Skip to main content

Deploying Databend Cluster

Databend recommends deploying a cluster with a minimum of three meta nodes and one query node for production environments. To gain a better understanding of Databend cluster deployment, see Understanding Databend Deployments, which will familiarize you with the concept. This topic aims to provide a practical guide for deploying a Databend cluster.

Before You Begin​

Before you start, make sure you have completed the following preparations:

  • Plan your deployment. This topic is based on the following cluster deployment plan, which involves setting up a meta cluster comprising three meta nodes and a query cluster consisting of two query nodes:
Node #IP AddressLeader Meta Node?Tenant IDCluster ID
Meta-1172.16.125.128/24Yes--
Meta-2172.16.125.129/24No--
Meta-3172.16.125.130/24No--
Query-1172.16.125.131/24-defaultdefault
Query-2172.16.125.132/24-defaultdefault
  • Download and extract the latest Databend package to each node.
Example:
root@meta-1:/usr# mkdir databend && cd databend
root@meta-1:/usr/databend# curl -O https://repo.databend.rs/databend/v1.2.410/databend-v1.2.410-aarch64-unknown-linux-gnu.tar.gz
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  333M  100  333M    0     0  18.5M      0  0:00:18  0:00:18 --:--:-- 16.4M
root@meta-1:/usr/databend# tar -xzvf databend-v1.2.410-aarch64-unknown-linux-gnu.tar.gz

Step 1: Deploy Meta Nodes​

  1. Configure the file databend-meta.toml in each meta node:

    • Ensure that the id parameter in [raft_config] is set to a unique value.
    • Set the single parameter to true for the leader meta node.
    • For follower meta nodes, comment out the single parameter using the # symbol, then add a parameter named join and provide an array of the IP addresses of the other meta nodes as its value.
ParameterMeta-1Meta-2Meta-3
grpc_api_advertise_host172.16.125.128172.16.125.129172.16.125.130
id123
raft_listen_host172.16.125.128172.16.125.129172.16.125.130
raft_advertise_host172.16.125.128172.16.125.129172.16.125.130
singletrue//
join/["172.16.125.128:28103","172.16.125.130:28103"]["172.16.125.128:28103","172.16.125.129:28103"]
cd configs && nano databend-meta.toml
databend-meta.toml
log_dir                 = "/var/log/databend"
admin_api_address       = "0.0.0.0:28101"
grpc_api_address        = "0.0.0.0:9191"
# databend-query fetch this address to update its databend-meta endpoints list,
# in case databend-meta cluster changes.
grpc_api_advertise_host = "172.16.125.128"

[raft_config]
id            = 1
raft_dir      = "/var/lib/databend/raft"
raft_api_port = 28103

# Assign raft_{listen|advertise}_host in test config.
# This allows you to catch a bug in unit tests when something goes wrong in raft meta nodes communication.
raft_listen_host = "172.16.125.128"
raft_advertise_host = "172.16.125.128"

# Start up mode: single node cluster
single        = true
  1. To start the meta nodes, run the following script on each node: Start with the leader node (Meta-1) and then proceed with the follower nodes sequentially.
cd .. && cd bin
./databend-meta -c ../configs/databend-meta.toml > meta.log 2>&1 &
  1. Once all the meta nodes have started, you can check them using the following curl command:
curl 172.16.125.128:28101/v1/cluster/nodes
[{"name":"1","endpoint":{"addr":"172.16.125.128","port":28103},"grpc_api_advertise_address":"172.16.125.128:9191"},{"name":"2","endpoint":{"addr":"172.16.125.129","port":28103},"grpc_api_advertise_address":"172.16.125.129:9191"},{"name":"3","endpoint":{"addr":"172.16.125.130","port":28103},"grpc_api_advertise_address":"172.16.125.130:9191"}]

Step 2: Deploy Query Nodes​

  1. Configure the file databend-query.toml in each query node. The following list only includes the parameters you need to set in each query node to reflect the deployment plan outlined in this document.

    • Set the tenant ID and cluster ID according to the deployment plan.
    • Set the endpoints parameter to an array of the IP addresses of the meta nodes.
ParameterQuery-1 / Query-2
tenant_iddefault
cluster_iddefault
endpoints["172.16.125.128:9191","172.16.125.129:9191","172.16.125.130:9191"]
cd configs/
nano databend-query.toml
databend-query.toml
...

tenant_id = "default"
cluster_id = "default"

...

[meta]
# It is a list of `grpc_api_advertise_host:<grpc-api-port>` of databend-meta config
endpoints = ["172.16.125.128:9191","172.16.125.129:9191","172.16.125.130:9191"]
...

  1. For each query node, you also need to configure the object storage and admin users in the file databend-query.toml. For detailed instructions, see here.

  2. Run the following script on each query node to start them:

cd .. && cd bin
./databend-query -c ../configs/databend-query.toml > query.log 2>&1 &

Step 3: Verify Deployment​

Connect to one of the query nodes using BendSQL, and retrieve information about the existing query nodes:

bendsql -h 172.16.125.131
Welcome to BendSQL 0.16.0-homebrew.
Connecting to 172.16.125.131:8000 as user root.
Connected to Databend Query v1.2.410-4b8cd16f0c(rust-1.77.0-nightly-2024-04-08T12:21:53.785045868Z)

root@172.16.125.131:8000/default> SELECT * FROM system.clusters;

SELECT
  *
FROM
  system.clusters

┌──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│          name          │ cluster │      host      │  port  │                                 version                                 │
├────────────────────────┼─────────┼────────────────┼────────┼─────────────────────────────────────────────────────────────────────────┤
│ 7rwadq5otY2AlBDdT25QL4 │ default │ 172.16.125.132 │   9091 │ v1.2.410-4b8cd16f0c(rust-1.77.0-nightly-2024-04-08T12:21:53.785045868Z) │
│ cH331pYsoFmvMSZXKRrn2  │ default │ 172.16.125.131 │   9091 │ v1.2.410-4b8cd16f0c(rust-1.77.0-nightly-2024-04-08T12:21:53.785045868Z) │
└──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘
2 rows read in 0.031 sec. Processed 2 rows, 327 B (64.1 rows/s, 10.23 KiB/s)

Next Steps​

After deploying Databend, you might need to learn about the following topics:

Explore Databend Cloud for FREE
Low-cost
Fast Analytics
Easy Data Ingestion
Elastic Scaling
Try it today