Shulou(Shulou.com)06/01 Report--

Monitoring status description

Cluster integrity check:

Wsrep_cluster_state_uuid: the values of all nodes in the cluster should be the same. Nodes with different values indicate that they are not connected to the cluster.

Wsrep_cluster_conf_id: normally this value is the same on all nodes. If the value is different, the node is temporarily partitioned. The same value should be restored when the network connection between nodes is restored.

Wsrep_cluster_size: if this value matches the expected number of nodes, then all cluster nodes are connected.

Wsrep_cluster_status: the state of the cluster composition. If it is not Primary, it means that there is a partition or split-brain condition.

Node status check:

Wsrep_ready: if the value is ON, the SQL load is acceptable. If Off, you need to check wsrep_connected.

Wsrep_connected: if the value is Off and the value of wsrep_ready is also Off, the node is not connected to the cluster. It may be caused by misconfiguration such as wsrep_cluster_address or wsrep_cluster_name. You need to check the error log for specific errors)

Wsrep_local_state_comment: if wsrep_connected is On, but wsrep_ready is OFF, you can check the reason from this item.

Copy the health check:

Wsrep_flow_control_paused: indicates how long replication has been stopped. That is to say, the extent to which the cluster is slow due to Slave latency. The value is 0-1, which is as close to 0 as possible, and a value of 1 means that replication stops completely. The value of wsrep_slave_threads can be optimized to improve.

Wsrep_cert_deps_distance: how many transactions can be processed in parallel. The wsrep _ slave_threads setting should not be too much higher than this value.

Wsrep_flow_control_sent: indicates how many times the node has stopped copying.

Wsrep_local_recv_queue_avg: indicates the average length of the slave transaction queue. A sign of bottleneck.

The slowest node has the highest wsrep_flow_control_sent and wsrep_local_recv_queue_avg values. If these two values are lower, they are relatively better.

Detect slow network problems:

Wsrep_local_send_queue_avg: a harbinger of network bottleneck. If this value is high, there may be a cyber bottle.

Number of conflicts or deadlocks:

Wsrep_last_committed: number of transactions last committed

Wsrep_local_cert_failures and wsrep_local_bf_aborts: rollback, number of conflicts detected

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.