Skip to content

Latest commit

 

History

History
446 lines (316 loc) · 24.5 KB

File metadata and controls

446 lines (316 loc) · 24.5 KB

Cluster Deployment

This guide describes how to manually deploy a cluster instance consisting of 3 ConfigNodes and 3 DataNodes (commonly referred to as a 3C3D cluster).

1. Prerequisites

  1. System Preparation: Ensure the system has been configured according to the System Requirements.

  2. IP Configuration: It is recommended to use hostnames for IP configuration to prevent issues caused by IP address changes. Set the hostname by editing the /etc/hosts file. For example, if the local IP is 192.168.1.3 and the hostname is iotdb-1, run:

echo "192.168.1.3  iotdb-1" >> /etc/hosts

Use the hostname for cn_internal_address and dn_internal_address in IoTDB configuration.

  1. Unmodifiable Parameters: Some parameters cannot be changed after the first startup. Refer to the Parameter Configuration section.

  2. Installation Path: Ensure the installation path contains no spaces or non-ASCII characters to prevent runtime issues.

  3. User Permissions: Choose one of the following permissions during installation and deployment:

    • Root User (Recommended): This avoids permission-related issues.
    • Non-Root User:
      • Use the same user for all operations, including starting, activating, and stopping services.
      • Avoid using sudo, which can cause permission conflicts.
  4. Monitoring Panel: Deploy a monitoring panel to track key performance metrics. Contact the Timecho team for access and refer to the Monitoring Board Install and Deploy.

2. Preparation

  1. Prepare the IoTDB database installation package::apache-iotdb-{version}-all-bin.zip(Please refer to the installation package for details:IoTDB-Package

  2. Configure the operating system environment according to environmental requirements (system environment configuration can be found in:Environment Requirement

3. Installation Steps

Taking a cluster with three Linux servers with the following information as example:

Node IP Hostname Services
11.101.17.224 iotdb-1 ConfigNode, DataNode
11.101.17.225 iotdb-2 ConfigNode, DataNode
11.101.17.226 iotdb-3 ConfigNode, DataNode

3.1 Configure Hostnames

On all three servers, configure the hostnames by editing the /etc/hosts file. Use the following commands:

echo "11.101.17.224  iotdb-1"  >> /etc/hosts
echo "11.101.17.225  iotdb-2"  >> /etc/hosts
echo "11.101.17.226  iotdb-3"  >> /etc/hosts 

3.2 Extract Installation Package

Unzip the installation package and navigate to the directory:

unzip  apache-iotdb-{version}-all-bin.zip 
cd  apache-iotdb-{version}-all-bin

3.3 Parameters Configuration

3.3.1 Memory Configuration

Edit the following files for memory allocation:

  • ConfigNode: ./conf/confignode-env.sh (or .bat for Windows)
  • DataNode: ./conf/datanode-env.sh (or .bat for Windows)
Parameter Description Default Recommended Notes
MEMORY_SIZE Total memory allocated to the node Empty As needed Effective after restarting the service.

3.3.2 General Configuration

Set the following parameters in ./conf/iotdb-system.properties. Refer to ./conf/iotdb-system.properties.template for a complete list.

Cluster-Level Parameters:

Parameter Description 11.101.17.224 11.101.17.225 11.101.17.226
cluster_name Name of the cluster defaultCluster defaultCluster defaultCluster
schema_replication_factor Metadata replication factor; DataNode count shall not be fewer than this value 3 3 3
data_replication_factor Data replication factor; DataNode count shall not be fewer than this value 2 2 2

ConfigNode Parameters:

Parameter Description Default Recommended 11.101.17.224 11.101.17.225 11.101.17.226 Notes
cn_internal_address Address used for internal communication within the cluster 127.0.0.1 Server's IPv4 address or hostname. Use hostname to avoid issues when the IP changes. iotdb-1 iotdb-2 iotdb-3 This parameter cannot be modified after the first startup.
cn_internal_port Port used for internal communication within the cluster 10710 10710 10710 10710 10710 This parameter cannot be modified after the first startup.
cn_consensus_port Port used for consensus protocol communication among ConfigNode replicas 10720 10720 10720 10720 10720 This parameter cannot be modified after the first startup.
cn_seed_config_node Address of the ConfigNode for registering and joining the cluster. (e.g.,cn_internal_address:cn_internal_port) 127.0.0.1:10710 Address and port of the seed ConfigNode (e.g., cn_internal_address:cn_internal_port) iotdb-1:10710 iotdb-1:10710 iotdb-1:10710 This parameter cannot be modified after the first startup.

DataNode Parameters:

Parameter Description Default Recommended 11.101.17.224 11.101.17.225 11.101.17.226 Notes
dn_rpc_address Address for the client RPC service 0.0.0.0 0.0.0.0 0.0.0.0 0.0.0.0 0.0.0.0 Effective after restarting the service.
dn_rpc_port Port for the client RPC service 6667 6667 6667 6667 6667 Effective after restarting the service.
dn_internal_address Address used for internal communication within the cluster 127.0.0.1 Server's IPv4 address or hostname. Use hostname to avoid issues when the IP changes. iotdb-1 iotdb-2 iotdb-3 This parameter cannot be modified after the first startup.
dn_internal_port Port used for internal communication within the cluster 10730 10730 10730 10730 10730 This parameter cannot be modified after the first startup.
dn_mpp_data_exchange_port Port used for receiving data streams 10740 10740 10740 10740 10740 This parameter cannot be modified after the first startup.
dn_data_region_consensus_port Port used for data replica consensus protocol communication 10750 10750 10750 10750 10750 This parameter cannot be modified after the first startup.
dn_schema_region_consensus_port Port used for metadata replica consensus protocol communication 10760 10760 10760 10760 10760 This parameter cannot be modified after the first startup.
dn_seed_config_node Address of the ConfigNode for registering and joining the cluster.(e.g.,cn_internal_address:cn_internal_port) 127.0.0.1:10710 Address of the first ConfigNode iotdb-1:10710 iotdb-1:10710 iotdb-1:10710 This parameter cannot be modified after the first startup.

Note: Ensure files are saved after editing. Tools like VSCode Remote do not save changes automatically.

3.4 Start ConfigNode Instances

  1. Start the first ConfigNode (iotdb-1) as the seed node
  cd sbin
  ./start-confignode.sh -d   # The "-d" flag starts the process in the background.
  1. Start the remaining ConfigNodes (iotdb-2 and iotdb-3) in sequence.

If the startup fails, refer to the Common Issues section below for troubleshooting.

3.5 Start DataNode Instances

On each server, navigate to the sbin directory and start the DataNode:

  cd sbin
  ./start-datanode.sh -d    # The "-d" flag starts the process in the background.

3.6 Start CLI

Enter the IoTDB CLI.

# For Linux or macOS
./start-cli.sh -sql_dialect table

# For Windows
./start-cli.bat -sql_dialect table

3.7 One-click Cluster Start and Stop

3.7.1 Overview

Within the root directory of IoTDB, the sbin subdirectory houses the start-all.sh and stop-all.sh scripts, which work in concert with the iotdb-cluster.properties configuration file located in the conf subdirectory. This synergy enables the one-click initiation or termination of all nodes within the cluster from a single node. This approach facilitates efficient management of the IoTDB cluster's lifecycle, streamlining the deployment and operational maintenance processes.

This following section will introduce the specific configuration items in the iotdb-cluster.properties file.

3.7.2 Configuration Items

Note:

  • When the cluster changes, this configuration file needs to be manually updated.
  • If the iotdb-cluster.properties configuration file is not set up and the start-all.sh or stop-all.sh scripts are executed, the scripts will, by default, start or stop the ConfigNode and DataNode nodes located in the IOTDB_HOME directory where the scripts reside.
  • It is recommended to configure SSH passwordless login: If not configured, the script will prompt for the server password after execution to facilitate subsequent start, stop, or destroy operations. If already configured, there is no need to enter the server password during script execution.
  • confignode_address_list
Name confignode_address_list
Description A list of IP addresses of the hosts where the ConfigNodes to be started/stopped are located. If there are multiple, they should be separated by commas.
Type String
Default None
Effective After restarting the system
  • datanode_address_list
Name datanode_address_list
Description A list of IP addresses of the hosts where the DataNodes to be started/stopped are located. If there are multiple, they should be separated by commas.
Type String
Default None
Effective After restarting the system
  • ssh_account
Name ssh_account
Description The username used to log in to the target hosts via SSH. All hosts must have the same username.
Type String
Default root
Effective After restarting the system
  • ssh_port
Name ssh_port
Description The SSH port exposed by the target hosts. All hosts must have the same SSH port.
Type int
Default 22
Effective After restarting the system
  • confignode_deploy_path
Name confignode_deploy_path
Description The path on the target hosts where all ConfigNodes to be started/stopped are located. All ConfigNodes must be in the same directory on their respective hosts.
Type String
Default None
Effective After restarting the system
  • datanode_deploy_path
Name datanode_deploy_path
Description The path on the target hosts where all DataNodes to be started/stopped are located. All DataNodes must be in the same directory on their respective hosts.
Type String
Default None
Effective After restarting the system

4. Maintenance

4.1 ConfigNode Maintenance

ConfigNode maintenance includes adding and removing ConfigNodes. Common use cases include:

  • Cluster Expansion: If the cluster contains only 1 ConfigNode, adding 2 more ConfigNodes enhances high availability, resulting in a total of 3 ConfigNodes.
  • Cluster Fault Recovery: If a ConfigNode's machine fails and it cannot function normally, remove the faulty ConfigNode and add a new one to the cluster.

Note: After completing ConfigNode maintenance, ensure that the cluster contains either 1 or 3 active ConfigNodes. Two ConfigNodes do not provide high availability, and more than three ConfigNodes can degrade performance.

4.1.1 Adding a ConfigNode

Linux / MacOS :

sbin/start-confignode.sh

Windows:

# Before version V2.0.4.x 
sbin\start-confignode.bat

# V2.0.4.x and later versions
sbin\windows\start-confignode.bat

4.1.2 Removing a ConfigNode

  1. Connect to the cluster using the CLI and confirm the internal address and port of the ConfigNode to be removed:
show confignodes;

Example output:

IoTDB> show confignodes
+------+-------+---------------+------------+--------+
|NodeID| Status|InternalAddress|InternalPort|    Role|
+------+-------+---------------+------------+--------+
|     0|Running|      127.0.0.1|       10710|  Leader|
|     1|Running|      127.0.0.1|       10711|Follower|
|     2|Running|      127.0.0.1|       10712|Follower|
+------+-------+---------------+------------+--------+
Total line number = 3
It costs 0.030s
  1. Remove the ConfigNode using the script:

Linux / MacOS:

sbin/remove-confignode.sh [confignode_id]
# Or:
sbin/remove-confignode.sh [cn_internal_address:cn_internal_port]

Windows:

# Before version V2.0.4.x 
sbin\remove-confignode.bat [confignode_id]
# Or:
sbin\remove-confignode.bat [cn_internal_address:cn_internal_port]

# V2.0.4.x and later versions
sbin\windows\remove-confignode.bat [confignode_id]
# Or:
sbin\windows\remove-confignode.bat [cn_internal_address:cn_internal_port]

4.2 DataNode Maintenance

DataNode maintenance includes adding and removing DataNodes. Common use cases include:

  • Cluster Expansion: Add new DataNodes to increase cluster capacity.
  • Cluster Fault Recovery: If a DataNode's machine fails and it cannot function normally, remove the faulty DataNode and add a new one to the cluster.

Note: During and after DataNode maintenance, ensure that the number of active DataNodes is not fewer than the data replication factor (usually 2) or the schema replication factor (usually 3).

4.2.1 Adding a DataNode

Linux / MacOS:

sbin/start-datanode.sh

Windows:

# Before version V2.0.4.x
sbin\start-datanode.bat

# V2.0.4.x and later versions
sbin\windows\start-datanode.bat

Note: After adding a DataNode, the cluster load will gradually balance across all nodes as new writes arrive and old data expires (if TTL is set).

4.2.2 Removing a DataNode

  1. Connect to the cluster using the CLI and confirm the RPC address and port of the DataNode to be removed:
show datanodes;

Example output:

IoTDB> show datanodes
+------+-------+----------+-------+-------------+---------------+
|NodeID| Status|RpcAddress|RpcPort|DataRegionNum|SchemaRegionNum|
+------+-------+----------+-------+-------------+---------------+
|     1|Running|   0.0.0.0|   6667|            0|              0|
|     2|Running|   0.0.0.0|   6668|            1|              1|
|     3|Running|   0.0.0.0|   6669|            1|              0|
+------+-------+----------+-------+-------------+---------------+
Total line number = 3
It costs 0.110s
  1. Remove the DataNode using the script:

Linux / MacOS:

sbin/remove-datanode.sh [dn_rpc_address:dn_rpc_port]

Windows:

# Before version V2.0.4.x
sbin\remove-datanode.bat [dn_rpc_address:dn_rpc_port]

# V2.0.4.x and later versions
sbin\windows\remove-datanode.bat [dn_rpc_address:dn_rpc_port]

5. Common Issues

  1. ConfigNode Fails to Start
  2. Review the startup logs to check if any parameters, which cannot be modified after the first startup, were changed.
  3. Check the logs for any other errors. If unresolved, contact technical support for assistance.
  4. If the deployment is fresh or data can be discarded, clean the environment and redeploy using the following steps:
**Clean the Environment**
  1. Stop all ConfigNode and DataNode processes: ```Bash sbin/stop-standalone.sh

    
    
  2. Check for any remaining processes: ```Bash jps

    or

    ps -ef | grep iotdb

    
    
  3. If processes remain, terminate them manually: ```Bash kill -9

    #For systems with a single IoTDB instance, you can clean up residual processes with: ps -ef | grep iotdb | grep -v grep | tr -s ' ' ' ' | cut -d ' ' -f2 | xargs kill -9

    
    
  4. Delete the data and logs directories: ```Bash cd /data/iotdb rm -rf data logs

    
    

6. Appendix

6.1 ConfigNode Parameters

Parameter Description Required
-d Starts the process in daemon mode (runs in the background). No

6.2 DataNode Parameters

Parameter Description Required
-v Displays version information. No
-f Runs the script in the foreground without backgrounding it. No
-d Starts the process in daemon mode (runs in the background). No
-p Specifies a file to store the process ID for process management. No
-c Specifies the path to the configuration folder; the script loads configuration files from this location. No
-g Prints detailed garbage collection (GC) information. No
-H Specifies the path for the Java heap dump file, used during JVM memory overflow. No
-E Specifies the file for JVM error logs. No
-D Defines system properties in the format key=value. No
-X Passes -XX options directly to the JVM. No
-h Displays the help instructions. No