How to perform a device tree cleanup with EMC storage, EMC Powerpath and Volume Manager in the configuration on a Solaris system

[해리]

https://www.veritas.com/support/en_US/article.TECH78463

Problem

How to perform a device tree cleanup with EMC storage, EMC Powerpath and Volume Manager in the configuration on a Solaris system

Solution

There may be some situations where a device tree cleanup may be required, and it is important to follow a sequence of steps, especially in environments using Third Party Drivers like EMC Powerpath for managing disk multi-pathing. This is to ensure that the device cleanup procedure addresses all the requirements from the OS, EMC Powerpath and Veritas Volume Manager perspective.

The following procedure is a Symantec Technical Support tested procedure to perform a device tree cleanup with EMC storage, EMC Powerpath and Volume Manager in the configuration on a Solaris system, where required.

Customers may want to consult EMC/SUN to verify the non Symantec steps involved in the following processes:

1.]     First need to confirm‎ if there is any Cluster in the setup before starting any device clean up procedure

If cluster exists:
Freeze all the service groups or freeze the system(s) themselves that are in the cluster configuration before performing any of the device tree cleanup procedure to avoid any  unnecessary and unexpected failovers.

Note:

In 4.1 VCS and below:

Freezing a system Freeze a system to prevent the service groups that it hosts from failing over to another system.

In 5.0 VCS and above:

Freezing a system Freeze a system to prevent service groups from coming online on the system.

# haconf -makerw

# hagrp -freeze <Service_Group_Name> -persistent
or freeze all of the systems in the cluster instead of freezing each service group individually
(Freezing a system only prevents a service group from coming online on that system in VCS 5.0 and above.)
# hasys -freeze -persistent <System_Name>  (repeat for other systems in the cluster)
# haconf -dump -makero

2.]     Stop the vxesd daemon to ensure that no device discovery is taking place by Volume Manager

# vxddladm stop eventsource

3.]     Clear up the EMC Powerpath configuration

# mv /kernel/drv/emcp.conf /kernel/drv/empc.conf.orig
# mv /etc/powermt.custom /etc/powermt.custom.orig
# mv /etc/emcp_devicesDB.dat /etc/emcp_devicesDB.dat.orig
# mv /etc/emcp_devicesDB.idx /etc/emcp_devicesDB.idx.orig
# rm -rf /dev/dsk/ecmpower*
 

NOTE:  If powerpath fails to start on reboot, you might have to copy these .orig files back into place.

4.]     Clear up the Volume Manager configuration

# cd /etc/vx
# mv disk.info disk.info.old
# mv array.info array.info.old
# mv jbod.info jbod.info.old
# rm -rf /dev/vx/rdmp/*
# rm -rf /dev/vx/dmp/*

5.]     Now, if possible, a reconfiguration reboot is recommended.

# reboot -- -r
or
# init 0
ok  boot -r


If a reconfiguration reboot is not possible, then proceed as below:

6.]     Device Tree Clean up

    a.)     OS level
    # devfsadm -Cv

    b.)     EMC Powerpath Level
    # powercf -q
    # powermt config
    # powermt display dev=all
    # powermt save

    c.)     Volume Manager Level

    # vxddladm set namingscheme=ebn persistence=no
    # vxddladm set namingscheme=ebn persistence=yes
    # vxconfigd -k -m enable -x cleartempdir

Please Note: If running VxVM version 5.0 MP3 or later, and, if using AVID=Yes (refer to "vxddladm get namingscheme"), then it is recommended to switch off persistence naming

7.]     Re-Start the vxesd daemon

# vxddladm start eventsource

8.]     Once everything is confirmed to be ok and "hastatus -sum" shows a good output, the service groups can be unfrozen
# haconf -makerw

# hagrp -unfreeze <Service_Group_Name> -persistent
or if systems were frozen then:
# hasys -unfreeze -persistent <System_Name>  (for all systems frozen.)

Finally save and dump the VCS configuration
# haconf -dump -makero