Sign InTry Free

Deploy and Maintain TiCDC

This document describes how to deploy and maintain a TiCDC cluster, including the hardware and software recommendations. You can either deploy TiCDC along with a new TiDB cluster or add the TiCDC component to an existing TiDB cluster.

Software and hardware recommendations

In production environments, the recommendations of software and hardware for TiCDC are as follows:

Linux OSVersion
Red Hat Enterprise Linux7.3 or later versions
CentOS7.3 or later versions
CPUMemoryDisk typeNetworkNumber of TiCDC cluster instances (minimum requirements for production environment)
16 core+64 GB+SSD10 Gigabit network card (2 preferred)2

For more information, see Software and Hardware Recommendations.

Deploy a new TiDB cluster that includes TiCDC using TiUP

When you deploy a new TiDB cluster using TiUP, you can also deploy TiCDC at the same time. You only need to add the cdc_servers section in the configuration file that TiUP uses to start the TiDB cluster. The following is an example:

cdc_servers: - host: gc-ttl: 86400 data_dir: "/cdc-data" - host: gc-ttl: 86400 data_dir: "/cdc-data"

More references:

Add or scale out TiCDC to an existing TiDB cluster using TiUP

The method of scaling out a TiCDC cluster is similar to that of deploying one. It is recommended to use TiUP to perform the scale-out.

  1. Create a scale-out.yml file to add the TiCDC node information. The following is an example:

    cdc_servers: - host: gc-ttl: 86400 data_dir: /tidb-data/cdc-8300 - host: gc-ttl: 86400 data_dir: /tidb-data/cdc-8300 - host: gc-ttl: 86400 data_dir: /tidb-data/cdc-8300
  2. Run the scale-out command on the TiUP control machine:

    tiup cluster scale-out <cluster-name> scale-out.yml

For more use cases, see Scale out a TiCDC cluster.

Delete or scale in TiCDC from an existing TiDB cluster using TiUP

It is recommended that you use TiUP to scale in TiCDC nodes. The following is the scale-in command:

tiup cluster scale-in <cluster-name> --node

For more use cases, see Scale in a TiCDC cluster.

Upgrade TiCDC using TiUP

You can upgrade TiDB clusters using TiUP, during which TiCDC is upgraded as well. After you execute the upgrade command, TiUP automatically upgrades the TiCDC component. The following is an example:

tiup update --self && \ tiup update --all && \ tiup cluster upgrade <cluster-name> <version> --transfer-timeout 600

Upgrade cautions

When you upgrade a TiCDC cluster, you need to pay attention to the following:

  • TiCDC v4.0.2 reconfigured changefeed. For details, see Configuration file compatibility notes.

  • If you encounter any problem during the upgrade, you can refer to upgrade FAQs for solutions.

  • Since v6.3.0, TiCDC supports rolling upgrade. During the upgrade, the replication latency is stable and does not fluctuate significantly. Rolling upgrade takes effect automatically if the following conditions are met:

  • TiCDC is v6.3.0 or later.

    • TiUP is v1.11.3 or later.
    • At least two TiCDC instances are running in the cluster.

Modify TiCDC cluster configurations using TiUP

This section describes how to use the tiup cluster edit-config command to modify the configurations of TiCDC. In the following example, it is assumed that you need to change the default value of gc-ttl from 86400 to 172800 (48 hours).

  1. Run the tiup cluster edit-config command. Replace <cluster-name> with the actual cluster name:

    tiup cluster edit-config <cluster-name>
  2. In the vi editor, modify the cdc server-configs:

    server_configs: tidb: {} tikv: {} pd: {} tiflash: {} tiflash-learner: {} pump: {} drainer: {} cdc: gc-ttl: 172800

    In the preceding command, gc-ttl is set to 48 hours.

  3. Run the tiup cluster reload -R cdc command to reload the configuration.

Stop and start TiCDC using TiUP

You can use TiUP to easily stop and start TiCDC nodes. The commands are as follows:

  • Stop TiCDC: tiup cluster stop -R cdc
  • Start TiCDC: tiup cluster start -R cdc
  • Restart TiCDC: tiup cluster restart -R cdc

Enable TLS for TiCDC

See Enable TLS Between TiDB Components.

View TiCDC status using the command-line tool

Run the following command to view the TiCDC cluster status. Note that you need to replace v<CLUSTER_VERSION> with the TiCDC cluster version, such as v6.5.0:

tiup ctl:v<CLUSTER_VERSION> cdc capture list --server=
[ { "id": "806e3a1b-0e31-477f-9dd6-f3f2c570abdd", "is-owner": true, "address": "", "cluster-id": "default" }, { "id": "ea2a4203-56fe-43a6-b442-7b295f458ebc", "is-owner": false, "address": "", "cluster-id": "default" } ]
  • id: Indicates the ID of the service process.
  • is-owner: Indicates whether the service process is the owner node.
  • address: Indicates the address via which the service process provides interface to the outside.
  • cluster-id: Indicates the ID of the TiCDC cluster. The default value is default.
Download PDF
One-stop & interactive experience of TiDB's capabilities WITHOUT registration.
TiDB Dedicated
TiDB Serverless
Get Demo
Get Started
© 2023 PingCAP. All Rights Reserved.