Cluster Diagnostics

To diagnose cluster-related issues, you can display the cluster membership history. The show cluster history command displays a history of the most recent 200 cluster-related events for a node. The cluster membership events include joins, leaves, membership updates, and initial configuration synchronizing.

The cluster membership history includes the following information (also found in the system log file):

Table 1: Cluster Membership History Information

Name	Format
[Index]	The entry index. The [0] entry is always displayed and provides information about when the log was started. The circular log maintains up to another 199 entries. For example: [140]
Timestamp	The date and time of the event, including the millisecond. The format is YYYY/MM/DD for the date, HH:MM:SS for the time, and .xxx for the millisecond. For example: 2015/02/06 10:47:19.918
Log Event	The short description of the event. For example: cfg sync
Cluster Node ID	The cluster node ID. Zero (0) is valid until a node is assigned a node ID. For example: 3
Current role	The name of current role. The valid roles are: unknown, leader, standby, and normal. In a stable cluster, there will be one leader, one standby, and the remaining nodes will be normal. For example: leader
Action	The detailed description of the event. For example: Cfg sync resp msg received (src nodeid 6, dst nodeid 3)

Use the following CLI command on a node in the cluster to display the cluster membership history for that node:

(config) # show cluster history

Use the following CLI command on the leader to display the cluster membership history for a specific node in the cluster:

(config) # show cluster history box-id 2

If you use the following command when you are not on the leader, an error message is displayed:

(config) # show cluster history box-id 2

Not leader - can only display cluster log for local box (1).