Notice: This website is an unofficial Microsoft Knowledge Base (hereinafter KB) archive and is intended to provide a reliable access to deleted content from Microsoft KB. All KB articles are owned by Microsoft Corporation. Read full disclaimer for more details.

Description of what to consider when you deploy Windows Server 2008 failover cluster nodes on different, routed subnets


INTRODUCTION

This article describes what to consider when you deploy Windows Server 2008 failover cluster nodes on different, routed subnets. Such deployments may cause re-connection delays when a failover of a highly available service or an application occurs.

↑ Back to the top


More Information

One of the new features that Windows Server 2008 failover clusters introduce is the ability to locate cluster nodes on different, routed subnets. Therefore, Client Access Point (CAP) configurations have cluster network name resources that depend on multiple IP address resources. The cluster logic that is used in this configuration determines whether a network name resource comes online if at least one of the IP Address resources on which it depends can successfully come online.

When a cluster network name resource comes online, it registers with a DNS server if the DNS server supports dynamic registrations. Any A-Record or AAAA-Record that is registered includes the following information:
  • The IP address information for the IP address resource that successfully comes online
  • A default Time to Live (TTL) entry
By default, the record is updated at least one time every 24-hour period. The default TTL that is registered is 20 minutes.

The default private property settings for cluster network name resources may cause delays when clients try to re-connect to cluster resources after a failover of a highly available application or a service group. There are network name resource private properties that can be modified to optimize name resolution behavior. Such optimization makes clients re-connect faster after a failover. The following is a list of the default settings for the private properties that may help in this scenario.
PropertyDefault setting
HostRecordTTL1200 (0x4b0) seconds
RegisterAllProvidersIP0
Notes
  • HostRecordTTL:

    This is the time during which clients cache DNS record information. The default is 20 minutes.
  • RegisterAllProvidersIP:

    This setting determines whether all IP addresses on which a network name resource depends are registered in DNS or whether only the IP address resource(s) that successfully come online are registered in DNS. The default setting is to register only those IP address resources that come online.

    Note This setting cannot be used in Microsoft Exchange Server 2007 clusters. It will adversely impact MAPI functionality.
You can modify these settings to help with client re-connections. However, it may also result in increased traffic to the DNS servers. For more information about configuring these settings, visit the following Web site:
Configure Heartbeat and DNS Settings in a Multi-Site Failover Cluster
http://technet.microsoft.com/en-us/library/dd197562.aspx
A variable that is not controlled in the context of failover clustering is the replication of DNS zone information. However, you should discuss the scenario in which replication of DNS zone information in not controlled in the context of failover clustering with those individuals who can modify DNS zone replication in an organization.

For more information about DNS and dynamic updates, visit the following Microsoft Web page:For more information, click the following article numbers to view the articles in the Microsoft Knowledge Base:

816592 How to configure DNS dynamic updates in Windows Server 2003

246804 How to enable or disable DNS updates in Windows 2000 and in Windows Server 2003

↑ Back to the top


Keywords: kbclustering, kbexpertiseinter, kbhowto, kbinfo, kb

↑ Back to the top

Article Info
Article ID : 947048
Revision : 3
Created on : 4/18/2018
Published on : 4/19/2018
Exists online : False
Views : 3080