Overview
Version
FineBI Version | Functional Change |
6.0 | / |
6.0.8 | Disabled data synchronization in ASYNC node. |
6.0.9 | Added information in the interface Node Management of Extract Cluster.
dbpath: Path to store extract data timestamp: System time
|
Functions
After the extract cluster is set up, you can manage it through the management interface.
Go to Manage > System > Extract Cluster. Once the extract cluster is configured and enabled, it can be seen.
This interface displays information of extract nodes, including node ID, node name, node IP, HTTP port, communication port between extract cluster nodes, node status, extract data storage path, and system time.
Node Status
Node Status | Description | Node Function | Available Operation |
UNKNOWN | The node has just started and is initializing the current node's status. | ||
ASYNC | Asynchronous node It supports high-concurrency queries. | Asynchronous nodes can only process query requests when there is data in the current node. (If there is no data, it will be forwarded to the synchronous node.) They can parse update tasks, but cannot execute update tasks. |
Query the status of all table data for the current node. |
SYNC | Synchronous node The status of data files in the synchronous nodes are always the latest (ensure data high consistency). During the update, it will be successful only when the data synchronization in the node is completed. | Synchronous nodes can handle all requests. |
Query the status of all table data for the current node. |
RECOVER | Restore node When a synchronous node has abnormal situations such as heartbeat timeout, which makes some data not be the latest, the synchronous node will be recovery status. The non-latest data will be synchronized. | Recovery nodes can process query requests when there is data in the current node. (If there is no data, it will be forwarded to the synchronous node.) They can parse update tasks, but cannot execute update tasks. |
Query the status of all table data for the current node.
Query the current node and the list of tables that is synchronizing (to be synchronized and currently synchronizing).
Force the current node to be set to SYNC status. Note: Other SYNC nodes may become RECOVER as a result. Note: When the synchronous node is started, it may first become RECOVER. After the final completion of judging the node status, it will become SYNC again. |
DOWN | Dead node This node is unavailable and there is a table in this node that no other nodes have. Consequently, other nodes are able to query the table of DOWN nodes from other configuration libraries, but unable to download it. It will always be RECOVER. | / | Delete Deleting DOWN nodes can remove the deleted nodes from all table states, thereby removing the tables that only DOWN nodes have from the internode table synchronization task, which allows other synchronous nodes to change from RECOVER to SYNC status. |
Debugging Tool
Description of the debugging tools in the management interface:
Function | Description |
Update status slice | Print the current scheduling status of the cluster to locate the problem that a stuck update in the extract cluster environment. |
All node table status | Print the table data status of all nodes, mainly used to determine which node has the latest data. |
Node status change record | Print the records of a synchronous node becoming RECOVER and its recovery for locating abnormal node status. |
Local stateless table | Print all tables in the node configuration library that do not have metadata to locate the problem of excess disk space usage. |
Using of Redis locks | Print the currently used Redis lock to facilitate locating cluster locks and related issues. |
