clockpipe

• An alternative to clickpipe for on-premise clickhouse users.
• Based on CDC, data from the original source is written to clickhouse.

Supported Source

• PostgreSQL

Install

Build from source code

git clone https://github.com/myyrakle/clockpipe
cd clockpipe
cargo install --path .

Using Cargo

cargo install clockpipe

PostgreSQL Setup

• Synchronization is implemented through PostgreSQL Publication.
• modify postgresql.conf and restart postgresql server.

postgres=# SHOW config_file;
                   config_file
-------------------------------------------------
 /opt/homebrew/var/postgresql@14/postgresql.conf

enable logical replica

sudo vim /opt/homebrew/var/postgresql@14/postgresql.conf

wal_level=logical

max_slot_wal_keep_size=-1
max_wal_size=10240

sudo systemctl restart postgresql

How to Run

• Prepare config file (example)
• Enter the information about the PostgreSQL table you want to synchronize.

    "tables": [
        {
            "schema_name": "public",
            "table_name": "foo"
        },
        {
            "schema_name": "public",
            "table_name": "nc_usr_account"
        }
    ]

• Then, Run it

clockpipe run --config-file ./clockpipe-config.json

• Pipe automatically creates and synchronizes tables in Clickhouse by querying table information.

• If you don't want the initial synchronization, use the skip_copy option. (CDC-based synchronization still works.)

    "tables": [
        {
            "schema_name": "public",
            "table_name": "user_table",
            "skip_copy": true
        }
    ]