Optimize Distributed Workload - Expert Oracle RAC Performance Diagnostics and Tuning - page 494

Table 15-1. ( continued )

Event Identifier Description

INSTANCE Name of the RAC instance for which the event is being raised.

HOST Name of the cluster node from where such an event was raised; the value matches the node

name known to Cluster Synchronization services (CSS).

STATUS The current status of the event type indicates what has occurred for the event type. The valid

status values are the following:
UP Managed resource is now up and available.

DOWN Managed resource is now down and is currently not available for access.

PRECONN_UP

The pre-connect application services is now up and available.

PRECONN_DOWN The pre-connect application service has failed or is down and is not

currently available.

NODEDOWN The Oracle RAC cluster node indicated by the host identifier is down and

is not reachable.

NOT_RESTARTING

Indicates that one of the managed resources that failed will not restart
either on the failed node or on another node after failover. For example,
VIP should failover to another and restart when a node fails.

UNKNOWN Status is unknown and no description is available.

REASON Indicates the reason for the event being raised and is normally related to the status. The

following are the possible reasons:
USER

Indicates that the down or up event raised is use initiated. Operations
performed using srvctl or from sqlplus belong to this category. This is a
planned type of event.

FAILURE During constant polling of the health of the various resources, when

a resource is not reachable, a failure event is triggered. This is an
unplanned type of event.

DEPENDENCY Availability of certain resources depends on other resources in the cluster

being up. For example, the Oracle instance on a specific node depends on
the database application being available. This is also an unplanned event.

UNKNOWN

State of the application was not known during the failure to determine the
actual cause of failure. This is also an unplanned event.

SYSTEM Change in system state, for example, when a node is started. This reason

is normally associated with only system events. For other event types, the
reason would be “boot.”

BOOT

The initial startup of the resource after the node was started, for example,
once all system related resources are started such as VIP, ONS, etc. All
user defined database services events have a reason code of boot.

AUTOSTART Unable to perform autorestart of the resource.

MEMBER_LEAVE Member of the resource has left the cluster, causing the resource to fail.

PUBLIC_NW_DOWN

The public network interface is not reachable/responding or is down.

( continued )