feat(iceberg): support iceberg engine table (in local env)

ci/docker-compose.yml

@@ -121,6 +121,13 @@ services:
     volumes:
       - ..:/risingwave
 
+  iceberg-engine-env:
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241030
+    depends_on:
+      - db
+    volumes:
+      - ..:/risingwave
+
   ci-flamegraph-env:
     image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241030
     # NOTE(kwannoel): This is used in order to permit

ci/scripts/e2e-iceberg-engine-test.sh

@@ -0,0 +1,49 @@
+#!/usr/bin/env bash
+
+# Exits as soon as any line fails.
+set -euo pipefail
+
+source ci/scripts/common.sh
+
+# prepare environment
+export CONNECTOR_LIBS_PATH="./connector-node/libs"
+export AWS_REGION="us-east-1"
+
+while getopts 'p:' opt; do
+    case ${opt} in
+        p )
+            profile=$OPTARG
+            ;;
+        \? )
+            echo "Invalid Option: -$OPTARG" 1>&2
+            exit 1
+            ;;
+        : )
+            echo "Invalid option: $OPTARG requires an argument" 1>&2
+            ;;
+    esac
+done
+shift $((OPTIND -1))
+
+download_and_prepare_rw "$profile" source
+
+echo "--- Download connector node package"
+buildkite-agent artifact download risingwave-connector.tar.gz ./
+mkdir ./connector-node
+tar xf ./risingwave-connector.tar.gz -C ./connector-node
+
+PGPASSWORD=postgres psql -h db -p 5432 -U postgres -c "DROP DATABASE IF EXISTS metadata;" -c "CREATE DATABASE metadata;"
+
+echo "--- starting risingwave cluster"
+mkdir -p .risingwave/log
+risedev ci-start ci-iceberg-engine
+sleep 1
+
+
+echo "--- testing iceberg engine"
+sqllogictest -p 4566 -d dev './e2e_test/iceberg/test_case/iceberg_engine.slt'
+sleep 1
+
+
+echo "--- Kill cluster"
+risedev ci-kill

ci/workflows/pull-request.yml

@@ -315,6 +315,21 @@ steps:
     timeout_in_minutes: 15
     retry: *auto-retry
 
+  - label: "end-to-end iceberg engine test"
+    if: build.pull_request.labels includes "ci/run-e2e-iceberg-engine-tests" || build.env("CI_STEPS") =~ /(^|,)e2e-iceberg-engine-tests?(,|$$)/
+    command: "ci/scripts/e2e-iceberg-engine-test.sh -p ci-dev"
+    depends_on:
+      - "build"
+      - "build-other"
+    plugins:
+      - docker-compose#v5.1.0:
+          run: iceberg-engine-env
+          config: ci/docker-compose.yml
+          mount-buildkite-agent: true
+      - ./ci/plugins/upload-failure-logs
+    timeout_in_minutes: 15
+    retry: *auto-retry
+
   - label: "end-to-end pulsar sink test"
     if: build.pull_request.labels includes "ci/run-e2e-pulsar-sink-tests" || build.env("CI_STEPS") =~ /(^|,)e2e-pulsar-sink-tests?(,|$$)/
     command: "ci/scripts/e2e-pulsar-sink-test.sh -p ci-dev"

e2e_test/iceberg/test_case/iceberg_engine.slt

@@ -0,0 +1,21 @@
+statement ok
+set sink_decouple = false;
+
+statement ok
+create table t(id int primary key, xxname varchar) engine = iceberg;
+
+statement ok
+insert into t values(1, 'xxx');
+
+statement ok
+FLUSH;
+
+sleep 5s
+
+query ??
+select * from t;
+----
+1 xxx
+
+statement ok
+DROP TABLE t;

grafana/risingwave-dev-dashboard.dashboard.py

@@ -4179,7 +4179,7 @@ def section_iceberg_metrics(outer_panels):
                             "read @ {{table_name}}",
                         ),
                         panels.target(
-                            f"sum({metric('nimtable_read_bytes')})",
+                            f"sum({metric('iceberg_read_bytes')})",
                             "total read",
                         ),
                     ],

java/connector-node/risingwave-sink-iceberg/src/main/java/com/risingwave/connector/catalog/JniCatalogWrapper.java

@@ -96,6 +96,17 @@ public boolean tableExists(String tableIdentifier) {
         return catalog.tableExists(id);
     }
 
+    /**
+     * Drop a table from the catalog.
+     *
+     * @param tableIdentifier The identifier of the table to drop.
+     * @return true if the table was dropped, false otherwise.
+     */
+    public boolean dropTable(String tableIdentifier) {
+        TableIdentifier id = TableIdentifier.parse(tableIdentifier);
+        return catalog.dropTable(id);
+    }
+
     /**
      * Create JniCatalogWrapper instance.
      *

proto/catalog.proto

@@ -369,6 +369,12 @@ message Table {
     int32 next_column_id = 2;
   }
 
+  enum Engine {
+    ENGINE_UNSPECIFIED = 0;
+    HUMMOCK = 1;
+    ICEBERG = 2;
+  }
+
   uint32 id = 1;
   uint32 schema_id = 2;
   uint32 database_id = 3;
@@ -474,6 +480,9 @@ message Table {
   // tables and tests. Not to be confused with the global catalog version for
   // notification service.
   TableVersion version = 100;
+
+  // Table Engine, currently only support hummock and iceberg
+  Engine engine = 200;
 }
 
 enum HandleConflictBehavior {

risedev.yml

@@ -307,6 +307,23 @@ profile:
       - use: compute-node
       - use: frontend
 
+  ci-iceberg-engine:
+    steps:
+      - use: minio
+      - use: postgres
+        port: 5432
+        address: db
+        database: metadata
+        user: postgres
+        password: postgres
+        user-managed: true
+        application: metastore
+      - use: meta-node
+        meta-backend: postgres
+      - use: compute-node
+      - use: frontend
+      - use: compactor
+
   meta-1cn-1fe-sqlite-with-recovery:
     config-path: src/config/ci-recovery.toml
     steps:

src/common/src/array/arrow/arrow_iceberg.rs

@@ -187,6 +187,14 @@ impl ToArrow for IcebergCreateTableArrowConvert {
         arrow_field
     }
 
+    fn jsonb_type_to_arrow(&self, name: &str) -> arrow_schema::Field {
+        let data_type = arrow_schema::DataType::Utf8;
+
+        let mut arrow_field = arrow_schema::Field::new(name, data_type, true);
+        self.add_field_id(&mut arrow_field);
+        arrow_field
+    }
+
     /// Convert RisingWave data type to Arrow data type.
     ///
     /// This function returns a `Field` instead of `DataType` because some may be converted to

src/common/src/catalog/mod.rs

@@ -27,6 +27,7 @@ use futures::stream::BoxStream;
 pub use internal_table::*;
 use parse_display::Display;
 pub use physical_table::*;
+use risingwave_pb::catalog::table::PbEngine;
 use risingwave_pb::catalog::{
     CreateType as PbCreateType, HandleConflictBehavior as PbHandleConflictBehavior,
     StreamJobStatus as PbStreamJobStatus,
@@ -541,6 +542,36 @@ impl ConflictBehavior {
     }
 }
 
+#[derive(Default, Debug, Clone, Copy, PartialEq, Eq, Hash)]
+pub enum Engine {
+    #[default]
+    Hummock,
+    Iceberg,
+}
+
+impl Engine {
+    pub fn from_protobuf(engine: &PbEngine) -> Self {
+        match engine {
+            PbEngine::Hummock | PbEngine::Unspecified => Engine::Hummock,
+            PbEngine::Iceberg => Engine::Iceberg,
+        }
+    }
+
+    pub fn to_protobuf(self) -> PbEngine {
+        match self {
+            Engine::Hummock => PbEngine::Hummock,
+            Engine::Iceberg => PbEngine::Iceberg,
+        }
+    }
+
+    pub fn debug_to_string(self) -> String {
+        match self {
+            Engine::Hummock => "Hummock".to_string(),
+            Engine::Iceberg => "Iceberg".to_string(),
+        }
+    }
+}
+
 #[derive(Clone, Copy, Debug, Default, Display, Hash, PartialOrd, PartialEq, Eq, Ord)]
 pub enum StreamJobStatus {
     #[default]

src/connector/Cargo.toml

@@ -43,6 +43,7 @@ chrono = { version = "0.4", default-features = false, features = [
     "clock",
     "std",
 ] }
+clap = { workspace = true }
 clickhouse = { git = "https://github.com/risingwavelabs/clickhouse.rs", rev = "d38c8b6391af098b724c114e5a4746aedab6ab8e", features = [
     "time",
 ] }

src/connector/src/connector_common/iceberg/jni_catalog.rs

@@ -285,7 +285,7 @@ impl CatalogV2 for JniCatalog {
         .map_err(|e| {
             iceberg::Error::new(
                 iceberg::ErrorKind::Unexpected,
-                "Failed to crete iceberg table.",
+                "Failed to create iceberg table.",
             )
             .with_source(e)
         })?
@@ -342,8 +342,29 @@ impl CatalogV2 for JniCatalog {
     }
 
     /// Drop a table from the catalog.
-    async fn drop_table(&self, _table: &TableIdent) -> iceberg::Result<()> {
-        todo!()
+    async fn drop_table(&self, table: &TableIdent) -> iceberg::Result<()> {
+        execute_with_jni_env(self.jvm, |env| {
+            let table_name_str = format!(
+                "{}.{}",
+                table.namespace().clone().inner().into_iter().join("."),
+                table.name()
+            );
+
+            let table_name_jstr = env.new_string(&table_name_str).unwrap();
+
+            call_method!(env, self.java_catalog.as_obj(), {boolean dropTable(String)},
+            &table_name_jstr)
+            .with_context(|| format!("Failed to drop iceberg table: {table_name_str}"))?;
+
+            Ok(())
+        })
+        .map_err(|e| {
+            iceberg::Error::new(
+                iceberg::ErrorKind::Unexpected,
+                "Failed to load iceberg table.",
+            )
+            .with_source(e)
+        })
     }
 
     /// Check if a table exists in the catalog.