Testing for Availability - Expert Oracle RAC Performance Diagnostics and Tuning - page 65

■ Note We discuss nhb and Dhb for node verification in much more detail in subsequent chapters.

CLUSter paraMeterS

the three parameters that control the node eviction rules and time out periods are

Misscount (MC)—defaults to 30 seconds. it's the maximum time in seconds that an nhb can be missed before
entering into a cluster reconfiguration to evict the node. (the default value of this parameter when using vendor
clusterware is 600 seconds to resolve any possible split-brain scenarios.)

Disktimeout (Dto)—defaults to 200 seconds. this is the time before an i/o operation against the voting disk
should complete.

reboottime (rbt)—defaults to 3 seconds. this is the time allowed for a node to complete a reboot after the
Cluster synchronization services (Css) daemon has been evicted.

this information can be obtained using the oracle Clusterware Control ( crsctl) utility:

$GRID_HOME/bin/crsctl get css
Usage:
crsctl get css <parameter>
Displays the value of a Cluster Synchronization Services parameter
clusterguid
diagwait
disktimeout
misscount
reboottime
priority
logfilesize
leafmisscount
$GRID_HOME/bin/crsctl get css reboottime
CRS-4678: Successful get reboottime 3 for Cluster Synchronization Services.

The voting disk, on the other hand, is updated by each node in the cluster with a time stamp entry once every
so many seconds. When a member arbitrates regarding the other members availability, the voting disk is consulted
by the OCW to verify this arbitration. Once this configuration is tested, the arbitrating instance uses the shared
disk to publish the proposed configuration to the other instances. All active instances then examine the published
configuration, and, if necessary, terminate themselves.
The reconfiguration process regroups the nodes that are accessible and removes the nodes that have failed. For
example, in a six-node cluster, if one node fails, the OCW will re-group among the remaining five nodes. The OCW
performs the reorganization of the cluster membership when a node is added to or removed from a cluster. This
information is exposed to the respective Oracle instances by the Global Enqueue Service Monitor ( LMON ) process on
the participating instances.