Designing the Compaq ProLiant Clusters HA/F100 and HA/F200
2-41
Compaq Confidential – Need to Know Required
Writer:
Bryan Hicks
Project:
Compaq ProLiant Clusters HA/F100 and HA/F200 Administrator Guide
Comments:
Part Number:
380362-003
File Name:
c-ch2 Designing the Compaq ProLiant Clusters HAF100 and HAF200.doc
Last Saved On:
8/24/00 12:00 PM
You can use the Windows performance tool to observe and track system
performance. Some applications may also have their own internal performance
measurement capabilities.
Microsoft Clustering Software Thresholds and
Periods
The Microsoft clustering software offers flexibility in configuring the
initiation of failover events. For resources, the Microsoft clustering software
allows you to set Restart Thresholds and Periods. For cluster groups, the
Microsoft clustering software allows you to set Failover Thresholds and
Periods.
Restart Threshold and Restart Period
A restart threshold defines the maximum number of times per restart period
that the Microsoft clustering software attempts to restart a resource before
failing over the resource and its corresponding cluster group. See the following
example:
Assume you have a disk resource (Disk1) that is part of a cluster group
(Group1). You set the restart threshold to 5 and the restart period to 10. If the
Disk1 resource fails, the Microsoft clustering software will attempt to restart
the resource on the group’s current cluster node five times within a 10-minute
period. If the resource cannot be brought online within the 10-minute restart
period, then Group1 will fail over to the partner cluster node.
Note that the Microsoft clustering software waits the length of the restart
period (for example, 10 minutes) before actually failing over the cluster group.
You must assess the likelihood that the group will successfully restart on its
present node against the time required to restart the cluster group before failing
it over. If it is appropriate to immediately fail over any group that encounters a
problem, set the restart threshold to 0 (zero). If the group will experience
severe performance limitations if failed over to a secondary server, set the
threshold and period so that the Microsoft clustering software attempts to
restart the group on its preferred server.
Summary of Contents for ProLiant Clusters HA/F100
Page 8: ......
