StarRocks Configuration

This guide explains how to configure a StarRocks sink in EZ-CDC.

Add StarRocks Sink

Via Portal

1. Navigate to Datasources → New Datasource
2. Select StarRocks as the type
3. Fill in the connection details:

Field	Description	Example
Name	Unique identifier	analytics-starrocks
Host	StarRocks FE hostname	starrocks.example.com
MySQL Port	Frontend MySQL protocol port	9030
HTTP Port	Backend HTTP port (Stream Load)	8040
Database	Target database	analytics
Username	StarRocks user	ezcdc_user
Password	User password	********

4. Click Test Connection
5. Click Save

Connection Options

Basic Options

Option	Required	Default	Description
host	Yes	-	FE hostname or IP
mysql_port	No	9030	FE MySQL protocol port
http_port	No	8040	BE HTTP port
database	Yes	-	Target database
username	Yes	-	StarRocks user
password	Yes	-	User password

Advanced Options

Option	Default	Description
load_url	Auto-generated	Custom Stream Load URL
timeout	300	Stream Load timeout (seconds)

Load URL Override

By default, EZ-CDC constructs the Stream Load URL as:

http://{host}:{http_port}/api/{database}/{table}/_stream_load

For custom configurations (e.g., load balancer):

{
  "load_url": "http://starrocks-lb.example.com:8040"
}

Cluster Configurations

Single Node

For development/testing:

{
  "host": "starrocks-standalone",
  "mysql_port": 9030,
  "http_port": 8040,
  "database": "analytics",
  "username": "root",
  "password": ""
}

Multi-Node Cluster

For production with multiple BEs, use a load balancer to distribute Stream Load requests across Backend nodes:

Architecture:

• Load Balancer (:8040) distributes requests to:
  • BE-1 (:8040)
  • BE-2 (:8040)
  • BE-3 (:8040)

Configure with load balancer:

{
  "host": "starrocks-fe",
  "mysql_port": 9030,
  "load_url": "http://starrocks-be-lb:8040",
  "database": "analytics",
  "username": "ezcdc_user",
  "password": "your_password"
}

StarRocks on Kubernetes

For K8s deployments:

{
  "host": "starrocks-fe.starrocks.svc.cluster.local",
  "mysql_port": 9030,
  "load_url": "http://starrocks-be.starrocks.svc.cluster.local:8040",
  "database": "analytics",
  "username": "ezcdc_user",
  "password": "your_password"
}

Test Connection

From Portal

Click Test Connection to verify:

1. MySQL connectivity (metadata queries)
2. HTTP connectivity (Stream Load)
3. Authentication
4. Database access

From CLI

# Test MySQL connection
mysql -h starrocks.example.com -P 9030 -u ezcdc_user -p

# Test Stream Load endpoint
curl -I http://starrocks.example.com:8040/api/analytics/test/_stream_load

Table Configuration

Automatic Table Creation

EZ-CDC can create tables automatically:

1. Reads source schema
2. Maps PostgreSQL types to StarRocks types
3. Creates Primary Key table
4. Adds audit columns

Example auto-created table:

CREATE TABLE orders (
  -- Source columns
  id BIGINT NOT NULL,
  customer_id BIGINT,
  total DECIMAL(18, 2),
  status VARCHAR(65535),
  created_at DATETIME,

  -- Audit columns (added by EZ-CDC)
  _cdc_updated_at DATETIME,
  _cdc_deleted BOOLEAN DEFAULT false
)
PRIMARY KEY (id)
DISTRIBUTED BY HASH(id) BUCKETS 8
PROPERTIES (
  "replication_num" = "1"
);

Manual Table Creation

For more control, create tables manually:

-- Create with specific settings
CREATE TABLE orders (
  id BIGINT NOT NULL,
  customer_id BIGINT,
  total DECIMAL(18, 2),
  status VARCHAR(100),  -- Specific length
  created_at DATETIME,
  _cdc_updated_at DATETIME,
  _cdc_deleted BOOLEAN DEFAULT false
)
PRIMARY KEY (id)
DISTRIBUTED BY HASH(id) BUCKETS 16  -- More buckets
PROPERTIES (
  "replication_num" = "3",  -- More replicas
  "enable_persistent_index" = "true"
);

Type Mapping

PostgreSQL → StarRocks

PostgreSQL Type	StarRocks Type	Notes
smallint	SMALLINT
integer	INT
bigint	BIGINT
numeric(p,s)	DECIMAL(p,s)	Max precision 38
real	FLOAT
double precision	DOUBLE
boolean	BOOLEAN
char(n)	CHAR(n)
varchar(n)	VARCHAR(n)	Max 65535
text	STRING
date	DATE
timestamp	DATETIME
timestamptz	DATETIME	Converted to UTC
jsonb	JSON	StarRocks 3.0+
uuid	VARCHAR(36)
bytea	VARBINARY

Troubleshooting

"Connection refused"

Error: Connection refused to starrocks.example.com:9030

Solutions:

1. Verify FE is running and listening
2. Check security groups allow port 9030
3. Try connecting with mysql client

"Access denied"

Error: Access denied for user 'ezcdc_user'

Solutions:

1. Verify password is correct
2. Check user has LOAD privilege
3. Verify user can access the database

"Stream Load failed"

Error: Failed to load data via Stream Load

Solutions:

1. Check BE is reachable on port 8040
2. Verify HTTP connectivity: curl http://be:8040
3. Check BE logs for errors
4. Verify user has LOAD privilege

"Database does not exist"

Error: Unknown database 'analytics'

Solutions:

-- Create the database
CREATE DATABASE analytics;
GRANT ALL ON analytics.* TO 'ezcdc_user';

Next Steps

• Stream Load Details
• Create a Job