PTF Cover Letter
PTF ( Program Temporary Fixes ) Cover letter
OSP Cluster Admin Domain Changes.
Pre/Co-Requisite PTF / Fix List
REQ LICENSED PTF/FIX LEVEL
TYPE PROGRAM REL NUMBER MIN/MAX OPTION
---- -------- --- ------- ------- ------
Application of this PTF may disable or render ineffective programs that
use system memory addresses not generated by the IBM translator,
including programs that circumvent control technology designed to limit
interactive capacity to purchased levels. This PTF may be a prerequisite
for future PTFs. By applying this PTF you authorize and agree to the
This PTF is subject to the terms of the license agreement which
accompanied, or was contained in, the Program for which you are obtaining
the PTF. You are not authorized to install or use the PTF except as part
of a Program for which you have a valid Proof of Entitlement.
SUBJECT TO ANY WARRANTIES WHICH CAN NOT BE EXCLUDED OR EXCEPT AS EXPLICITLY
AGREED TO IN THE APPLICABLE LICENSE AGREEMENT OR AN APPLICABLE SUPPORT
AGREEMENT, IBM MAKES NO WARRANTIES OR CONDITIONS EITHER EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OR CONDITIONS OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON INFRINGEMENT,
REGARDING THE PTF.
The applicable license agreement may have been provided to you in printed
form and/or may be viewed using the Work with Software Agreements (WRKSFWAGR)
APAR Error Description / Circumvention
This fix changes the behavior of an admin domain when adding an
MRE. Old behavior is to fail the add when an object cannot be
created on all nodes, if it does not exist. New behavior is to
allow the MRE to be added to an Admin Domain even if the
underlying object does not exist and cannot be created during
In addition, user profiles could not be added to a cluster
administrative domain if the user profile exists on the target
nodes and the password is not available.
Adding an *NWSD MRE with a network server storage link (NWSSTGL)
attached with the ACCESS parameter set to *SHRUPD will not work.
MSGCPD0084 willbe seen in the admin domain CRG job log on the
CORRECTION FOR APAR SE47523 :
The behavior of the admin domain is changed so that if an object
does not exist, and cannot be created during an add MRE
operation, the MRE for the resource will still be added to the
cluster administrative domain.
In addition, user profiles can now be added to a cluster
administrative domain even if the password is not available on
the target nodes. If the PTF is applied on all nodes in your
cluster administrative domain, a user profile can be added to
the domain if the following conditions are met.
1. QRETSVRSEC must be set to 1 on all nodes in the domain.
2. The password for the user profile on the node where the add
MRE operation is initiated (the source node) must have been
changed after QRETSVRSEC was set to 1.
3. The cluster administrative domain must be ACTIVE. You can
use the STRCAD command to start the administrative domain.
If the cluster administrative domain is not active, and the
password is not available on all nodes in the domain, the add
MRE operation will fail as before.
If the cluster administrative domain is active, the user profile
will be created on any node where it did not exist. On nodes
where the user profile existed prior to the add MRE operation,
the password will be changed using the password from the source
node, even if the password was not previously available on the
An *NWSD with attribute NWSSTGL will now be able to be added to
the admin domain when the ACCESS parameter for the NWSSTGL is
The PTF that changes this behavior must be applied to all nodes
in the admin domain, else the "old behavior" e.g., the add MRE
operation fails if the object does not exist and cannot be
created on all nodes in the domain, will prevail.
CIRCUMVENTION FOR APAR SE47523 :
You must follow the Special Instructions for the superseded PTF
THE FOLLOWING ARE SUPERSEDED SPECIAL INSTRUCTIONS. IF THE SUPERSEDED
PTF HAS ALREADY BEEN APPLIED AND ITS SPECIAL INSTRUCTION FOLLOWED,
IT IS NOT NECESSARY TO FOLLOW THAT SPECIAL INSTRUCTION AGAIN.
SPECIAL INSTRUCTIONS FOR SUPERSEDED PTF SI39264 :
You must follow the Special Instructions for the superseded PTF
SPECIAL INSTRUCTIONS FOR SUPERSEDED PTF SI31138 :
In order to fix the problems described by this PTF, the PTF must be
applied on every node in the cluster. You do not have to end the
cluster to do this. However, you should end the CRG jobs associated
with the cluster administrative domain on every cluster node before you
apply the PTF. You can do this one node at a time.
Note, you must end the cluster administrative domain CRG jobs on each
cluster node, apply the PTF, and restart the CRG job before the fix is
The cluster administrative domain CRG jobs will be ended if you end
Cluster Resource Services for each node in your cluster. Cluster
Resource Services can be ended using the End Cluster Node API or CL
command. Optionally, you can end just the cluster administrative domain
jobs without ending your cluster nodes by calling the Change Cluster
Recovery CL command.
To determine if any cluster administrative domain jobs are active on
your system, use the WRKACTJOB command from the command line of your
system. Look for any system jobs which have the same name as any
cluster administrative domains which you have created in your cluster.
If there are any cluster administrative domain jobs on your system,
they must be ended. You can end just the cluster administrative domain
jobs on every node with this command:
CHGCLURCY CLUSTER(<cluster>) CRG(<admin_domain>) NODE(*ALL)
where <cluster> is the name of your cluster and <admin_domain> is the
name of a cluster administrative domain. You should call this command
once for each cluster administrative domain in your cluster.
You must sign off all sessions where any cluster administrative domain
commands or APIs were used before the PTF can be applied.
After the PTF has been applied on a system, you can restart the cluster
jobs that were ended before the PTF was applied. If Cluster Resource
Services was ended, you can use the Start Cluster Node API or CL
command to start Cluster Resource Services for each cluster node. If
you used the Change Cluster Recovery CL command to end just the cluster
administrative domain jobs, they can be restarted on each cluster node
using the Change Cluster Recovery command as follows:
CHGCLURCY CLUSTER(<cluster>) CRG(<admin_domain>) NODE(<node>)
where <cluster> is the name of your cluster, <admin_domain> is the name
of the cluster administrative domain, and <node> is the cluster node
name for the system. You should call this command once for each cluster
administrative domain in your cluster.
After the PTF has been applied, you should also restart the QPRFSYNCH
job on every node in your cluster. You can restart the QPRFSYNCH job
with the following commands:
CALL PGM(QSYS/QFPAPRFJ) PARM(END)
CALL PGM(QSYS/QFPAPRFJ) PARM(START)
THIS PTF CAN BE APPLIED IMMEDIATE OR DELAYED.
PTF/FIX NO(S). APAR TITLE LINE
SI39264 OSP Cluster Admin Domain Changes.
SI31138 OSP-INCORROUT Admin domain incorrectly shows *NWSD is Consis
SI31138 OSP-MSGCPD0019-T/QWCSVALOH CHGSYSVAL QCURSYM to %
SI29033 OSP-INCORROUT Add MRE fails in mixed release admin domain
SI29033 OSP INCORROUT Failed to add *USRPRF MRE when QRETSVRSEC set
|Superseded by:......................||View fix details for PTF SI56943|
|MRI Feature ........................||NONE|
05 May 2011