Date: 2009-05-15
OS: Windows
OS Version: 2000, 2003
Packages: Private_5_0_0301_1587279_x86.zip, Private_5_0_0301_1587279_ia64.zip, Private_5_0_0301_1587279_x64.zip
Etrack Incidents: 1587279
Errors/Problems Fixed:
If there is a resource fault while taking a service group offline on a node, VCS fails to update the service group state details in the internal configuration. As a result, if that node is faulted, VCS triggers a local failover and brings the service group online on an alternate node. 
This failover is unintentional and should not have occurred as the service group was taken offline intentionally, either for modifying the resources or for some other maintenance purpose. 
This problem exaggerates in a global cluster (GCO) environment, when the global group is being taken offline on the primary site to bring it online on the secondary site. The group remains in offline state at the primary site and online at secondary site until the node at the primary site is in RUNNING state.
As soon as the primary site node is faulted the service group comes online at the primary site. This results in a concurrency violation and possible data corruption.
Fixes Applied for Products:
Veritas Storage Foundation and High Availability Solutions 5.0 RP2 for Windows
Install/Uninstall Instructions:
1.Download the private zip file to a temporary location on your cluster node.
See the table below to determine the file you must download:
       ---------------------------------------------------------
       | Operating System  | Name of Private File to Download  | 
       ---------------------------------------------------------
       | Windows x86       | Private_5_0_0301_1587279_x86.zip  |
       ---------------------------------------------------------
       | Windows ia64      | Private_5_0_0301_1587279_ia64.zip |
       ---------------------------------------------------------
       | Windows x64       | Private_5_0_0301_1587279_x64.zip  |
       ---------------------------------------------------------
Each .zip file contains the following file:
       -------------------------------------------
       | Name of the file   | New Version number |
       -------------------------------------------
       | had.exe            |  5.0.301.497       | 
       -------------------------------------------
2.Extract the file to a temporary location on your cluster node.
3.Stop Veritas high availability engine (HAD) on the cluster node.
Type the following on the command prompt:
	hastop -local -force
4.Copy the file, had.exe, to %VCS_HOME%in directory.
Here, the variable %VCS_HOME% is the default installation directory for VCS, typically C:Program FilesVeritasCluster Server.
5.Repeat steps 3 and 4 on all the cluster nodes.
6.Start the Veritas high availability engine (HAD) on all the cluster nodes. 
Type the following on the command prompt:
	hastart