* * * READ ME * * *
               * * * Veritas Volume Manager 4.1 MP2 * * *
                      * * * Rolling Patch 11 * * *
                         Patch Date: 2012-03-21


This document provides the following information:

   * PATCH NAME
   * PACKAGES AFFECTED BY THE PATCH
   * BASE PRODUCT VERSIONS FOR THE PATCH
   * OPERATING SYSTEMS SUPPORTED BY THE PATCH
   * INCIDENTS FIXED BY THE PATCH
   * INSTALLATION PRE-REQUISITES
   * INSTALLING THE PATCH
   * REMOVING THE PATCH


PATCH NAME
----------
Veritas Volume Manager 4.1 MP2 Rolling Patch 11


PACKAGES AFFECTED BY THE PATCH
------------------------------
VRTSvxvm
VRTSvxvm


BASE PRODUCT VERSIONS FOR THE PATCH
-----------------------------------
   * Veritas Volume Manager 4.1 MP2
   * Veritas Storage Foundation for Oracle RAC 4.1 MP2
   * Veritas Storage Foundation Cluster File System 4.1 MP2
   * Veritas Volume Replicator 4.1 MP2
   * Veritas Storage Foundation 4.1 MP2
   * Veritas Storage Foundation High Availability 4.1 MP2


OPERATING SYSTEMS SUPPORTED BY THE PATCH
----------------------------------------
HP-UX 11i v2 (11.23)


INCIDENTS FIXED BY THE PATCH
----------------------------
This patch fixes the following Symantec incidents:

Patch ID: PHCO_42767, PHKL_42786

* 2325986 (Tracking ID: 622508)

SYMPTOM:
LUNs with different serial numbers are grouped together under a single DMP node.

DESCRIPTION:
Device discovery layer (DDL) differentiates each LUN based on its serial 
number. However, DDL treats comma (",") as a delimiter within a serial number 
and ignores the preceding characters before the comma. For example, two 
different LUNs having serial numbers "57#i, 000Q" and "57#j, 000Q" are wrongly 
claimed under a single DMP node.

RESOLUTION:
DDL routines are modified not to consider the comma as a delimiter for LUN 
serial numbers.

* 2327632 (Tracking ID: 1545835)

SYMPTOM:
Vxconfigd dumps core when disk groups (DG) with unsupported DG versions are 
imported.
 
Stack trace looks like:
 
voldbsup_rtyp2vers ()
voldbsup_unformat_record () 
voldb_unformat_record () 
extract_new_record () 
ddb_extract () 
dg_config_read ()
dg_import_start ()
dg_auto_import_all ()
mode_set ()
setup_mode ()
startup ()
main ()

DESCRIPTION:
Vxconfigd dumps core with "Segmentation Fault" error while validating the disk 
group version as part of its import. During this validation, when a disk group 
with unsupported version is found, vxconfigd fails the validation by returning 
NULL thereby causing NULL pointer dereference within the code. This results in 
vxconfigd dumping core.

RESOLUTION:
Code Changes have been made to handle the NULL pointer dereference problem.

* 2566195 (Tracking ID: 2566192)

SYMPTOM:
In case of HP ALUA disk array, slave node is not able to join the cluster with 
the below messages in syslog:
    VxVM vxconfigd WARNING V-5-1-8222 slave: missing disk 1173484051.196.rx3600-
1
    VxVM vxconfigd WARNING V-5-1-7830 cannot find disk 1173484051.196.rx3600-1

DESCRIPTION:
As part of node join protocol, Master exchanges Dynamic Multipathing (DMP) 
database with node that is joining the cluster. As part of this, slave node 
updates the preferred path which will be selected for further I/Os. But slave 
node cannot find the paths through portid sent by Master. As a result slave node 
fails to join the cluster assuming disk is not found.

RESOLUTION:
Modified the code to use target port groupid instead of portid during node 
join protocol.

* 2672190 (Tracking ID: 2672194)

SYMPTOM:
PHCO_39779 introduced behavior that on systems with VxVM installed but not
configured will cause empty directories to be left around in the /tmp directory.
Here is an examples of such directories:
	
	    drwx------ 2 root root 96 Jun 4 13:18 \
		vx.12683.24477.8798.2336
	    drwx------ 2 root root 96 Jun 4 13:18 \
		vx.1780.8286.5780.2312
	    drwx------ 2 root root 96 Jun 4 13:16 \
		vx.18308.4973.32167.437
	    drwx------ 2 root sys 96 Jun 4 13:11 \
		vx.25660.20106.11892.9397

DESCRIPTION:
In general the script should call quit(), in which it call do the cleanup when
errors detected. /sbin/init.d/vras-vradmind.sh and /sbin/init.d/vxrsyncd.sh call
exit() directly instead of quit() to leave some random-created directories not
cleaned.

RESOLUTION:
These script should be restored to call quit() instead of exit() directly.

* 2695313 (Tracking ID: 2695315)

SYMPTOM:
VxVM Command Patch PHCO_42767 patches the following kernel modules in the VxVM
command fileset, VRTSvxvm.VXVM-RUN.

/usr/conf/mod/dmphpalua
/usr/conf/mod/dmphdsalua
 
Installing the patch will rebuild the kernel but a reboot is not enforced.

DESCRIPTION:
Typically VxVM command patches do not issue a reboot.  In order to continue 
following that, PHCO_42767 is not flagged for reboot. As such, PHCO_42767 should
be installed along with a co-dependent VxVM kernel patch, which will serve to
trigger the reboot.

RESOLUTION:
A new VxVM kernel patch is released to be installed with PHCO_42767 to enforce a
reboot.


INSTALLING THE PATCH
--------------------
$ swinstall -x autoreboot=true <patch id> 
Please do swverify after installing the patches in order to make sure
   that the patches are installed correctly using:

   $ swverify <patch id>


REMOVING THE PATCH
------------------
To remove the patch, enter the following command:

        # swremove  -x autoreboot=true <patch id>


SPECIAL INSTRUCTIONS
--------------------
For PHCO_42767:
---------------

If you are using Serviceguard with CVM 4.1 or CFS 4.1
products, and are not installing on an inactive DRD system
image, you must halt your cluster before applying this
patch. All the nodes within your cluster must have the
same level of patches.

The patch rebuilds the kernel. A system reboot is needed
to put the new kernel in effect. Please install it
together with the dependent VxVM kernel patch.


For PHKL_42786:
---------------

JAGag22549 fix brings in a new tunable
voldrl_volumemax_drtregs which can be used along with the
prior existing tunable voldrl_max_drtregs to maximize
performance on VxVM volumes with DRL logs.

In order to fix JAGaf65522, the patch recreates DRL logs
on all fully enabled volumes on all imported disk groups.
If a volume has been stopped for some reason and/or disk
group deported at the moment of patch installation, a user
has to manually remove and recreate DRL logs. This
includes systems with disk groups being a part of a
cluster package which have packages brought down during
patch installation. Otherwise, resync problems can still
be seen on volumes with non-modified DRL logs.

If you are using Serviceguard with CVM 4.1 or CFS 4.1
products, and are not installing on an inactive DRD system
image, you must halt your cluster before applying this
patch. All the nodes within your cluster must have the
same level of patches.


OTHERS
------
NONE