Cluster Management

tip

Only Enterprise Edition supports

View node list

SHOW DATANODES

The SHOW DATANODES command will return a list of all nodes in the cluster as follows:

+---------+-----------+-----------+---------+-----------+------------+---------------------+
| NODE_ID | HOST      | ATTRIBUTE | STATUS  | DISK_FREE | DISK_TOTAL | LAST_UPDATED_TIME   |
+---------+-----------+-----------+---------+-----------+------------+---------------------+
| 1001    | localhost | HOT       | HEALTHY | 26.03 GB  | 62.67 GB   | 2023-12-26 09:15:29 |
+---------+-----------+-----------+---------+-----------+------------+---------------------+

Add node

To add a new node, simply set the meta_service_addr in the cluster configuration to point to the meta endpoint in the target cluster, and ensure that the name matches the name of the target cluster.Each node_id in the cluster must be kept unique, and the host must ensure that any node in the cluster can be accessed.

[global]
node_id = 888888
host = "localhost"
... ...
[cluster]
name = cluster_xxx
meta_service_addr = [127.0.0.1:8901]
... ...

Remove node

If the node is a storage node (tskv or query_tskv), its data will be reasonably migrated to other nodes that can store data. The operation of taking offline a storage node requires two or more storage nodes with the same ATTRIBUTE.

REMOVENODE nodeid [FORCE]

Example

Remove the node with NODE_ID as 1001

If there is data in node 1001, it will be marked as read-only (ReadOnly) and will no longer write new data until the data in the node disappears completely according to the TTL setting or the data disappears based on the tiered storage function. The cluster will completely destroy the 1001 node, and the removed node cannot rejoin the cluster.

REMOVENODE 1001

Force remove node

REMOVENODE 1001 FORCE

When the FORCE parameter is attached, the data will be actively migrated to other nodes, and the following rules will be followed during migration:

• Do not migrate to other nodes that are not part of the replica set.
• Choose the node with the largest remaining storage space that can be migrated.
• If there is no appropriate node migration, the data is discarded.

During migration the node will be marked as Migrationn in the state, stop writing new data, after completion the node will be completely destroyed, the node will not be able to rejoin the cluster.

View Vnode list

SHOW VNODE

Display a list of all Vnodes in the current database

+----------+---------+-----------+--------------------+---------------+-------------+---------------------+---------------------+---------+
| VNODE_ID | NODE_ID | BUCKET_ID | REPLICATION_SET_ID | DATABASE_NAME | TENANT_NAME | START_TIME          | END_TIME            | STATUS  |
+----------+---------+-----------+--------------------+---------------+-------------+---------------------+---------------------+---------+
| 3        | 1001    | 1         | 2                  | usage_schema  | cnosdb      | 1671408000000000000 | 1702944000000000000 | Running |
+----------+---------+-----------+--------------------+---------------+-------------+---------------------+---------------------+---------+

CnosDB supports using SQL to perform operations such as MOVE, COPY, DROP, COMPACT, CHECKSUM GROUP on Vnode

Example

Move VNODE with VNODE_ID of 5 to node 1001

MOVE VNODE 5 TO NODE 1001

Copy the VNODE with VNODE_ID of 5 to node 1001

COPY VNODE 5 TO NODE 1001

Remove the VNODE with 5

DROP VNODE 5

Force compress VNODE with VNODE_ID of 5, 6, 7 respectively

COMPACT VNODE 5 6 7

Check the data consistency of REPLICATION_SET_ID for 5 Group

CHECKSUM GROUP 5

Resource Management

CnosDB supports resource management, will monitor some complex and multi-step resources, and will retry when the operation fails until successful.

The currently supported resources include: DropTenant, DropDatabase, DropTable, DropColumn, AddColumn, AlterColumn, RenameTagName, UpdateTagValue

The corresponding commands are: DROP TENANT, DROP DATABASE, DROP TABLE, DROP COLUMN, ADD COLUMN, ALTER COLUMN, REMOTE COLUMN

The system table information_schema.resource_status stores the status of resources:

SELECT * FROM information_schema.resource_status;
+---------------------+------------+--------------+-----------+----------+---------+
| time                | name       | action       | try_count | status   | comment |
+---------------------+------------+--------------+-----------+----------+---------+
| 2023-10-18 08:15:16 | cnosdb/db1 | DropDatabase | 0         | Schedule |         |
+---------------------+------------+--------------+-----------+----------+---------+

note

The community edition only supports some features:

Command	Enterprise	Community
SHOW DATANODES	✅	❌
REMOVENODE	✅	❌
SHOW VNODE	✅	❌
MOVE VNODE	✅	✅
COPY VNODE	✅	✅
DROP VNODE	✅	✅
COMPACT VNODE	✅	✅
CHECKSUM GROUP	✅	✅
Resource Management	✅	✅