ClustrixDB provides all of the information necessary for monitoring the cluster. The system database contains low-level information about all portions of the cluster, and proprietary SQL commands return high-level summary information. The following sections describe these features in detail.
As a best practice, unix sysadmin tools such as /usr/bin/top should not be used on ClustrixDB to monitor node status, as their results may not be reliable.
Listing Current Sessions
The following query will list all sessions in the cluster:
sql> select * from system.sessions\G *************************** 1. row *************************** nodeid: 1 session_id: 1560577 source_ip: NULL source_port: NULL local_ip: NULL local_port: NULL user: 4106 database: NULL trx_state: closed statement_state: finished xid: NULL cpu: 3 isolation: REPEATABLE-READ last_statement: delete from system.rebalancer_activity_log where finished < subdate(now(), interval @global.rebalancer_activity_log_hours * 60 minute) time_in_state_s: 161 created: 2016-01-11 20:31:41 heap_id: 288230379201751147 trx_age: NULL trx_mode: autocommit trx_counter_select: 0 trx_counter_insert: 0 trx_counter_update: 0 trx_counter_delete: 1 trx_is_writer: 1 1 row in set (0.01 sec)
Monitoring CPU Load
To display CPU load percentages (0 to 100) for CPU utilization, disk read/write utilization, and buffer cache miss rates, issue the SHOW LOAD command. In the cluster, CPU core 0 is used for specific tasks and therefore omitted from the load average when calculating the overall load.
Disk read and write values are calculated as a percentage using estimated maximums for solid state drives and might exceed 100% under certain conditions.
Monitoring Space Utilization per Table
The system.table_sizes table shows the size in bytes of each table in every database, including indexes, replicas, and row data. The value in the size column represents total storage on disk, rather than just the size of the data.
Monitoring Data Rebalancing Activity
ClustrixDB continuously monitors how data is distributed in the cluster. When it detects an imbalance, it rebalances data by moving, copying, and splitting as required. Data is also copied for protective purposes, such as when a disk or node is lost.
See ClustrixGUI Administration UI for more information on monitoring your cluster.