vcs-esx30_i686-5.1MP2P4

 Basic information
Release type: P-patch
Release date: 2009-03-16
OS update support: None
Technote: 321204
Documentation: None
Popularity: 707 viewed    downloaded
Download size: 7.7 MB
Checksum: 939362949

 Applies to one or more of the following products:
Cluster Server 5.1MP2 On VMware ESX 3.0 i686
Cluster Server 5.1MP2 On VMware ESX 3.5 i686

 Obsolete patches, incompatibilities, superseded patches, or other requirements:
None.

 Fixes the following incidents:
1477058, 1542375

 Patch ID:
None.

Readme file
POINT PATCH FOR ESX AGENTS FOR VERITAS CLUSTER SERVER 5.1MP2 on ESX
===================================================================
Name:5.1MP2P4

Date:2009-03-16

OS:VMware ESX

OS Version:3.0 Update 2 and 3.5 Update 2

Packages:

Etrack Incidents: 1477058, 1542375


RELEVANT ARCHITECTURES:x86

Patches obsoleted by this patch: 5.1MP2P2

Errors/Problems Fixed:
1. To create a new set of ESX agents that will consume less CPU on the ESX
   console OS.
2. GuestOSApp resources go into UNKNOWN state after virtual machine is migrated.
3. ESXDatastore agent reports the following error while bringing a resource
   online:
   V-97-10061-16559 Failed to retrieve values of properties [12:-1:Permission to
   perform this operation was denied.]

Install Instructions:
=====================
Install this patch after installing Veritas Cluster Server 5.1MP2 or
Veritas Cluster Server 5.1MP2P1 or Veritas Cluster Server 5.1MP2P2.

----------------------------------------------------
*** Steps to be run on any one system in the cluster:
----------------------------------------------------

01. Log in as superuser on a system in the cluster where the point patch is to
    be installed.

02. Run upgrade-conf.pl, which is included with this patch, while VCS is
    running.

03. Take all the resources of type GuestOSApp and VMIP offline.

04. Upgrade the configuration by running the command file generated by the
    upgrade-conf.pl.

    Review and modify the configuration, if required, and enable the resources
    of type ESXVM and ESXDatastore.


---------------------------------------------------
*** Steps to be run on each system in the cluster:
---------------------------------------------------

05. Log in as superuser on the system where the point patch is to be installed.

06. Run install-patch.sh, which is included with this patch.


Uninstall Instructions:
=======================

----------------------------------------------------
*** Steps to be run on any one system in the cluster:
----------------------------------------------------

07. Log on as superuser on a system in the cluster where the point patch is to
    be uninstalled.

08. Delete all resources of type ESXDatastore, ESXVM, and ESXVSwitch.

09. Delete the ESXDatastore, ESXVM, and ESXVSwitch resource types.
	$ hatype -delete ESXDatastore
	$ hatype -delete ESXVM
	$ hatype -delete ESXVSwitch

---------------------------------------------------
*** Steps to be run on each system in the cluster:
---------------------------------------------------

10. Log on as superuser on the system where the point patch is to be
    uninstalled.

11. Run uninstall-patch.sh, which is included with this patch.

----------------------------------------------------
*** Steps to be run on any one system in the cluster:
----------------------------------------------------

12. Run the following commands to reset the ESXHost type definition.
	$ hatype -modify ESXHost ArgList -delete -keys
	$ haattr -delete ESXHost HostName
	$ haattr -delete ESXHost User
	$ haattr -delete ESXHost Password
	$ haattr -delete ESXHost NodeName

13. Run the following commands to reset the GuestOSApp type definition.
	$ hatype -modify GuestOSApp ArgList VMwareResName "VMwareResName:CfgFile" "VMwareResName:State" IState
	$ haattr -delete GuestOSApp ESXHostRes


Notes:
======
01. The resources of type ESXVM and ESXDatastore must be enabled before you
    online the service groups.

02. Taking the GuestOSApp and VMIP resources offline in step#3 does not stop
    the actual applications running inside the virtual machine. This step is
    required to prevent any false failover due to the upgrade.
    The resources will go online on their own after the agents are upgraded on
    all cluster nodes, i.e. after step#6. If not, the resources can be brought
    online manually after step#6.

03. Perform the following steps if you modify the resource name of type ESXHost
    after the upgrade:
    a. Update the value of the ESXHostRes attribute for resources of the
       following type: ESXDatastore, ESXVM, ESXVSwitch
       To do so, type the following command:
            $ hares -modify <res-name> ESXHostRes <ESXHost-res-name>
    b. Update the ESXHostRes attribute of the GuestOSApp resource type. 
            $ hatype -modify GuestOSApp ESXHostRes <ESXHost-res-name>
    c. Restart the GuestOSApp agent on all the nodes where resources of this
       type are configured.
            $ haagent -stop GuestOSApp -force -sys <system>
            $ haagent -start GuestOSApp -sys <system>