert_eq!(options.enable_telemetry, default_options.enable_telemetry);
         assert_eq!(options.http, default_options.http);
         assert_eq!(options.grpc, default_options.grpc);

src/cmd/tests/cli.rs

@@ -1,148 +0,0 @@
-// Copyright 2023 Greptime Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-#[cfg(target_os = "macos")]
-mod tests {
-    use std::path::PathBuf;
-    use std::process::{Command, Stdio};
-    use std::time::Duration;
-
-    use common_test_util::temp_dir::create_temp_dir;
-    use rexpect::session::PtyReplSession;
-
-    struct Repl {
-        repl: PtyReplSession,
-    }
-
-    impl Repl {
-        fn send_line(&mut self, line: &str) {
-            let _ = self.repl.send_line(line).unwrap();
-
-            // read a line to consume the prompt
-            let _ = self.read_line();
-        }
-
-        fn read_line(&mut self) -> String {
-            self.repl.read_line().unwrap()
-        }
-
-        fn read_expect(&mut self, expect: &str) {
-            assert_eq!(self.read_line(), expect);
-        }
-
-        fn read_contains(&mut self, pat: &str) {
-            assert!(self.read_line().contains(pat));
-        }
-    }
-
-    // TODO(LFC): Un-ignore this REPL test.
-    // Ignore this REPL test because some logical plans like create database are not supported yet in Datanode.
-    #[ignore]
-    #[test]
-    fn test_repl() {
-        let data_home = create_temp_dir("data");
-        let wal_dir = create_temp_dir("wal");
-
-        let mut bin_path = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
-        bin_path.push("../../target/debug");
-        let bin_path = bin_path.to_str().unwrap();
-
-        let mut datanode = Command::new("./greptime")
-            .current_dir(bin_path)
-            .args([
-                "datanode",
-                "start",
-                "--rpc-bind-addr=0.0.0.0:4321",
-                "--node-id=1",
-                &format!("--data-home={}", data_home.path().display()),
-                &format!("--wal-dir={}", wal_dir.path().display()),
-            ])
-            .stdout(Stdio::null())
-            .spawn()
-            .unwrap();
-
-        // wait for Datanode actually started
-        std::thread::sleep(Duration::from_secs(3));
-
-        let mut repl_cmd = Command::new("./greptime");
-        let _ = repl_cmd.current_dir(bin_path).args([
-            "--log-level=off",
-            "cli",
-            "attach",
-            "--grpc-bind-addr=0.0.0.0:4321",
-            // history commands can sneaky into stdout and mess up our tests, so disable it
-            "--disable-helper",
-        ]);
-        let pty_session = rexpect::session::spawn_command(repl_cmd, Some(5_000)).unwrap();
-        let repl = PtyReplSession {
-            prompt: "> ".to_string(),
-            pty_session,
-            quit_command: None,
-            echo_on: false,
-        };
-        let repl = &mut Repl { repl };
-        repl.read_expect("Ready for commands. (Hint: try 'help')");
-
-        test_create_database(repl);
-
-        test_use_database(repl);
-
-        test_create_table(repl);
-
-        test_insert(repl);
-
-        test_select(repl);
-
-        datanode.kill().unwrap();
-        let _ = datanode.wait().unwrap();
-    }
-
-    fn test_create_database(repl: &mut Repl) {
-        repl.send_line("CREATE DATABASE db;");
-        repl.read_expect("Affected Rows: 1");
-        repl.read_contains("Cost");
-    }
-
-    fn test_use_database(repl: &mut Repl) {
-        repl.send_line("USE db");
-        repl.read_expect("Total Rows: 0");
-        repl.read_contains("Cost");
-        repl.read_expect("Using db");
-    }
-
-    fn test_create_table(repl: &mut Repl) {
-        repl.send_line("CREATE TABLE t(x STRING, ts TIMESTAMP TIME INDEX);");
-        repl.read_expect("Affected Rows: 0");
-        repl.read_contains("Cost");
-    }
-
-    fn test_insert(repl: &mut Repl) {
-        repl.send_line("INSERT INTO t(x, ts) VALUES ('hello', 1676895812239);");
-        repl.read_expect("Affected Rows: 1");
-        repl.read_contains("Cost");
-    }
-
-    fn test_select(repl: &mut Repl) {
-        repl.send_line("SELECT * FROM t;");
-
-        repl.read_expect("+-------+-------------------------+");
-        repl.read_expect("| x     | ts                      |");
-        repl.read_expect("+-------+-------------------------+");
-        repl.read_expect("| hello | 2023-02-20T12:23:32.239 |");
-        repl.read_expect("+-------+-------------------------+");
-        repl.read_expect("Total Rows: 1");
-
-        repl.read_contains("Cost");
-    }
-}

src/cmd/tests/load_config_test.rs

@@ -168,8 +168,8 @@ fn test_load_metasrv_example_config() {
                 tracing_sample_ratio: Some(Default::default()),
                 slow_query: SlowQueryOptions {
                     enable: false,
-                    threshold: Some(Duration::from_secs(10)),
-                    sample_ratio: Some(1.0),
+                    threshold: None,
+                    sample_ratio: None,
                 },
                 ..Default::default()
             },

src/common/catalog/src/consts.rs

@@ -137,4 +137,12 @@ pub const SPAN_ID_COLUMN: &str = "span_id";
 pub const SPAN_NAME_COLUMN: &str = "span_name";
 pub const SERVICE_NAME_COLUMN: &str = "service_name";
 pub const PARENT_SPAN_ID_COLUMN: &str = "parent_span_id";
+pub const TRACE_TABLE_NAME: &str = "opentelemetry_traces";
+pub const TRACE_TABLE_NAME_SESSION_KEY: &str = "trace_table_name";
+// ---- End of special table and fields ----
+
+/// Generate the trace services table name from the trace table name by adding `_services` suffix.
+pub fn trace_services_table_name(trace_table_name: &str) -> String {
+    format!("{}_services", trace_table_name)
+}
 // ---- End of special table and fields ----

src/common/datasource/Cargo.toml

@@ -31,7 +31,8 @@ derive_builder.workspace = true
 futures.workspace = true
 lazy_static.workspace = true
 object-store.workspace = true
-orc-rust = { version = "0.5", default-features = false, features = [
+object_store_opendal.workspace = true
+orc-rust = { git = "https://github.com/datafusion-contrib/orc-rust", rev = "3134cab581a8e91b942d6a23aca2916ea965f6bb", default-features = false, features = [
     "async",
 ] }
 parquet.workspace = true

src/common/datasource/src/compression.rs

@@ -19,6 +19,7 @@ use std::str::FromStr;
 use async_compression::tokio::bufread::{BzDecoder, GzipDecoder, XzDecoder, ZstdDecoder};
 use async_compression::tokio::write;
 use bytes::Bytes;
+use datafusion::datasource::file_format::file_compression_type::FileCompressionType;
 use futures::Stream;
 use serde::{Deserialize, Serialize};
 use strum::EnumIter;
@@ -192,3 +193,15 @@ macro_rules! impl_compression_type {
 }
 
 impl_compression_type!((Gzip, Gzip), (Bzip2, Bz), (Xz, Xz), (Zstd, Zstd));
+
+impl From<CompressionType> for FileCompressionType {
+    fn from(t: CompressionType) -> Self {
+        match t {
+            CompressionType::Gzip => FileCompressionType::GZIP,
+            CompressionType::Bzip2 => FileCompressionType::BZIP2,
+            CompressionType::Xz => FileCompressionType::XZ,
+            CompressionType::Zstd => FileCompressionType::ZSTD,
+            CompressionType::Uncompressed => FileCompressionType::UNCOMPRESSED,
+        }
+    }
+}

src/common/datasource/src/file_format/csv.rs

@@ -14,28 +14,23 @@
 
 use std::collections::HashMap;
 use std::str::FromStr;
-use std::sync::Arc;
 
 use arrow::csv;
 use arrow::csv::reader::Format;
 use arrow::record_batch::RecordBatch;
-use arrow_schema::{Schema, SchemaRef};
+use arrow_schema::Schema;
 use async_trait::async_trait;
 use common_runtime;
-use datafusion::datasource::physical_plan::{FileMeta, FileOpenFuture, FileOpener};
-use datafusion::error::Result as DataFusionResult;
 use datafusion::physical_plan::SendableRecordBatchStream;
-use derive_builder::Builder;
 use object_store::ObjectStore;
 use snafu::ResultExt;
 use tokio_util::compat::FuturesAsyncReadCompatExt;
 use tokio_util::io::SyncIoBridge;
 
-use super::stream_to_file;
 use crate::buffered_writer::DfRecordBatchEncoder;
 use crate::compression::CompressionType;
 use crate::error::{self, Result};
-use crate::file_format::{self, open_with_decoder, FileFormat};
+use crate::file_format::{self, stream_to_file, FileFormat};
 use crate::share_buffer::SharedBuffer;
 
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
@@ -100,66 +95,6 @@ impl Default for CsvFormat {
     }
 }
 
-#[derive(Debug, Clone, Builder)]
-pub struct CsvConfig {
-    batch_size: usize,
-    file_schema: SchemaRef,
-    #[builder(default = "None")]
-    file_projection: Option<Vec<usize>>,
-    #[builder(default = "true")]
-    has_header: bool,
-    #[builder(default = "b','")]
-    delimiter: u8,
-}
-
-impl CsvConfig {
-    fn builder(&self) -> csv::ReaderBuilder {
-        let mut builder = csv::ReaderBuilder::new(self.file_schema.clone())
-            .with_delimiter(self.delimiter)
-            .with_batch_size(self.batch_size)
-            .with_header(self.has_header);
-
-        if let Some(proj) = &self.file_projection {
-            builder = builder.with_projection(proj.clone());
-        }
-
-        builder
-    }
-}
-
-#[derive(Debug, Clone)]
-pub struct CsvOpener {
-    config: Arc<CsvConfig>,
-    object_store: Arc<ObjectStore>,
-    compression_type: CompressionType,
-}
-
-impl CsvOpener {
-    /// Return a new [`CsvOpener`]. The caller must ensure [`CsvConfig`].file_schema must correspond to the opening file.
-    pub fn new(
-        config: CsvConfig,
-        object_store: ObjectStore,
-        compression_type: CompressionType,
-    ) -> Self {
-        CsvOpener {
-            config: Arc::new(config),
-            object_store: Arc::new(object_store),
-            compression_type,
-        }
-    }
-}
-
-impl FileOpener for CsvOpener {
-    fn open(&self, meta: FileMeta) -> DataFusionResult<FileOpenFuture> {
-        open_with_decoder(
-            self.object_store.clone(),
-            meta.location().to_string(),
-            self.compression_type,
-            || Ok(self.config.builder().build_decoder()),
-        )
-    }
-}
-
 #[async_trait]
 impl FileFormat for CsvFormat {
     async fn infer_schema(&self, store: &ObjectStore, path: &str) -> Result<Schema> {

src/common/datasource/src/file_format/json.rs

@@ -15,29 +15,24 @@
 use std::collections::HashMap;
 use std::io::BufReader;
 use std::str::FromStr;
-use std::sync::Arc;
 
-use arrow::datatypes::SchemaRef;
+use arrow::json;
 use arrow::json::reader::{infer_json_schema_from_iterator, ValueIter};
 use arrow::json::writer::LineDelimited;
-use arrow::json::{self, ReaderBuilder};
 use arrow::record_batch::RecordBatch;
 use arrow_schema::Schema;
 use async_trait::async_trait;
 use common_runtime;
-use datafusion::datasource::physical_plan::{FileMeta, FileOpenFuture, FileOpener};
-use datafusion::error::{DataFusionError, Result as DataFusionResult};
 use datafusion::physical_plan::SendableRecordBatchStream;
 use object_store::ObjectStore;
 use snafu::ResultExt;
 use tokio_util::compat::FuturesAsyncReadCompatExt;
 use tokio_util::io::SyncIoBridge;
 
-use super::stream_to_file;
 use crate::buffered_writer::DfRecordBatchEncoder;
 use crate::compression::CompressionType;
 use crate::error::{self, Result};
-use crate::file_format::{self, open_with_decoder, FileFormat};
+use crate::file_format::{self, stream_to_file, FileFormat};
 use crate::share_buffer::SharedBuffer;
 
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
@@ -114,47 +109,6 @@ impl FileFormat for JsonFormat {
     }
 }
 
-#[derive(Debug, Clone)]
-pub struct JsonOpener {
-    batch_size: usize,
-    projected_schema: SchemaRef,
-    object_store: Arc<ObjectStore>,
-    compression_type: CompressionType,
-}
-
-impl JsonOpener {
-    /// Return a new [`JsonOpener`]. Any fields not present in `projected_schema` will be ignored.
-    pub fn new(
-        batch_size: usize,
-        projected_schema: SchemaRef,
-        object_store: ObjectStore,
-        compression_type: CompressionType,
-    ) -> Self {
-        Self {
-            batch_size,
-            projected_schema,
-            object_store: Arc::new(object_store),
-            compression_type,
-        }
-    }
-}
-
-impl FileOpener for JsonOpener {
-    fn open(&self, meta: FileMeta) -> DataFusionResult<FileOpenFuture> {
-        open_with_decoder(
-            self.object_store.clone(),
-            meta.location().to_string(),
-            self.compression_type,
-            || {
-                ReaderBuilder::new(self.projected_schema.clone())
-                    .with_batch_size(self.batch_size)
-                    .build_decoder()
-                    .map_err(DataFusionError::from)
-            },
-        )
-    }
-}
-
 pub async fn stream_to_json(
     stream: SendableRecordBatchStream,
     store: ObjectStore,

src/common/datasource/src/file_format/tests.rs

@@ -19,7 +19,10 @@ use std::vec;
 
 use common_test_util::find_workspace_path;
 use datafusion::assert_batches_eq;
-use datafusion::datasource::physical_plan::{FileOpener, FileScanConfig, FileStream, ParquetExec};
+use datafusion::datasource::file_format::file_compression_type::FileCompressionType;
+use datafusion::datasource::physical_plan::{
+    CsvConfig, CsvOpener, FileOpener, FileScanConfig, FileStream, JsonOpener, ParquetExec,
+};
 use datafusion::execution::context::TaskContext;
 use datafusion::physical_plan::metrics::ExecutionPlanMetricsSet;
 use datafusion::physical_plan::ExecutionPlan;
@@ -27,14 +30,11 @@ use datafusion::prelude::SessionContext;
 use futures::StreamExt;
 
 use super::FORMAT_TYPE;
-use crate::compression::CompressionType;
-use crate::error;
-use crate::file_format::csv::{CsvConfigBuilder, CsvOpener};
-use crate::file_format::json::JsonOpener;
 use crate::file_format::orc::{OrcFormat, OrcOpener};
 use crate::file_format::parquet::DefaultParquetFileReaderFactory;
 use crate::file_format::{FileFormat, Format};
-use crate::test_util::{self, scan_config, test_basic_schema, test_store};
+use crate::test_util::{scan_config, test_basic_schema, test_store};
+use crate::{error, test_util};
 
 struct Test<'a, T: FileOpener> {
     config: FileScanConfig,
@@ -62,15 +62,18 @@ impl<T: FileOpener> Test<'_, T> {
 #[tokio::test]
 async fn test_json_opener() {
     let store = test_store("/");
+    let store = Arc::new(object_store_opendal::OpendalStore::new(store));
 
     let schema = test_basic_schema();
 
-    let json_opener = JsonOpener::new(
-        100,
-        schema.clone(),
-        store.clone(),
-        CompressionType::Uncompressed,
-    );
+    let json_opener = || {
+        JsonOpener::new(
+            test_util::TEST_BATCH_SIZE,
+            schema.clone(),
+            FileCompressionType::UNCOMPRESSED,
+            store.clone(),
+        )
+    };
 
     let path = &find_workspace_path("/src/common/datasource/tests/json/basic.json")
         .display()
@@ -78,7 +81,7 @@ async fn test_json_opener() {
     let tests = [
         Test {
             config: scan_config(schema.clone(), None, path),
-            opener: json_opener.clone(),
+            opener: json_opener(),
             expected: vec![
                 "+-----+-------+",
                 "| num | str   |",
@@ -91,7 +94,7 @@ async fn test_json_opener() {
         },
         Test {
             config: scan_config(schema.clone(), Some(1), path),
-            opener: json_opener.clone(),
+            opener: json_opener(),
             expected: vec![
                 "+-----+------+",
                 "| num | str  |",
@@ -110,23 +113,30 @@ async fn test_json_opener() {
 #[tokio::test]
 async fn test_csv_opener() {
     let store = test_store("/");
+    let store = Arc::new(object_store_opendal::OpendalStore::new(store));
 
     let schema = test_basic_schema();
     let path = &find_workspace_path("/src/common/datasource/tests/csv/basic.csv")
         .display()
         .to_string();
-    let csv_conf = CsvConfigBuilder::default()
-        .batch_size(test_util::TEST_BATCH_SIZE)
-        .file_schema(schema.clone())
-        .build()
-        .unwrap();
+    let csv_config = Arc::new(CsvConfig::new(
+        test_util::TEST_BATCH_SIZE,
+        schema.clone(),
+        None,
+        true,
+        b',',
+        b'"',
+        None,
+        store,
+        None,
+    ));
 
-    let csv_opener = CsvOpener::new(csv_conf, store, CompressionType::Uncompressed);
+    let csv_opener = || CsvOpener::new(csv_config.clone(), FileCompressionType::UNCOMPRESSED);
 
     let tests = [
         Test {
             config: scan_config(schema.clone(), None, path),
-            opener: csv_opener.clone(),
+            opener: csv_opener(),
             expected: vec![
                 "+-----+-------+",
                 "| num | str   |",
@@ -139,7 +149,7 @@ async fn test_csv_opener() {
         },
         Test {
             config: scan_config(schema.clone(), Some(1), path),
-            opener: csv_opener.clone(),
+            opener: csv_opener(),
             expected: vec![
                 "+-----+------+",
                 "| num | str  |",

src/common/datasource/src/test_util.rs

@@ -16,17 +16,19 @@ use std::sync::Arc;
 
 use arrow_schema::{DataType, Field, Schema, SchemaRef};
 use common_test_util::temp_dir::{create_temp_dir, TempDir};
-use datafusion::common::Statistics;
+use datafusion::common::{Constraints, Statistics};
+use datafusion::datasource::file_format::file_compression_type::FileCompressionType;
 use datafusion::datasource::listing::PartitionedFile;
 use datafusion::datasource::object_store::ObjectStoreUrl;
-use datafusion::datasource::physical_plan::{FileScanConfig, FileStream};
+use datafusion::datasource::physical_plan::{
+    CsvConfig, CsvOpener, FileScanConfig, FileStream, JsonOpener,
+};
 use datafusion::physical_plan::metrics::ExecutionPlanMetricsSet;
 use object_store::services::Fs;
 use object_store::ObjectStore;
 
-use crate::compression::CompressionType;
-use crate::file_format::csv::{stream_to_csv, CsvConfigBuilder, CsvOpener};
-use crate::file_format::json::{stream_to_json, JsonOpener};
+use crate::file_format::csv::stream_to_csv;
+use crate::file_format::json::stream_to_json;
 use crate::test_util;
 
 pub const TEST_BATCH_SIZE: usize = 100;
@@ -74,6 +76,7 @@ pub fn scan_config(file_schema: SchemaRef, limit: Option<usize>, filename: &str)
         object_store_url: ObjectStoreUrl::parse("empty://").unwrap(), // won't be used
         file_schema,
         file_groups: vec![vec![PartitionedFile::new(filename.to_string(), 10)]],
+        constraints: Constraints::empty(),
         statistics,
         projection: None,
         limit,
@@ -90,8 +93,8 @@ pub async fn setup_stream_to_json_test(origin_path: &str, threshold: impl Fn(usi
     let json_opener = JsonOpener::new(
         test_util::TEST_BATCH_SIZE,
         schema.clone(),
-        store.clone(),
-        CompressionType::Uncompressed,
+        FileCompressionType::UNCOMPRESSED,
+        Arc::new(object_store_opendal::OpendalStore::new(store.clone())),
     );
 
     let size = store.read(origin_path).await.unwrap().len();
@@ -124,13 +127,19 @@ pub async fn setup_stream_to_csv_test(origin_path: &str, threshold: impl Fn(usiz
 
     let schema = test_basic_schema();
 
-    let csv_conf = CsvConfigBuilder::default()
-        .batch_size(test_util::TEST_BATCH_SIZE)
-        .file_schema(schema.clone())
-        .build()
-        .unwrap();
+    let csv_config = Arc::new(CsvConfig::new(
+        TEST_BATCH_SIZE,
+        schema.clone(),
+        None,
+        true,
+        b',',
+        b'"',
+        None,
+        Arc::new(object_store_opendal::OpendalStore::new(store.clone())),
+        None,
+    ));
 
-    let csv_opener = CsvOpener::new(csv_conf, store.clone(), CompressionType::Uncompressed);
+    let csv_opener = CsvOpener::new(csv_config, FileCompressionType::UNCOMPRESSED);
 
     let size = store.read(origin_path).await.unwrap().len();
 

src/common/error/Cargo.toml

@@ -12,3 +12,6 @@ http.workspace = true
 snafu.workspace = true
 strum.workspace = true
 tonic.workspace = true
+
+[dev-dependencies]
+common-macro.workspace = true

src/common/error/src/ext.rs

@@ -42,7 +42,7 @@ pub trait ErrorExt: StackError {
                 if let Some(external_error) = error.source() {
                     let external_root = external_error.sources().last().unwrap();
 
-                    if error.to_string().is_empty() {
+                    if error.transparent() {
                         format!("{external_root}")
                     } else {
                         format!("{error}: {external_root}")
@@ -86,6 +86,14 @@ pub trait StackError: std::error::Error {
         }
         result
     }
+
+    /// Indicates whether this error is "transparent", that it delegates its "display" and "source"
+    /// to the underlying error. Could be useful when you are just wrapping some external error,
+    /// **AND** can not or would not provide meaningful contextual info. For example, the
+    /// `DataFusionError`.
+    fn transparent(&self) -> bool {
+        false
+    }
 }
 
 impl<T: ?Sized + StackError> StackError for Arc<T> {

src/common/error/src/status_code.rs

@@ -34,12 +34,14 @@ pub enum StatusCode {
     Internal = 1003,
     /// Invalid arguments.
     InvalidArguments = 1004,
-    /// The task is cancelled.
+    /// The task is cancelled (typically caller-side).
     Cancelled = 1005,
     /// Illegal state, can be exposed to users.
     IllegalState = 1006,
     /// Caused by some error originated from external system.
     External = 1007,
+    /// The request is deadline exceeded (typically server-side).
+    DeadlineExceeded = 1008,
     // ====== End of common status code ================
 
     // ====== Begin of SQL related status code =========
@@ -142,6 +144,7 @@ impl StatusCode {
             | StatusCode::Unexpected
             | StatusCode::InvalidArguments
             | StatusCode::Cancelled
+            | StatusCode::DeadlineExceeded
             | StatusCode::InvalidSyntax
             | StatusCode::DatabaseAlreadyExists
             | StatusCode::PlanQuery
@@ -177,6 +180,7 @@ impl StatusCode {
             | StatusCode::Unexpected
             | StatusCode::Internal
             | StatusCode::Cancelled
+            | StatusCode::DeadlineExceeded
             | StatusCode::IllegalState
             | StatusCode::EngineExecuteQuery
             | StatusCode::StorageUnavailable
@@ -272,6 +276,7 @@ pub fn status_to_tonic_code(status_code: StatusCode) -> Code {
             Code::InvalidArgument
         }
         StatusCode::Cancelled => Code::Cancelled,
+        StatusCode::DeadlineExceeded => Code::DeadlineExceeded,
         StatusCode::TableAlreadyExists
         | StatusCode::TableColumnExists
         | StatusCode::RegionAlreadyExists
@@ -299,6 +304,15 @@ pub fn status_to_tonic_code(status_code: StatusCode) -> Code {
     }
 }
 
+/// Converts tonic [Code] to [StatusCode].
+pub fn convert_tonic_code_to_status_code(code: Code) -> StatusCode {
+    match code {
+        Code::Cancelled => StatusCode::Cancelled,
+        Code::DeadlineExceeded => StatusCode::DeadlineExceeded,
+        _ => StatusCode::Internal,
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use strum::IntoEnumIterator;

src/common/error/tests/ext.rs

@@ -0,0 +1,115 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::any::Any;
+
+use common_error::ext::{ErrorExt, PlainError, StackError};
+use common_error::status_code::StatusCode;
+use common_macro::stack_trace_debug;
+use snafu::{Location, ResultExt, Snafu};
+
+#[derive(Snafu)]
+#[stack_trace_debug]
+enum MyError {
+    #[snafu(display(r#"A normal error with "display" attribute, message "{}""#, message))]
+    Normal {
+        message: String,
+        #[snafu(source)]
+        error: PlainError,
+        #[snafu(implicit)]
+        location: Location,
+    },
+
+    #[snafu(transparent)]
+    Transparent {
+        #[snafu(source)]
+        error: PlainError,
+        #[snafu(implicit)]
+        location: Location,
+    },
+}
+
+impl ErrorExt for MyError {
+    fn status_code(&self) -> StatusCode {
+        StatusCode::Unexpected
+    }
+
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+}
+
+fn normal_error() -> Result<(), MyError> {
+    let plain_error = PlainError::new("<root cause>".to_string(), StatusCode::Unexpected);
+    Err(plain_error).context(NormalSnafu { message: "blabla" })
+}
+
+fn transparent_error() -> Result<(), MyError> {
+    let plain_error = PlainError::new("<root cause>".to_string(), StatusCode::Unexpected);
+    Err(plain_error)?
+}
+
+#[test]
+fn test_output_msg() {
+    let result = normal_error();
+    assert_eq!(
+        result.unwrap_err().output_msg(),
+        r#"A normal error with "display" attribute, message "blabla": <root cause>"#
+    );
+
+    let result = transparent_error();
+    assert_eq!(result.unwrap_err().output_msg(), "<root cause>");
+}
+
+#[test]
+fn test_to_string() {
+    let result = normal_error();
+    assert_eq!(
+        result.unwrap_err().to_string(),
+        r#"A normal error with "display" attribute, message "blabla""#
+    );
+
+    let result = transparent_error();
+    assert_eq!(result.unwrap_err().to_string(), "<root cause>");
+}
+
+#[test]
+fn test_debug_format() {
+    let result = normal_error();
+    let debug_output = format!("{:?}", result.unwrap_err());
+    let normalized_output = debug_output.replace('\\', "/");
+    assert_eq!(
+        normalized_output,
+        r#"0: A normal error with "display" attribute, message "blabla", at src/common/error/tests/ext.rs:55:22
+1: PlainError { msg: "<root cause>", status_code: Unexpected }"#
+    );
+
+    let result = transparent_error();
+    let debug_output = format!("{:?}", result.unwrap_err());
+    let normalized_output = debug_output.replace('\\', "/");
+    assert_eq!(
+        normalized_output,
+        r#"0: <transparent>, at src/common/error/tests/ext.rs:60:5
+1: PlainError { msg: "<root cause>", status_code: Unexpected }"#
+    );
+}
+
+#[test]
+fn test_transparent_flag() {
+    let result = normal_error();
+    assert!(!result.unwrap_err().transparent());
+
+    let result = transparent_error();
+    assert!(result.unwrap_err().transparent());
+}

src/common/function/Cargo.toml

@@ -8,6 +8,7 @@ license.workspace = true
 workspace = true
 
 [features]
+testing = []
 default = ["geo"]
 geo = ["geohash", "h3o", "s2", "wkt", "geo-types", "dep:geo"]
 
@@ -17,6 +18,7 @@ api.workspace = true
 arc-swap = "1.0"
 async-trait.workspace = true
 bincode = "1.3"
+catalog.workspace = true
 chrono.workspace = true
 common-base.workspace = true
 common-catalog.workspace = true
@@ -39,6 +41,7 @@ geohash = { version = "0.13", optional = true }
 h3o = { version = "0.6", optional = true }
 hyperloglogplus = "0.4"
 jsonb.workspace = true
+memchr = "2.7"
 nalgebra.workspace = true
 num = "0.4"
 num-traits = "0.2"

src/common/function/src/admin.rs

@@ -12,15 +12,21 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+mod add_region_follower;
 mod flush_compact_region;
 mod flush_compact_table;
+mod metadata_snaphost;
 mod migrate_region;
+mod remove_region_follower;
 
 use std::sync::Arc;
 
+use add_region_follower::AddRegionFollowerFunction;
 use flush_compact_region::{CompactRegionFunction, FlushRegionFunction};
 use flush_compact_table::{CompactTableFunction, FlushTableFunction};
+use metadata_snaphost::{DumpMetadataFunction, RestoreMetadataFunction};
 use migrate_region::MigrateRegionFunction;
+use remove_region_follower::RemoveRegionFollowerFunction;
 
 use crate::flush_flow::FlushFlowFunction;
 use crate::function_registry::FunctionRegistry;
@@ -32,10 +38,14 @@ impl AdminFunction {
     /// Register all table functions to [`FunctionRegistry`].
     pub fn register(registry: &FunctionRegistry) {
         registry.register_async(Arc::new(MigrateRegionFunction));
+        registry.register_async(Arc::new(AddRegionFollowerFunction));
+        registry.register_async(Arc::new(RemoveRegionFollowerFunction));
         registry.register_async(Arc::new(FlushRegionFunction));
         registry.register_async(Arc::new(CompactRegionFunction));
         registry.register_async(Arc::new(FlushTableFunction));
         registry.register_async(Arc::new(CompactTableFunction));
         registry.register_async(Arc::new(FlushFlowFunction));
+        registry.register_async(Arc::new(DumpMetadataFunction));
+        registry.register_async(Arc::new(RestoreMetadataFunction));
     }
 }

src/common/function/src/admin/add_region_follower.rs

@@ -0,0 +1,129 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use common_macro::admin_fn;
+use common_meta::rpc::procedure::AddRegionFollowerRequest;
+use common_query::error::{
+    InvalidFuncArgsSnafu, MissingProcedureServiceHandlerSnafu, Result,
+    UnsupportedInputDataTypeSnafu,
+};
+use common_query::prelude::{Signature, TypeSignature, Volatility};
+use datatypes::prelude::ConcreteDataType;
+use datatypes::value::{Value, ValueRef};
+use session::context::QueryContextRef;
+use snafu::ensure;
+
+use crate::handlers::ProcedureServiceHandlerRef;
+use crate::helper::cast_u64;
+
+/// A function to add a follower to a region.
+/// Only available in cluster mode.
+///
+/// - `add_region_follower(region_id, peer_id)`.
+///
+/// The parameters:
+/// - `region_id`:  the region id
+/// - `peer_id`:  the peer id
+#[admin_fn(
+    name = AddRegionFollowerFunction,
+    display_name = add_region_follower,
+    sig_fn = signature,
+    ret = uint64
+)]
+pub(crate) async fn add_region_follower(
+    procedure_service_handler: &ProcedureServiceHandlerRef,
+    _ctx: &QueryContextRef,
+    params: &[ValueRef<'_>],
+) -> Result<Value> {
+    ensure!(
+        params.len() == 2,
+        InvalidFuncArgsSnafu {
+            err_msg: format!(
+                "The length of the args is not correct, expect exactly 2, have: {}",
+                params.len()
+            ),
+        }
+    );
+
+    let Some(region_id) = cast_u64(&params[0])? else {
+        return UnsupportedInputDataTypeSnafu {
+            function: "add_region_follower",
+            datatypes: params.iter().map(|v| v.data_type()).collect::<Vec<_>>(),
+        }
+        .fail();
+    };
+    let Some(peer_id) = cast_u64(&params[1])? else {
+        return UnsupportedInputDataTypeSnafu {
+            function: "add_region_follower",
+            datatypes: params.iter().map(|v| v.data_type()).collect::<Vec<_>>(),
+        }
+        .fail();
+    };
+
+    procedure_service_handler
+        .add_region_follower(AddRegionFollowerRequest { region_id, peer_id })
+        .await?;
+
+    Ok(Value::from(0u64))
+}
+
+fn signature() -> Signature {
+    Signature::one_of(
+        vec![
+            // add_region_follower(region_id, peer)
+            TypeSignature::Uniform(2, ConcreteDataType::numerics()),
+        ],
+        Volatility::Immutable,
+    )
+}
+
+#[cfg(test)]
+mod tests {
+    use std::sync::Arc;
+
+    use common_query::prelude::TypeSignature;
+    use datatypes::vectors::{UInt64Vector, VectorRef};
+
+    use super::*;
+    use crate::function::{AsyncFunction, FunctionContext};
+
+    #[test]
+    fn test_add_region_follower_misc() {
+        let f = AddRegionFollowerFunction;
+        assert_eq!("add_region_follower", f.name());
+        assert_eq!(
+            ConcreteDataType::uint64_datatype(),
+            f.return_type(&[]).unwrap()
+        );
+        assert!(matches!(f.signature(),
+                         Signature {
+                             type_signature: TypeSignature::OneOf(sigs),
+                             volatility: Volatility::Immutable
+                         } if sigs.len() == 1));
+    }
+
+    #[tokio::test]
+    async fn test_add_region_follower() {
+        let f = AddRegionFollowerFunction;
+        let args = vec![1, 1];
+        let args = args
+            .into_iter()
+            .map(|arg| Arc::new(UInt64Vector::from_slice([arg])) as _)
+            .collect::<Vec<_>>();
+
+        let result = f.eval(FunctionContext::mock(), &args).await.unwrap();
+        let expect: VectorRef = Arc::new(UInt64Vector::from_slice([0u64]));
+        assert_eq!(result, expect);
+    }
+}

src/common/function/src/admin/metadata_snaphost.rs

@@ -0,0 +1,56 @@
+use common_macro::admin_fn;
+use common_query::error::{MissingMetadataSnapshotHandlerSnafu, Result};
+use common_query::prelude::{Signature, Volatility};
+use datatypes::prelude::*;
+use session::context::QueryContextRef;
+
+use crate::handlers::MetadataSnapshotHandlerRef;
+
+const METADATA_DIR: &str = "/snaphost/";
+const METADATA_FILE_NAME: &str = "dump_metadata";
+const METADATA_FILE_EXTENSION: &str = "metadata.fb";
+
+#[admin_fn(
+    name = DumpMetadataFunction,
+    display_name = dump_metadata,
+    sig_fn = dump_signature,
+    ret = string
+)]
+pub(crate) async fn dump_metadata(
+    metadata_snapshot_handler: &MetadataSnapshotHandlerRef,
+    _query_ctx: &QueryContextRef,
+    _params: &[ValueRef<'_>],
+) -> Result<Value> {
+    let filename = metadata_snapshot_handler
+        .dump(METADATA_DIR, METADATA_FILE_NAME)
+        .await?;
+    Ok(Value::from(filename))
+}
+
+fn dump_signature() -> Signature {
+    Signature::uniform(0, vec![], Volatility::Immutable)
+}
+
+#[admin_fn(
+    name = RestoreMetadataFunction,
+    display_name = restore_metadata,
+    sig_fn = restore_signature,
+    ret = uint64,
+)]
+pub(crate) async fn restore_metadata(
+    metadata_snapshot_handler: &MetadataSnapshotHandlerRef,
+    _query_ctx: &QueryContextRef,
+    _params: &[ValueRef<'_>],
+) -> Result<Value> {
+    let num_keyvalues = metadata_snapshot_handler
+        .restore(
+            METADATA_DIR,
+            &format!("{METADATA_FILE_NAME}.{METADATA_FILE_EXTENSION}"),
+        )
+        .await?;
+    Ok(Value::from(num_keyvalues))
+}
+
+fn restore_signature() -> Signature {
+    Signature::uniform(0, vec![], Volatility::Immutable)
+}

src/common/function/src/admin/migrate_region.rs

@@ -25,12 +25,13 @@ use session::context::QueryContextRef;
 use crate::handlers::ProcedureServiceHandlerRef;
 use crate::helper::cast_u64;
 
-const DEFAULT_TIMEOUT_SECS: u64 = 30;
+/// The default timeout for migrate region procedure.
+const DEFAULT_TIMEOUT_SECS: u64 = 300;
 
 /// A function to migrate a region from source peer to target peer.
 /// Returns the submitted procedure id if success. Only available in cluster mode.
 ///
-/// - `migrate_region(region_id, from_peer, to_peer)`, with timeout(30 seconds).
+/// - `migrate_region(region_id, from_peer, to_peer)`, with timeout(300 seconds).
 /// - `migrate_region(region_id, from_peer, to_peer, timeout(secs))`.
 ///
 /// The parameters:

src/common/function/src/admin/remove_region_follower.rs

@@ -0,0 +1,129 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use common_macro::admin_fn;
+use common_meta::rpc::procedure::RemoveRegionFollowerRequest;
+use common_query::error::{
+    InvalidFuncArgsSnafu, MissingProcedureServiceHandlerSnafu, Result,
+    UnsupportedInputDataTypeSnafu,
+};
+use common_query::prelude::{Signature, TypeSignature, Volatility};
+use datatypes::prelude::ConcreteDataType;
+use datatypes::value::{Value, ValueRef};
+use session::context::QueryContextRef;
+use snafu::ensure;
+
+use crate::handlers::ProcedureServiceHandlerRef;
+use crate::helper::cast_u64;
+
+/// A function to remove a follower from a region.
+//// Only available in cluster mode.
+///
+/// - `remove_region_follower(region_id, peer_id)`.
+///
+/// The parameters:
+/// - `region_id`:  the region id
+/// - `peer_id`:  the peer id
+#[admin_fn(
+    name = RemoveRegionFollowerFunction,
+    display_name = remove_region_follower,
+    sig_fn = signature,
+    ret = uint64
+)]
+pub(crate) async fn remove_region_follower(
+    procedure_service_handler: &ProcedureServiceHandlerRef,
+    _ctx: &QueryContextRef,
+    params: &[ValueRef<'_>],
+) -> Result<Value> {
+    ensure!(
+        params.len() == 2,
+        InvalidFuncArgsSnafu {
+            err_msg: format!(
+                "The length of the args is not correct, expect exactly 2, have: {}",
+                params.len()
+            ),
+        }
+    );
+
+    let Some(region_id) = cast_u64(&params[0])? else {
+        return UnsupportedInputDataTypeSnafu {
+            function: "add_region_follower",
+            datatypes: params.iter().map(|v| v.data_type()).collect::<Vec<_>>(),
+        }
+        .fail();
+    };
+    let Some(peer_id) = cast_u64(&params[1])? else {
+        return UnsupportedInputDataTypeSnafu {
+            function: "add_region_follower",
+            datatypes: params.iter().map(|v| v.data_type()).collect::<Vec<_>>(),
+        }
+        .fail();
+    };
+
+    procedure_service_handler
+        .remove_region_follower(RemoveRegionFollowerRequest { region_id, peer_id })
+        .await?;
+
+    Ok(Value::from(0u64))
+}
+
+fn signature() -> Signature {
+    Signature::one_of(
+        vec![
+            // remove_region_follower(region_id, peer_id)
+            TypeSignature::Uniform(2, ConcreteDataType::numerics()),
+        ],
+        Volatility::Immutable,
+    )
+}
+
+#[cfg(test)]
+mod tests {
+    use std::sync::Arc;
+
+    use common_query::prelude::TypeSignature;
+    use datatypes::vectors::{UInt64Vector, VectorRef};
+
+    use super::*;
+    use crate::function::{AsyncFunction, FunctionContext};
+
+    #[test]
+    fn test_remove_region_follower_misc() {
+        let f = RemoveRegionFollowerFunction;
+        assert_eq!("remove_region_follower", f.name());
+        assert_eq!(
+            ConcreteDataType::uint64_datatype(),
+            f.return_type(&[]).unwrap()
+        );
+        assert!(matches!(f.signature(),
+                         Signature {
+                             type_signature: TypeSignature::OneOf(sigs),
+                             volatility: Volatility::Immutable
+                         } if sigs.len() == 1));
+    }
+
+    #[tokio::test]
+    async fn test_remove_region_follower() {
+        let f = RemoveRegionFollowerFunction;
+        let args = vec![1, 1];
+        let args = args
+            .into_iter()
+            .map(|arg| Arc::new(UInt64Vector::from_slice([arg])) as _)
+            .collect::<Vec<_>>();
+
+        let result = f.eval(FunctionContext::mock(), &args).await.unwrap();
+        let expect: VectorRef = Arc::new(UInt64Vector::from_slice([0u64]));
+        assert_eq!(result, expect);
+    }
+}

src/common/function/src/function.rs

@@ -32,7 +32,7 @@ pub struct FunctionContext {
 
 impl FunctionContext {
     /// Create a mock [`FunctionContext`] for test.
-    #[cfg(test)]
+    #[cfg(any(test, feature = "testing"))]
     pub fn mock() -> Self {
         Self {
             query_ctx: QueryContextBuilder::default().build().into(),

src/common/function/src/function_registry.rs

@@ -27,6 +27,7 @@ use crate::scalars::hll_count::HllCalcFunction;
 use crate::scalars::ip::IpFunctions;
 use crate::scalars::json::JsonFunction;
 use crate::scalars::matches::MatchesFunction;
+use crate::scalars::matches_term::MatchesTermFunction;
 use crate::scalars::math::MathFunction;
 use crate::scalars::timestamp::TimestampFunction;
 use crate::scalars::uddsketch_calc::UddSketchCalcFunction;
@@ -116,6 +117,7 @@ pub static FUNCTION_REGISTRY: Lazy<Arc<FunctionRegistry>> = Lazy::new(|| {
 
     // Full text search function
     MatchesFunction::register(&function_registry);
+    MatchesTermFunction::register(&function_registry);
 
     // System and administration functions
     SystemFunction::register(&function_registry);

src/common/function/src/handlers.rs

@@ -15,8 +15,12 @@
 use std::sync::Arc;
 
 use async_trait::async_trait;
+use catalog::CatalogManagerRef;
 use common_base::AffectedRows;
-use common_meta::rpc::procedure::{MigrateRegionRequest, ProcedureStateResponse};
+use common_meta::rpc::procedure::{
+    AddRegionFollowerRequest, MigrateRegionRequest, ProcedureStateResponse,
+    RemoveRegionFollowerRequest,
+};
 use common_query::error::Result;
 use common_query::Output;
 use session::context::QueryContextRef;
@@ -63,6 +67,15 @@ pub trait ProcedureServiceHandler: Send + Sync {
 
     /// Query the procedure' state by its id
     async fn query_procedure_state(&self, pid: &str) -> Result<ProcedureStateResponse>;
+
+    /// Add a region follower to a region.
+    async fn add_region_follower(&self, request: AddRegionFollowerRequest) -> Result<()>;
+
+    /// Remove a region follower from a region.
+    async fn remove_region_follower(&self, request: RemoveRegionFollowerRequest) -> Result<()>;
+
+    /// Get the catalog manager
+    fn catalog_manager(&self) -> &CatalogManagerRef;
 }
 
 /// This flow service handler is only use for flush flow for now.
@@ -76,8 +89,18 @@ pub trait FlowServiceHandler: Send + Sync {
     ) -> Result<api::v1::flow::FlowResponse>;
 }
 
+/// This metadata snapshot handler is only use for dump and restore metadata for now.
+#[async_trait]
+pub trait MetadataSnapshotHandler: Send + Sync {
+    async fn dump(&self, path: &str, filename: &str) -> Result<String>;
+
+    async fn restore(&self, path: &str, filename: &str) -> Result<u64>;
+}
+
 pub type TableMutationHandlerRef = Arc<dyn TableMutationHandler>;
 
 pub type ProcedureServiceHandlerRef = Arc<dyn ProcedureServiceHandler>;
 
 pub type FlowServiceHandlerRef = Arc<dyn FlowServiceHandler>;
+
+pub type MetadataSnapshotHandlerRef = Arc<dyn MetadataSnapshotHandler>;

src/common/function/src/scalars.rs

@@ -19,6 +19,7 @@ pub mod expression;
 pub mod geo;
 pub mod json;
 pub mod matches;
+pub mod matches_term;
 pub mod math;
 pub mod vector;
 

src/common/function/src/scalars/geo/encoding.rs

@@ -27,7 +27,7 @@ use datatypes::value::{ListValue, Value};
 use datatypes::vectors::VectorRef;
 use snafu::{ensure, ResultExt};
 
-use super::helpers::{ensure_columns_len, ensure_columns_n};
+use crate::scalars::geo::helpers::{ensure_columns_len, ensure_columns_n};
 
 /// Accumulator of lat, lng, timestamp tuples
 #[derive(Debug)]

src/common/function/src/scalars/geo/h3.rs

@@ -31,8 +31,8 @@ use h3o::{CellIndex, LatLng, Resolution};
 use once_cell::sync::Lazy;
 use snafu::ResultExt;
 
-use super::helpers::{ensure_and_coerce, ensure_columns_len, ensure_columns_n};
 use crate::function::{Function, FunctionContext};
+use crate::scalars::geo::helpers::{ensure_and_coerce, ensure_columns_len, ensure_columns_n};
 
 static CELL_TYPES: Lazy<Vec<ConcreteDataType>> = Lazy::new(|| {
     vec![

src/common/function/src/scalars/geo/measure.rs

@@ -26,9 +26,9 @@ use geo::{Area, Distance, Haversine};
 use geo_types::Geometry;
 use snafu::ResultExt;
 
-use super::helpers::{ensure_columns_len, ensure_columns_n};
-use super::wkt::parse_wkt;
 use crate::function::{Function, FunctionContext};
+use crate::scalars::geo::helpers::{ensure_columns_len, ensure_columns_n};
+use crate::scalars::geo::wkt::parse_wkt;
 
 /// Return WGS84(SRID: 4326) euclidean distance between two geometry object, in degree
 #[derive(Clone, Debug, Default, Display)]

src/common/function/src/scalars/geo/relation.rs

@@ -23,9 +23,9 @@ use geo::algorithm::contains::Contains;
 use geo::algorithm::intersects::Intersects;
 use geo::algorithm::within::Within;
 
-use super::helpers::{ensure_columns_len, ensure_columns_n};
-use super::wkt::parse_wkt;
 use crate::function::{Function, FunctionContext};
+use crate::scalars::geo::helpers::{ensure_columns_len, ensure_columns_n};
+use crate::scalars::geo::wkt::parse_wkt;
 
 /// Test if spatial relationship: contains
 #[derive(Clone, Debug, Default, Display)]

src/common/function/src/scalars/geo/wkt.rs

@@ -26,8 +26,8 @@ use once_cell::sync::Lazy;
 use snafu::ResultExt;
 use wkt::{ToWkt, TryFromWkt};
 
-use super::helpers::{ensure_columns_len, ensure_columns_n};
 use crate::function::{Function, FunctionContext};
+use crate::scalars::geo::helpers::{ensure_columns_len, ensure_columns_n};
 
 static COORDINATE_TYPES: Lazy<Vec<ConcreteDataType>> = Lazy::new(|| {
     vec![

src/common/function/src/scalars/matches_term.rs

@@ -0,0 +1,375 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::Arc;
+use std::{fmt, iter};
+
+use common_query::error::{InvalidFuncArgsSnafu, Result};
+use common_query::prelude::Volatility;
+use datatypes::prelude::ConcreteDataType;
+use datatypes::scalars::ScalarVectorBuilder;
+use datatypes::vectors::{BooleanVector, BooleanVectorBuilder, MutableVector, VectorRef};
+use memchr::memmem;
+use snafu::ensure;
+
+use crate::function::{Function, FunctionContext};
+use crate::function_registry::FunctionRegistry;
+
+/// Exact term/phrase matching function for text columns.
+///
+/// This function checks if a text column contains exact term/phrase matches
+/// with non-alphanumeric boundaries. Designed for:
+/// - Whole-word matching (e.g. "cat" in "cat!" but not in "category")
+/// - Phrase matching (e.g. "hello world" in "note:hello world!")
+///
+/// # Signature
+/// `matches_term(text: String, term: String) -> Boolean`
+///
+/// # Arguments
+/// * `text` - String column to search
+/// * `term` - Search term/phrase
+///
+/// # Returns
+/// BooleanVector where each element indicates if the corresponding text
+/// contains an exact match of the term, following these rules:
+/// 1. Exact substring match found (case-sensitive)
+/// 2. Match boundaries are either:
+///    - Start/end of text
+///    - Any non-alphanumeric character (including spaces, hyphens, punctuation, etc.)
+///
+/// # Examples
+/// ```
+/// -- SQL examples --
+/// -- Match phrase with space --
+/// SELECT matches_term(column, 'hello world') FROM table;
+/// -- Text: "warning:hello world!" => true
+/// -- Text: "hello-world"          => false (hyphen instead of space)
+/// -- Text: "hello world2023"      => false (ending with numbers)
+///
+/// -- Match multiple words with boundaries --
+/// SELECT matches_term(column, 'critical error') FROM logs;
+/// -- Match in: "ERROR:critical error!"
+/// -- No match: "critical_errors"
+///
+/// -- Empty string handling --
+/// SELECT matches_term(column, '') FROM table;
+/// -- Text: "" => true
+/// -- Text: "any" => false
+///
+/// -- Case sensitivity --
+/// SELECT matches_term(column, 'Cat') FROM table;
+/// -- Text: "Cat" => true
+/// -- Text: "cat" => false
+/// ```
+pub struct MatchesTermFunction;
+
+impl MatchesTermFunction {
+    pub fn register(registry: &FunctionRegistry) {
+        registry.register(Arc::new(MatchesTermFunction));
+    }
+}
+
+impl fmt::Display for MatchesTermFunction {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "MATCHES_TERM")
+    }
+}
+
+impl Function for MatchesTermFunction {
+    fn name(&self) -> &str {
+        "matches_term"
+    }
+
+    fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::boolean_datatype())
+    }
+
+    fn signature(&self) -> common_query::prelude::Signature {
+        common_query::prelude::Signature::exact(
+            vec![
+                ConcreteDataType::string_datatype(),
+                ConcreteDataType::string_datatype(),
+            ],
+            Volatility::Immutable,
+        )
+    }
+
+    fn eval(&self, _func_ctx: &FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
+        ensure!(
+            columns.len() == 2,
+            InvalidFuncArgsSnafu {
+                err_msg: format!(
+                    "The length of the args is not correct, expect exactly 2, have: {}",
+                    columns.len()
+                ),
+            }
+        );
+
+        let text_column = &columns[0];
+        if text_column.is_empty() {
+            return Ok(Arc::new(BooleanVector::from(Vec::<bool>::with_capacity(0))));
+        }
+
+        let term_column = &columns[1];
+        let compiled_finder = if term_column.is_const() {
+            let term = term_column.get_ref(0).as_string().unwrap();
+            match term {
+                None => {
+                    return Ok(Arc::new(BooleanVector::from_iter(
+                        iter::repeat(None).take(text_column.len()),
+                    )));
+                }
+                Some(term) => Some(MatchesTermFinder::new(term)),
+            }
+        } else {
+            None
+        };
+
+        let len = text_column.len();
+        let mut result = BooleanVectorBuilder::with_capacity(len);
+        for i in 0..len {
+            let text = text_column.get_ref(i).as_string().unwrap();
+            let Some(text) = text else {
+                result.push_null();
+                continue;
+            };
+
+            let contains = match &compiled_finder {
+                Some(finder) => finder.find(text),
+                None => {
+                    let term = match term_column.get_ref(i).as_string().unwrap() {
+                        None => {
+                            result.push_null();
+                            continue;
+                        }
+                        Some(term) => term,
+                    };
+                    MatchesTermFinder::new(term).find(text)
+                }
+            };
+            result.push(Some(contains));
+        }
+
+        Ok(result.to_vector())
+    }
+}
+
+/// A compiled finder for `matches_term` function that holds the compiled term
+/// and its metadata for efficient matching.
+///
+/// A term is considered matched when:
+/// 1. The exact sequence appears in the text
+/// 2. It is either:
+///    - At the start/end of text with adjacent non-alphanumeric character
+///    - Surrounded by non-alphanumeric characters
+///
+/// # Examples
+/// ```
+/// let finder = MatchesTermFinder::new("cat");
+/// assert!(finder.find("cat!"));      // Term at end with punctuation
+/// assert!(finder.find("dog,cat"));   // Term preceded by comma
+/// assert!(!finder.find("category")); // Partial match rejected
+///
+/// let finder = MatchesTermFinder::new("world");
+/// assert!(finder.find("hello-world")); // Hyphen boundary
+/// ```
+#[derive(Clone, Debug)]
+pub struct MatchesTermFinder {
+    finder: memmem::Finder<'static>,
+    term: String,
+    starts_with_non_alnum: bool,
+    ends_with_non_alnum: bool,
+}
+
+impl MatchesTermFinder {
+    /// Create a new `MatchesTermFinder` for the given term.
+    pub fn new(term: &str) -> Self {
+        let starts_with_non_alnum = term.chars().next().is_some_and(|c| !c.is_alphanumeric());
+        let ends_with_non_alnum = term.chars().last().is_some_and(|c| !c.is_alphanumeric());
+
+        Self {
+            finder: memmem::Finder::new(term).into_owned(),
+            term: term.to_string(),
+            starts_with_non_alnum,
+            ends_with_non_alnum,
+        }
+    }
+
+    /// Find the term in the text.
+    pub fn find(&self, text: &str) -> bool {
+        if self.term.is_empty() {
+            return text.is_empty();
+        }
+
+        if text.len() < self.term.len() {
+            return false;
+        }
+
+        let mut pos = 0;
+        while let Some(found_pos) = self.finder.find(text[pos..].as_bytes()) {
+            let actual_pos = pos + found_pos;
+
+            let prev_ok = self.starts_with_non_alnum
+                || text[..actual_pos]
+                    .chars()
+                    .last()
+                    .map(|c| !c.is_alphanumeric())
+                    .unwrap_or(true);
+
+            if prev_ok {
+                let next_pos = actual_pos + self.finder.needle().len();
+                let next_ok = self.ends_with_non_alnum
+                    || text[next_pos..]
+                        .chars()
+                        .next()
+                        .map(|c| !c.is_alphanumeric())
+                        .unwrap_or(true);
+
+                if next_ok {
+                    return true;
+                }
+            }
+
+            if let Some(next_char) = text[actual_pos..].chars().next() {
+                pos = actual_pos + next_char.len_utf8();
+            } else {
+                break;
+            }
+        }
+
+        false
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn matches_term_example() {
+        let finder = MatchesTermFinder::new("hello world");
+        assert!(finder.find("warning:hello world!"));
+        assert!(!finder.find("hello-world"));
+        assert!(!finder.find("hello world2023"));
+
+        let finder = MatchesTermFinder::new("critical error");
+        assert!(finder.find("ERROR:critical error!"));
+        assert!(!finder.find("critical_errors"));
+
+        let finder = MatchesTermFinder::new("");
+        assert!(finder.find(""));
+        assert!(!finder.find("any"));
+
+        let finder = MatchesTermFinder::new("Cat");
+        assert!(finder.find("Cat"));
+        assert!(!finder.find("cat"));
+    }
+
+    #[test]
+    fn matches_term_with_punctuation() {
+        assert!(MatchesTermFinder::new("cat").find("cat!"));
+        assert!(MatchesTermFinder::new("dog").find("!dog"));
+    }
+
+    #[test]
+    fn matches_phrase_with_boundaries() {
+        assert!(MatchesTermFinder::new("hello-world").find("hello-world"));
+        assert!(MatchesTermFinder::new("'foo bar'").find("test: 'foo bar'"));
+    }
+
+    #[test]
+    fn matches_at_text_boundaries() {
+        assert!(MatchesTermFinder::new("start").find("start..."));
+        assert!(MatchesTermFinder::new("end").find("...end"));
+    }
+
+    // Negative cases
+    #[test]
+    fn rejects_partial_matches() {
+        assert!(!MatchesTermFinder::new("cat").find("category"));
+        assert!(!MatchesTermFinder::new("boot").find("rebooted"));
+    }
+
+    #[test]
+    fn rejects_missing_term() {
+        assert!(!MatchesTermFinder::new("foo").find("hello world"));
+    }
+
+    // Edge cases
+    #[test]
+    fn handles_empty_inputs() {
+        assert!(!MatchesTermFinder::new("test").find(""));
+        assert!(!MatchesTermFinder::new("").find("text"));
+    }
+
+    #[test]
+    fn different_unicode_boundaries() {
+        assert!(MatchesTermFinder::new("café").find("café>"));
+        assert!(!MatchesTermFinder::new("café").find("口café>"));
+        assert!(!MatchesTermFinder::new("café").find("café口"));
+        assert!(!MatchesTermFinder::new("café").find("cafémore"));
+        assert!(MatchesTermFinder::new("русский").find("русский!"));
+        assert!(MatchesTermFinder::new("русский").find("русский！"));
+    }
+
+    #[test]
+    fn case_sensitive_matching() {
+        assert!(!MatchesTermFinder::new("cat").find("Cat"));
+        assert!(MatchesTermFinder::new("CaT").find("CaT"));
+    }
+
+    #[test]
+    fn numbers_in_term() {
+        assert!(MatchesTermFinder::new("v1.0").find("v1.0!"));
+        assert!(!MatchesTermFinder::new("v1.0").find("v1.0a"));
+    }
+
+    #[test]
+    fn adjacent_alphanumeric_fails() {
+        assert!(!MatchesTermFinder::new("cat").find("cat5"));
+        assert!(!MatchesTermFinder::new("dog").find("dogcat"));
+    }
+
+    #[test]
+    fn empty_term_text() {
+        assert!(!MatchesTermFinder::new("").find("text"));
+        assert!(MatchesTermFinder::new("").find(""));
+        assert!(!MatchesTermFinder::new("text").find(""));
+    }
+
+    #[test]
+    fn leading_non_alphanumeric() {
+        assert!(MatchesTermFinder::new("/cat").find("dog/cat"));
+        assert!(MatchesTermFinder::new("dog/").find("dog/cat"));
+        assert!(MatchesTermFinder::new("dog/cat").find("dog/cat"));
+    }
+
+    #[test]
+    fn continues_searching_after_boundary_mismatch() {
+        assert!(!MatchesTermFinder::new("log").find("bloglog!"));
+        assert!(MatchesTermFinder::new("log").find("bloglog log"));
+        assert!(MatchesTermFinder::new("log").find("alogblog_log!"));
+
+        assert!(MatchesTermFinder::new("error").find("errorlog_error_case"));
+        assert!(MatchesTermFinder::new("test").find("atestbtestc_test_end"));
+        assert!(MatchesTermFinder::new("data").find("database_data_store"));
+        assert!(!MatchesTermFinder::new("data").find("database_datastore"));
+        assert!(MatchesTermFinder::new("log.txt").find("catalog.txt_log.txt!"));
+        assert!(!MatchesTermFinder::new("log.txt").find("catalog.txtlog.txt!"));
+        assert!(MatchesTermFinder::new("data-set").find("bigdata-set_data-set!"));
+
+        assert!(MatchesTermFinder::new("中文").find("这是中文测试，中文！"));
+        assert!(MatchesTermFinder::new("error").find("错误errorerror日志_error!"));
+    }
+}

src/common/function/src/scalars/udf.rs

@@ -163,7 +163,7 @@ mod tests {
         ];
 
         let args = ScalarFunctionArgs {
-            args: &args,
+            args,
             number_rows: 4,
             return_type: &ConcreteDataType::boolean_datatype().as_arrow_type(),
         };

src/common/function/src/state.rs

@@ -12,7 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use crate::handlers::{FlowServiceHandlerRef, ProcedureServiceHandlerRef, TableMutationHandlerRef};
+use crate::handlers::{
+    FlowServiceHandlerRef, MetadataSnapshotHandlerRef, ProcedureServiceHandlerRef,
+    TableMutationHandlerRef,
+};
 
 /// Shared state for SQL functions.
 /// The handlers in state may be `None` in cli command-line or test cases.
@@ -24,18 +27,24 @@ pub struct FunctionState {
     pub procedure_service_handler: Option<ProcedureServiceHandlerRef>,
     // The flownode handler
     pub flow_service_handler: Option<FlowServiceHandlerRef>,
+    // The metadata snapshot handler
+    pub metadata_snapshot_handler: Option<MetadataSnapshotHandlerRef>,
 }
 
 impl FunctionState {
     /// Create a mock [`FunctionState`] for test.
-    #[cfg(test)]
+    #[cfg(any(test, feature = "testing"))]
     pub fn mock() -> Self {
         use std::sync::Arc;
 
         use api::v1::meta::ProcedureStatus;
         use async_trait::async_trait;
+        use catalog::CatalogManagerRef;
         use common_base::AffectedRows;
-        use common_meta::rpc::procedure::{MigrateRegionRequest, ProcedureStateResponse};
+        use common_meta::rpc::procedure::{
+            AddRegionFollowerRequest, MigrateRegionRequest, ProcedureStateResponse,
+            RemoveRegionFollowerRequest,
+        };
         use common_query::error::Result;
         use common_query::Output;
         use session::context::QueryContextRef;
@@ -44,10 +53,14 @@ impl FunctionState {
             CompactTableRequest, DeleteRequest, FlushTableRequest, InsertRequest,
         };
 
-        use crate::handlers::{FlowServiceHandler, ProcedureServiceHandler, TableMutationHandler};
+        use crate::handlers::{
+            FlowServiceHandler, MetadataSnapshotHandler, ProcedureServiceHandler,
+            TableMutationHandler,
+        };
         struct MockProcedureServiceHandler;
         struct MockTableMutationHandler;
         struct MockFlowServiceHandler;
+        struct MockMetadataServiceHandler;
         const ROWS: usize = 42;
 
         #[async_trait]
@@ -66,6 +79,21 @@ impl FunctionState {
                     ..Default::default()
                 })
             }
+
+            async fn add_region_follower(&self, _request: AddRegionFollowerRequest) -> Result<()> {
+                Ok(())
+            }
+
+            async fn remove_region_follower(
+                &self,
+                _request: RemoveRegionFollowerRequest,
+            ) -> Result<()> {
+                Ok(())
+            }
+
+            fn catalog_manager(&self) -> &CatalogManagerRef {
+                unimplemented!()
+            }
         }
 
         #[async_trait]
@@ -131,10 +159,22 @@ impl FunctionState {
             }
         }
 
+        #[async_trait]
+        impl MetadataSnapshotHandler for MockMetadataServiceHandler {
+            async fn dump(&self, _path: &str, _filename: &str) -> Result<String> {
+                Ok("test_filename".to_string())
+            }
+
+            async fn restore(&self, _path: &str, _filename: &str) -> Result<u64> {
+                Ok(100)
+            }
+        }
+
         Self {
             table_mutation_handler: Some(Arc::new(MockTableMutationHandler)),
             procedure_service_handler: Some(Arc::new(MockProcedureServiceHandler)),
             flow_service_handler: Some(Arc::new(MockFlowServiceHandler)),
+            metadata_snapshot_handler: Some(Arc::new(MockMetadataServiceHandler)),
         }
     }
 }

src/common/function/src/system.rs

@@ -22,7 +22,9 @@ mod version;
 use std::sync::Arc;
 
 use build::BuildFunction;
-use database::{CurrentSchemaFunction, DatabaseFunction, SessionUserFunction};
+use database::{
+    CurrentSchemaFunction, DatabaseFunction, ReadPreferenceFunction, SessionUserFunction,
+};
 use pg_catalog::PGCatalogFunction;
 use procedure_state::ProcedureStateFunction;
 use timezone::TimezoneFunction;
@@ -39,6 +41,7 @@ impl SystemFunction {
         registry.register(Arc::new(CurrentSchemaFunction));
         registry.register(Arc::new(DatabaseFunction));
         registry.register(Arc::new(SessionUserFunction));
+        registry.register(Arc::new(ReadPreferenceFunction));
         registry.register(Arc::new(TimezoneFunction));
         registry.register_async(Arc::new(ProcedureStateFunction));
         PGCatalogFunction::register(registry);

src/common/function/src/system/database.rs

@@ -30,9 +30,12 @@ pub struct DatabaseFunction;
 pub struct CurrentSchemaFunction;
 pub struct SessionUserFunction;
 
+pub struct ReadPreferenceFunction;
+
 const DATABASE_FUNCTION_NAME: &str = "database";
 const CURRENT_SCHEMA_FUNCTION_NAME: &str = "current_schema";
 const SESSION_USER_FUNCTION_NAME: &str = "session_user";
+const READ_PREFERENCE_FUNCTION_NAME: &str = "read_preference";
 
 impl Function for DatabaseFunction {
     fn name(&self) -> &str {
@@ -94,6 +97,26 @@ impl Function for SessionUserFunction {
     }
 }
 
+impl Function for ReadPreferenceFunction {
+    fn name(&self) -> &str {
+        READ_PREFERENCE_FUNCTION_NAME
+    }
+
+    fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::string_datatype())
+    }
+
+    fn signature(&self) -> Signature {
+        Signature::nullary(Volatility::Immutable)
+    }
+
+    fn eval(&self, func_ctx: &FunctionContext, _columns: &[VectorRef]) -> Result<VectorRef> {
+        let read_preference = func_ctx.query_ctx.read_preference();
+
+        Ok(Arc::new(StringVector::from_slice(&[read_preference.as_ref()])) as _)
+    }
+}
+
 impl fmt::Display for DatabaseFunction {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "DATABASE")
@@ -112,6 +135,12 @@ impl fmt::Display for SessionUserFunction {
     }
 }
 
+impl fmt::Display for ReadPreferenceFunction {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "READ_PREFERENCE")
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use std::sync::Arc;

src/common/grpc-expr/src/alter.rs

@@ -15,11 +15,13 @@
 use api::helper::ColumnDataTypeWrapper;
 use api::v1::add_column_location::LocationType;
 use api::v1::alter_table_expr::Kind;
-use api::v1::column_def::{as_fulltext_option, as_skipping_index_type};
+use api::v1::column_def::{
+    as_fulltext_option_analyzer, as_fulltext_option_backend, as_skipping_index_type,
+};
 use api::v1::{
     column_def, AddColumnLocation as Location, AlterTableExpr, Analyzer, CreateTableExpr,
-    DropColumns, ModifyColumnTypes, RenameTable, SemanticType,
-    SkippingIndexType as PbSkippingIndexType,
+    DropColumns, FulltextBackend as PbFulltextBackend, ModifyColumnTypes, RenameTable,
+    SemanticType, SkippingIndexType as PbSkippingIndexType,
 };
 use common_query::AddColumnLocation;
 use datatypes::schema::{ColumnSchema, FulltextOptions, RawSchema, SkippingIndexOptions};
@@ -126,11 +128,15 @@ pub fn alter_expr_to_request(table_id: TableId, expr: AlterTableExpr) -> Result<
                         column_name: f.column_name.clone(),
                         options: FulltextOptions {
                             enable: f.enable,
-                            analyzer: as_fulltext_option(
+                            analyzer: as_fulltext_option_analyzer(
                                 Analyzer::try_from(f.analyzer)
                                     .context(InvalidSetFulltextOptionRequestSnafu)?,
                             ),
                             case_sensitive: f.case_sensitive,
+                            backend: as_fulltext_option_backend(
+                                PbFulltextBackend::try_from(f.backend)
+                                    .context(InvalidSetFulltextOptionRequestSnafu)?,
+                            ),
                         },
                     },
                 },

src/common/grpc/Cargo.toml

@@ -23,8 +23,11 @@ flatbuffers = "24"
 hyper.workspace = true
 lazy_static.workspace = true
 prost.workspace = true
+serde.workspace = true
+serde_json.workspace = true
 snafu.workspace = true
 tokio.workspace = true
+tokio-util.workspace = true
 tonic.workspace = true
 tower.workspace = true
 

src/common/grpc/benches/channel_manager.rs

@@ -25,7 +25,7 @@ async fn do_bench_channel_manager() {
         let m_clone = m.clone();
         let join = tokio::spawn(async move {
             for _ in 0..10000 {
-                let idx = rand::random::<usize>() % 100;
+                let idx = rand::random::<u32>() % 100;
                 let ret = m_clone.get(format!("{idx}"));
                 let _ = ret.unwrap();
             }

src/common/grpc/src/channel_manager.rs

@@ -22,6 +22,7 @@ use dashmap::mapref::entry::Entry;
 use dashmap::DashMap;
 use lazy_static::lazy_static;
 use snafu::{OptionExt, ResultExt};
+use tokio_util::sync::CancellationToken;
 use tonic::transport::{
     Certificate, Channel as InnerChannel, ClientTlsConfig, Endpoint, Identity, Uri,
 };
@@ -39,18 +40,48 @@ lazy_static! {
     static ref ID: AtomicU64 = AtomicU64::new(0);
 }
 
-#[derive(Clone, Debug)]
+#[derive(Clone, Debug, Default)]
 pub struct ChannelManager {
+    inner: Arc<Inner>,
+}
+
+#[derive(Debug)]
+struct Inner {
     id: u64,
     config: ChannelConfig,
     client_tls_config: Option<ClientTlsConfig>,
     pool: Arc<Pool>,
-    channel_recycle_started: Arc<AtomicBool>,
+    channel_recycle_started: AtomicBool,
+    cancel: CancellationToken,
 }
 
-impl Default for ChannelManager {
+impl Default for Inner {
     fn default() -> Self {
-        ChannelManager::with_config(ChannelConfig::default())
+        Self::with_config(ChannelConfig::default())
+    }
+}
+
+impl Drop for Inner {
+    fn drop(&mut self) {
+        // Cancel the channel recycle task.
+        self.cancel.cancel();
+    }
+}
+
+impl Inner {
+    fn with_config(config: ChannelConfig) -> Self {
+        let id = ID.fetch_add(1, Ordering::Relaxed);
+        let pool = Arc::new(Pool::default());
+        let cancel = CancellationToken::new();
+
+        Self {
+            id,
+            config,
+            client_tls_config: None,
+            pool,
+            channel_recycle_started: AtomicBool::new(false),
+            cancel,
+        }
     }
 }
 
@@ -60,19 +91,14 @@ impl ChannelManager {
     }
 
     pub fn with_config(config: ChannelConfig) -> Self {
-        let id = ID.fetch_add(1, Ordering::Relaxed);
-        let pool = Arc::new(Pool::default());
+        let inner = Inner::with_config(config);
         Self {
-            id,
-            config,
-            client_tls_config: None,
-            pool,
-            channel_recycle_started: Arc::new(AtomicBool::new(false)),
+            inner: Arc::new(inner),
         }
     }
 
     pub fn with_tls_config(config: ChannelConfig) -> Result<Self> {
-        let mut cm = Self::with_config(config.clone());
+        let mut inner = Inner::with_config(config.clone());
 
         // setup tls
         let path_config = config.client_tls.context(InvalidTlsConfigSnafu {
@@ -88,17 +114,23 @@ impl ChannelManager {
             .context(InvalidConfigFilePathSnafu)?;
         let client_identity = Identity::from_pem(client_cert, client_key);
 
-        cm.client_tls_config = Some(
+        inner.client_tls_config = Some(
             ClientTlsConfig::new()
                 .ca_certificate(server_root_ca_cert)
                 .identity(client_identity),
         );
 
-        Ok(cm)
+        Ok(Self {
+            inner: Arc::new(inner),
+        })
     }
 
     pub fn config(&self) -> &ChannelConfig {
-        &self.config
+        &self.inner.config
+    }
+
+    fn pool(&self) -> &Arc<Pool> {
+        &self.inner.pool
     }
 
     pub fn get(&self, addr: impl AsRef<str>) -> Result<InnerChannel> {
@@ -106,12 +138,12 @@ impl ChannelManager {
 
         let addr = addr.as_ref();
         // It will acquire the read lock.
-        if let Some(inner_ch) = self.pool.get(addr) {
+        if let Some(inner_ch) = self.pool().get(addr) {
             return Ok(inner_ch);
         }
 
         // It will acquire the write lock.
-        let entry = match self.pool.entry(addr.to_string()) {
+        let entry = match self.pool().entry(addr.to_string()) {
             Entry::Occupied(entry) => {
                 entry.get().increase_access();
                 entry.into_ref()
@@ -150,7 +182,7 @@ impl ChannelManager {
             access: AtomicUsize::new(1),
             use_default_connector: false,
         };
-        self.pool.put(addr, channel);
+        self.pool().put(addr, channel);
 
         Ok(inner_channel)
     }
@@ -159,11 +191,11 @@ impl ChannelManager {
     where
         F: FnMut(&String, &mut Channel) -> bool,
     {
-        self.pool.retain_channel(f);
+        self.pool().retain_channel(f);
     }
 
     fn build_endpoint(&self, addr: &str) -> Result<Endpoint> {
-        let http_prefix = if self.client_tls_config.is_some() {
+        let http_prefix = if self.inner.client_tls_config.is_some() {
             "https"
         } else {
             "http"
@@ -172,51 +204,52 @@ impl ChannelManager {
         let mut endpoint =
             Endpoint::new(format!("{http_prefix}://{addr}")).context(CreateChannelSnafu)?;
 
-        if let Some(dur) = self.config.timeout {
+        if let Some(dur) = self.config().timeout {
             endpoint = endpoint.timeout(dur);
         }
-        if let Some(dur) = self.config.connect_timeout {
+        if let Some(dur) = self.config().connect_timeout {
             endpoint = endpoint.connect_timeout(dur);
         }
-        if let Some(limit) = self.config.concurrency_limit {
+        if let Some(limit) = self.config().concurrency_limit {
             endpoint = endpoint.concurrency_limit(limit);
         }
-        if let Some((limit, dur)) = self.config.rate_limit {
+        if let Some((limit, dur)) = self.config().rate_limit {
             endpoint = endpoint.rate_limit(limit, dur);
         }
-        if let Some(size) = self.config.initial_stream_window_size {
+        if let Some(size) = self.config().initial_stream_window_size {
             endpoint = endpoint.initial_stream_window_size(size);
         }
-        if let Some(size) = self.config.initial_connection_window_size {
+        if let Some(size) = self.config().initial_connection_window_size {
             endpoint = endpoint.initial_connection_window_size(size);
         }
-        if let Some(dur) = self.config.http2_keep_alive_interval {
+        if let Some(dur) = self.config().http2_keep_alive_interval {
             endpoint = endpoint.http2_keep_alive_interval(dur);
         }
-        if let Some(dur) = self.config.http2_keep_alive_timeout {
+        if let Some(dur) = self.config().http2_keep_alive_timeout {
             endpoint = endpoint.keep_alive_timeout(dur);
         }
-        if let Some(enabled) = self.config.http2_keep_alive_while_idle {
+        if let Some(enabled) = self.config().http2_keep_alive_while_idle {
             endpoint = endpoint.keep_alive_while_idle(enabled);
         }
-        if let Some(enabled) = self.config.http2_adaptive_window {
+        if let Some(enabled) = self.config().http2_adaptive_window {
             endpoint = endpoint.http2_adaptive_window(enabled);
         }
-        if let Some(tls_config) = &self.client_tls_config {
+        if let Some(tls_config) = &self.inner.client_tls_config {
             endpoint = endpoint
                 .tls_config(tls_config.clone())
                 .context(CreateChannelSnafu)?;
         }
 
         endpoint = endpoint
-            .tcp_keepalive(self.config.tcp_keepalive)
-            .tcp_nodelay(self.config.tcp_nodelay);
+            .tcp_keepalive(self.config().tcp_keepalive)
+            .tcp_nodelay(self.config().tcp_nodelay);
 
         Ok(endpoint)
     }
 
     fn trigger_channel_recycling(&self) {
         if self
+            .inner
             .channel_recycle_started
             .compare_exchange(false, true, Ordering::Relaxed, Ordering::Relaxed)
             .is_err()
@@ -224,13 +257,15 @@ impl ChannelManager {
             return;
         }
 
-        let pool = self.pool.clone();
-        let _handle = common_runtime::spawn_global(async {
-            recycle_channel_in_loop(pool, RECYCLE_CHANNEL_INTERVAL_SECS).await;
+        let pool = self.pool().clone();
+        let cancel = self.inner.cancel.clone();
+        let id = self.inner.id;
+        let _handle = common_runtime::spawn_global(async move {
+            recycle_channel_in_loop(pool, id, cancel, RECYCLE_CHANNEL_INTERVAL_SECS).await;
         });
         info!(
             "ChannelManager: {}, channel recycle is started, running in the background!",
-            self.id
+            self.inner.id
         );
     }
 }
@@ -443,11 +478,23 @@ impl Pool {
     }
 }
 
-async fn recycle_channel_in_loop(pool: Arc<Pool>, interval_secs: u64) {
+async fn recycle_channel_in_loop(
+    pool: Arc<Pool>,
+    id: u64,
+    cancel: CancellationToken,
+    interval_secs: u64,
+) {
     let mut interval = tokio::time::interval(Duration::from_secs(interval_secs));
 
     loop {
-        let _ = interval.tick().await;
+        tokio::select! {
+            _ = cancel.cancelled() => {
+                info!("Stop channel recycle, ChannelManager id: {}", id);
+                break;
+            },
+            _ = interval.tick() => {}
+        }
+
         pool.retain_channel(|_, c| c.access.swap(0, Ordering::Relaxed) != 0)
     }
 }
@@ -461,11 +508,7 @@ mod tests {
     #[should_panic]
     #[test]
     fn test_invalid_addr() {
-        let pool = Arc::new(Pool::default());
-        let mgr = ChannelManager {
-            pool,
-            ..Default::default()
-        };
+        let mgr = ChannelManager::default();
         let addr = "http://test";
 
         let _ = mgr.get(addr).unwrap();
@@ -475,7 +518,9 @@ mod tests {
     async fn test_access_count() {
         let mgr = ChannelManager::new();
         // Do not start recycle
-        mgr.channel_recycle_started.store(true, Ordering::Relaxed);
+        mgr.inner
+            .channel_recycle_started
+            .store(true, Ordering::Relaxed);
         let mgr = Arc::new(mgr);
         let addr = "test_uri";
 
@@ -493,12 +538,12 @@ mod tests {
             join.await.unwrap();
         }
 
-        assert_eq!(1000, mgr.pool.get_access(addr).unwrap());
+        assert_eq!(1000, mgr.pool().get_access(addr).unwrap());
 
-        mgr.pool
+        mgr.pool()
             .retain_channel(|_, c| c.access.swap(0, Ordering::Relaxed) != 0);
 
-        assert_eq!(0, mgr.pool.get_access(addr).unwrap());
+        assert_eq!(0, mgr.pool().get_access(addr).unwrap());
     }
 
     #[test]
@@ -624,4 +669,49 @@ mod tests {
             true
         });
     }
+
+    #[tokio::test]
+    async fn test_pool_release_with_channel_recycle() {
+        let mgr = ChannelManager::new();
+
+        let pool_holder = mgr.pool().clone();
+
+        // start channel recycle task
+        let addr = "test_addr";
+        let _ = mgr.get(addr);
+
+        let mgr_clone_1 = mgr.clone();
+        let mgr_clone_2 = mgr.clone();
+        assert_eq!(3, Arc::strong_count(mgr.pool()));
+
+        drop(mgr_clone_1);
+        drop(mgr_clone_2);
+        assert_eq!(3, Arc::strong_count(mgr.pool()));
+
+        drop(mgr);
+
+        // wait for the channel recycle task to finish
+        tokio::time::sleep(Duration::from_millis(10)).await;
+
+        assert_eq!(1, Arc::strong_count(&pool_holder));
+    }
+
+    #[tokio::test]
+    async fn test_pool_release_without_channel_recycle() {
+        let mgr = ChannelManager::new();
+
+        let pool_holder = mgr.pool().clone();
+
+        let mgr_clone_1 = mgr.clone();
+        let mgr_clone_2 = mgr.clone();
+        assert_eq!(2, Arc::strong_count(mgr.pool()));
+
+        drop(mgr_clone_1);
+        drop(mgr_clone_2);
+        assert_eq!(2, Arc::strong_count(mgr.pool()));
+
+        drop(mgr);
+
+        assert_eq!(1, Arc::strong_count(&pool_holder));
+    }
 }