event status show
Display event status
Availability: This command is available to cluster administrators at the admin privilege level.
Description
The event status show
command summarizes information about occurrences of events. For detailed information about specific occurrences of events, use the event log show command.
Parameters
- {
[-fields <fieldname>,…]
-
If you specify the
-fields <fieldname>, …
parameter, the command output also includes the specified field or fields. You can use '-fields ?' to display the fields to specify. - |
[-instance ]
} -
If you specify the
-instance
parameter, the command displays detailed information about all fields. [-node {<nodename>|local}]
- Node-
Selects the event records that match this parameter value. Events are tracked on a node-by-node basis, rather than being rolled up cluster-wide.
[-message-name <Message Name>]
- Message Name-
Selects the event records that match this parameter value. The message name is a short descriptive string. Filtering output by message name displays messages of a specific type.
[-indications <integer>]
- Number of Indications-
Selects the event records that match this parameter value. This parameter is most useful when used with a range, such as using the range ">20" to display only events that have been posted more than 20 times.
[-drops <integer>]
- Number of Drops-
Selects the event records that match this parameter value.
[-last-time-occurred <MM/DD/YYYY HH:MM:SS>]
- Last Indication Time-
Selects the event records that match this parameter value.
[-last-time-dropped <MM/DD/YYYY HH:MM:SS>]
- Last Suppressed Indication Time-
Selects the event records that match this parameter value.
[-last-time-processed <MM/DD/YYYY HH:MM:SS>]
- Last Processed Indication Time-
Selects the event records that match this parameter value.
[-stat-starting-time <MM/DD/YYYY HH:MM:SS>]
- Stat Starting Time-
Selects the event records that match this parameter value.
[-last-hour-histogram <integer>,…]
- 60-minute Histogram (privilege: advanced)-
Use this parameter with the
-fields
parameter to display the "last hour" histogram for each event type. The last hour histogram records the number of times each event occurred in the last hour. The histogram is divided into sixty buckets, and each bucket collects one minute's events. The buckets display with the most recent event first. [-last-day-histogram <integer>,…]
- 24-hour Histogram (privilege: advanced)-
Use this parameter with the
-fields
parameter to display the "last day" histogram for each event type. The last day histogram records the number of times each event occurred in the last day. The histogram is divided into 24 buckets, and each bucket collects one hour's events. The buckets display with the most recent event first. [-last-week-histogram <integer>,…]
- 7-day Histogram (privilege: advanced)-
Use this parameter with the
-fields
parameter to display the "last week" histogram for each event type. The last week histogram records the number of times each event occurred in the last week. The histogram is divided into 7 buckets, and each bucket collects one day's events. The buckets display with the most recent event first. [-severity {NODE_FAULT|SVC_FAULT|NODE_ERROR|SVC_ERROR|WARNING|NOTICE|INFO|DEBUG|VAR}]
- Severity-
Selects events that have the event severity you specify. Severity levels sort with the most severe levels first. Severity levels:
-
NODE_FAULT - The node has detected data corruption, or is unable to provide client service.
-
SVC_FAULT - The node has detected a temporary loss of service. Typically, this is caused by a transient software fault.
-
NODE_ERROR - The node has detected a hardware error that is not immediately fatal.
-
SVC_ERROR - The node has detected a software error that is not immediately fatal.
-
WARNING - A high-priority message that does not indicate a fault.
-
NOTICE - A normal-priority message that does not indicate a fault.
-
INFO - A low-priority message that does not indicate a fault.
-
DEBUG - A debugging message. These messages are typically suppressed.
-
VAR - These messages have variable severity. Severity level for these messages is selected at runtime.
The examples below illustrate how to query on severity.
-
Examples
The following example displays recent event-occurrence status for node1:
cluster1::> event status show -node node1 Node Message Occurs Drops Last Time ----------------- ---------------------------- ------ ----- ------------------- node1 raid.spares.media_scrub.start 6 0 3/11/2010 15:59:00 node1 raid.uninitialized.parity.vol 3 0 3/11/2010 15:58:28 node1 raid.vol.state.online 3 0 3/11/2010 15:58:29 node1 reg.defaultCommit.set.timeTaken 1 0 3/11/2010 15:58:28 node1 scsitgt.ha.state.changed 2 0 3/11/2010 15:58:28 node1 ses.multipath.notSupported 2 0 3/11/2010 15:58:43 node1 shelf.config.mpha 1 0 3/11/2010 15:58:48 node1 sk.hog.runtime 1 0 3/11/2010 15:58:28 node1 snmp.agent.msg.access.denied 1 0 3/11/2010 15:58:28 node1 snmp.link.up 6 0 3/11/2010 15:58:28 node1 tar.csum.mismatch 2 0 3/11/2010 15:58:28 node1 tar.extract.success 2 0 3/11/2010 15:58:28 node1 vifmgr.lifsuccessfullymoved 3 0 3/11/2010 15:58:46 node1 vifmgr.portdown 1 0 3/11/2010 15:58:48 node1 vifmgr.portup 5 0 3/11/2010 15:58:48 node1 vifmgr.startedsuccessfully 1 0 3/11/2010 15:58:43
The following example displays a summary of events which are warnings or more severe:
cluster1::> event status show -node node1 -severity <=warning -fields indications,drops,severity node message-name indications drops severity ------- ------------------------ ----------- ----- -------- node1 api.output.invalidSchema 5463 840 WARNING node1 callhome.dsk.config 1 0 WARNING node1 callhome.sys.config 1 0 SVC_ERROR node1 cecc_log.dropped 145 0 WARNING node1 cecc_log.entry 5 0 WARNING node1 cecc_log.entry_no_syslog 4540 218 WARNING node1 cecc_log.summary 5 0 WARNING node1 cf.fm.noPartnerVariable 5469 839 WARNING node1 cf.fm.notkoverBadMbox 1 0 WARNING node1 cf.fm.notkoverClusterDisable 1 0 WARNING node1 cf.fsm.backupMailboxError 1 0 WARNING node1 cf.takeover.disabled 23 0 WARNING node1 cmds.sysconf.logErr 1 0 NODE_ERROR node1 config.noPartnerDisks 1 0 NODE_ERROR node1 fci.initialization.failed 2 0 NODE_ERROR node1 fcp.service.adapter 1 0 WARNING node1 fmmb.BlobNotFound 1 0 WARNING node1 ha.takeoverImpNotDef 1 0 WARNING node1 httpd.config.mime.missing 2 0 WARNING node1 mgr.opsmgr.autoreg.norec 1 0 WARNING node1 monitor.globalStatus.critical 1 0 NODE_ERROR node1 raid.mirror.vote.versionZero 1 0 SVC_ERROR node1 ses.multipath.notSupported 2 0 NODE_ERROR node1 snmp.agent.msg.access.denied 1 0 WARNING 24 entries were displayed.
The above example makes use of several features which are common to all show
commands:
-
A query is specified for the severity parameter. A query restricts the output of the show command; only rows matching the query will be displayed. In this case, the query indicates that only events which have a severity of "WARNING" or more severe will be displayed.
-
The fields parameter selects the fields to display. Note that the severity field is not displayed in the default output.