Date:2008-11-19
OS:Linux x86_64
OS Version:RHEL 4 Update 6
Packages:VRTSvcsag
Etrack Incidents: 1450300

POINT PATCH FOR VERITAS CLUSTER SERVER 5.0 MP3
===============================================
NAME: 				5.0MP3+e1450300
DATE:                           2008-Nov-19
VCS RELEASE:			5.0 MP3
LINUX RELEASE:                  RHEL 4 Update 6
RELEVANT ARCHITECTURES:         x86
BUGID'S FIXED WITH THIS PATCH:  e1450300

PROBLEM DESCRIPTION:
	Escalation: 281-428-099
	-----------------------
 	NFSRestart can't be probed when it's not in a service group
	Etrack Ids:
	-----------
	1. 1450300: NFSRestart resource is not probed if not in a service group


Prerequisite: 
-------------

	Ensure that you have VCS 5.0 MP3 installed on every node before you 
	install the patch.


PATCH INSTALLATION INSTRUCTIONS:
--------------------------------

To install the patch, perform the following steps:

1. Log on as superuser on the system where the point patch needs to be
   installed.

2. Open the VCS configuration in read-write mode using the following command:
   # haconf -makerw

3. Freeze all the service groups persistently using the following command:
   # hagrp -freeze group_name -persistent

4. Save the VCS configuration using the following command:
   # haconf -dump -makero

5. Shut down VCS on all nodes using the following command:

   # hastop -all -force

6. Run the following command to change the directory to rpm location:

   # cd <rpm_location>

7. Run the following command to apply the NFS hotfix:

   # rpm -Uvh VRTSvcsag-5.0.30.02-MP3HF2_RHEL4.i686.rpm

8. Repeat step 4 and 5 on each node of the cluster.

9. Start VCS on all the nodes

10. Open the VCS configuration in read-write mode using the following command:
   # haconf -makerw

11. Unfreeze all service groups using the following command:
   # hagrp -unfreeze group_name -persistent

12. Save the VCS configuration using the following command:

   # haconf -dump -makero"


Change in configuration for NFS lock failover:
==============================================

You must modify the configurations for the NFSRestart agent to include a DNS
resource. This DNS resource depends on the IP resource, and the NFSRestart
resource depends on the DNS resource.
These dependencies are shown in the following extract from a sample
configuration for the NFSRestart resource. The sample configuration also
includes the DNS resource definition for use with the NFSRestart agent.

	DNS dnsres (
		Domain = "marketing.db.com"
		StealthMasters = {"11.123.10.10"}
		ResRecord@sysA = {sysA= "11.123.175.11"}
		ResRecord@sysB = {sysB= "11.123.175.11"}
		TSIGKeyFile="/var/tsig/Kveritas.com.+157+00000.private"
	)
	ip1 requires nic1
	ip1 requires Share_dir1
	Mount_dir1 requires dg01_vol01
	nfsres1 requires dnsres
	dnsres requires ip1
	Share_dir1 requires Mount_dir1
	Share_dir1 requires nfs1
	dg01_vol01 requires dg01
The DNS record enables NFS clients to identify the associated NFS server node
in a cluster. The NFS clients that require lock recovery must be able to map
this virtual IP address to hostname.