Synchronize from Kafka

Log into your UCloud Global account and go to the User Console , then search for “Kafka Message Queue UKafka” under all products or under Big Data, and enter the Kafka Message Queue UKafka Console . Click the Create Cluster button. On the create cluster page, make selections based on the configurations provided and place your order.

Friendly reminder:

UKafka is a paid product. Please refer to the pricing on the create page.

Ensure that the UKafka cluster and the UClickhouse cloud data warehouse cluster are in the same region (same VPC). If they are different, use Private Network VPC -> Network Interconnection.

Prerequisites

The UKafka product has been created, and a Topic and Group have been created. Please refer to UKafka Creation.

Steps

Connect to the Cluster

Set up a cloud host in the same region as the cluster (same subnet), and install the Clickhouse-client on the cloud host. Older version download: Download Clickhouse-client . Newer version download: Download Clickhouse-client .

It is recommended to choose the Clickhouse-client version corresponding to the kernel version actually created. For example, if the cluster kernel version created is 21.8.14.5, download the following rpm packages:
```
wget https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/clickhouse-client-21.8.14.5-2.noarch.rpm
wget https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/clickhouse-common-static-21.8.14.5-2.x86_64.rpm
```

Execute the installation


rpm -ivh clickhouse-common-static-21.8.14.5-2.x86_64.rpm
rpm -ivh clickhouse-client-21.8.14.5-2.noarch.rpm

Connect to the cluster with clickhouse-client
```
clickhouse-client --host=any node IP address --port=9000 --user=admin --password=password set during cluster creation
```
The above command will enter interactive mode. The default username is admin, the default port is 9000, and the node IP can be checked in the cluster details.

Create Database


CREATE DATABASE IF NOT EXISTS ck_test ON CLUSTER ck_cluster;

Create Kafka Consumer Table


CREATE TABLE ck_test.lineorder_kafka ON CLUSTER ck_cluster 
(
    LO_ORDERKEY             UInt32,
    LO_LINENUMBER           UInt8,
    LO_CUSTKEY              UInt32,
    LO_PARTKEY              UInt32,
    LO_SUPPKEY              UInt32,
    LO_ORDERDATE            Date,
    LO_ORDERPRIORITY        LowCardinality(String),
    LO_SHIPPRIORITY         UInt8,
    LO_QUANTITY             UInt8,
    LO_EXTENDEDPRICE        UInt32,
    LO_ORDTOTALPRICE        UInt32,
    LO_DISCOUNT             UInt8,
    LO_REVENUE              UInt32,
    LO_SUPPLYCOST           UInt32,
    LO_TAX                  UInt8,
    LO_COMMITDATE           Date,
    LO_SHIPMODE             LowCardinality(String)
) ENGINE = Kafka()
SETTINGS
    kafka_broker_list = 'broker list',
    kafka_topic_list = 'topic list',
    kafka_group_name = 'consumer group name',
    kafka_format = 'CSV',
    kafka_num_consumers = 1,
    kafka_max_block_size = 65536,
    kafka_skip_broken_messages = 0,
    kafka_auto_offset_reset = 'latest';


Note:
  The Kafka consumer table cannot be used directly as a result table. Its function is only to consume data from Kafka; it does not actually store the data.

Parameter description:

Name	Required	Description
kafka_broker_list	Yes	A comma-separated list of Kafka broker addresses.
kafka_topic_list	Yes	A comma-separated list of Topic names.
kafka_group_name	Yes	The consumer group name for Kafka.
kafka_format	Yes	The message format supported by UClickHouse. Supported formats can be found at Input and Output Formats .
kafka_row_delimiter	No	Row delimiter for separating different data rows. Default is “\n”, but you can set it according to the actual separation format used when writing data.
kafka_num_consumers	No	Number of consumers for a single table. Default value is 1. If one consumer’s throughput is not sufficient, specify more consumers. The total number of consumers should not exceed the number of partitions in the Topic, because each partition can only be assigned one consumer.
kafka_max_block_size	No	Maximum batch size of Kafka messages in bytes. Default value is 65536 bytes.
kafka_skip_broken_messages	No	The tolerance level of kafka message parser for dirty data. Default value is 0. If `kafka_skip_broken_messages=N`, the engine will skip N unparseable Kafka messages (one message equals one line of data).
kafka_commit_every_batch	No	Frequency of executing Kafka commit, described below. Commit is executed after writing all the data of a whole Block data block. A commit is executed after writing each Batch of data.
kafka_ auto_offset_reset	No	Where to start reading Kafka data from. Options are described below. earliest: Start reading Kafka data from the earliest offset. latest: Start reading Kafka data from the latest offset. Note: The 21.8 version of ClickHouse cloud database clusters does not support this parameter.

For more parameters, refer to the official Kafka documentation .

Create UClickhouse Local Table

Create Local Table


CREATE TABLE ck_test.lineorder_local ON CLUSTER ck_cluster
(
    LO_ORDERKEY             UInt32,
    LO_LINENUMBER           UInt8,
    LO_CUSTKEY              UInt32,
    LO_PARTKEY              UInt32,
    LO_SUPPKEY              UInt32,
    LO_ORDERDATE            Date,
    LO_ORDERPRIORITY        LowCardinality(String),
    LO_SHIPPRIORITY         UInt8,
    LO_QUANTITY             UInt8,
    LO_EXTENDEDPRICE        UInt32,
    LO_ORDTOTALPRICE        UInt32,
    LO_DISCOUNT             UInt8,
    LO_REVENUE              UInt32,
    LO_SUPPLYCOST           UInt32,
    LO_TAX                  UInt8,
    LO_COMMITDATE           Date,
    LO_SHIPMODE             LowCardinality(String)
)
ENGINE = MergeTree ORDER BY (LO_ORDERKEY);

Create Distributed Table

If you only need to sync data to the local table, you can skip this step.


CREATE TABLE IF NOT EXISTS ck_test.lineorder on cluster ck_cluster AS ck_test.lineorder_local  ENGINE = Distributed(
       ck_cluster,
       ck_test,
       lineorder_local,
       rand()
);

Create Materialized View

Create a materialized view to synchronize data consumed from the Kafka consumer table to the UClickHouse target table. If your target table is a local table, replace the distributed table name with the local table name before synchronization.


CREATE MATERIALIZED VIEW consumer ON CLUSTER ck_cluster TO lineorder AS SELECT * FROM lineorder_kafka;

Write Messages in UKafka

Please refer to UKafka.

Query Verification


clickhouse-client --host=any node IP address --port=9000 --user=admin --password=password set during cluster creation --query="select count(*) from lineorder"