In Solaris , Sometimes veritas volume manger device tree will be messed up with newly assigned LUNS and you have to perform the device tree cleanup in a order to fix the device tree mess up.
Option #1: Perform device tree reset procedure (non-reboot method) on each participating node.
Step #1: Rename ‘.info’ files.
# mv /etc/vx/array.info /etc/vx/old_array.info
# mv /etc/vx/disk.info /etc/vx/old_disk.info
# mv /etc/vx/jbod.info /etc/vx/old_jbod.info
# mv /etc/vx/dmppolicy.info /etc/vx/old_dmppolicy.info (if present)
Step #2: Remove ‘unwanted” entries from the /dev/dsk, /dev/rdsk, /dev/vx/dmp, and /dev/vx/rdmp directories. Be careful NOT to remove the bootdisk or bootmirror from any of these device directories.
Step #3: Update device tree.
# devfsadm -C
# vxconfigd -k -s syslog (freeze HA cluster before running this command)
# vxdctl initdmp
# vxdctl enable
Below are two options for resolving device tree inconsistencies in VXVM. The first option does not require a reboot. The second option does require reboot but has a higher success rate as the OS kernel will be flushed on reboot.
If the first option fails to resolve issue you will need to perform option#2 to resolve the matter. If the system is within a VCS Cluster, this procedure must be performed on each node so that the device tree is consistent across the cluster.
It is recommended that you collect a VRTSexplorer from each of the affected nodes prior to performing either of the two options.
If the first option fails to resolve issue you will need to perform option#2 to resolve the matter. If the system is within a VCS Cluster, this procedure must be performed on each node so that the device tree is consistent across the cluster.
It is recommended that you collect a VRTSexplorer from each of the affected nodes prior to performing either of the two options.
Option #1: Perform device tree reset procedure (non-reboot method) on each participating node.
Step #1: Rename ‘.info’ files.
# mv /etc/vx/array.info /etc/vx/old_array.info
# mv /etc/vx/disk.info /etc/vx/old_disk.info
# mv /etc/vx/jbod.info /etc/vx/old_jbod.info
# mv /etc/vx/dmppolicy.info /etc/vx/old_dmppolicy.info (if present)
Step #2: Remove ‘unwanted” entries from the /dev/dsk, /dev/rdsk, /dev/vx/dmp, and /dev/vx/rdmp directories. Be careful NOT to remove the bootdisk or bootmirror from any of these device directories.
Step #3: Update device tree.
# devfsadm -C
# vxconfigd -k -s syslog (freeze HA cluster before running this command)
# vxdctl initdmp
# vxdctl enable
Option #2: Perform device tree reset procedure (reboot method) on each participating node.
Perform device tree reset procedure.
To fully update device tree perform the following procedure.
To reset invalid device entries remove all entries except the rootdisk, rootmirror, or any of the ‘rootdg’ devices from /dev/dsk, /dev/rdsk, /dev/vx/dmp, and /dev/vx/rdmp directories.
Again, do not remove the bootdisk, bootmirror, or ‘rootdg’ devices from any of these directories.
IMPORTANT: If the boot disk or boot mirror device tree entries are removed the system will not boot and the operating system will need to be re-installed. Lastly, rename the ‘array.info’, ‘disk.info’, ‘jbod.info’, and ‘dmppolicy.info’ (if present) files located in the /etc/vx directory. Once this is done perform reconfiguration reboot.
See below.
# mv /etc/vx/array.info /etc/vx/old_array.info
# mv /etc/vx/disk.info /etc/vx/old_disk.info (if present)
# mv /etc/vx/array.info /etc/vx/old_array.info
# mv /etc/vx/disk.info /etc/vx/old_disk.info (if present)
# mv /etc/vx/jbod.info /etc/vx/old_jbod.info
# mv /etc/vx/dmppolicy.info /etc/vx/old_dmppolicy.info
NOTE: If EMC PowerPath is in use run powermt config and powermt save.
Once all the above items have been addressed perform a reconfiguration reboot.
# touch /reconfigure
# shutdown -i6 -y -g0
The article is based on Symantec technical support.
Thank you for reading this article.Please leave a comment if you have any doubt ,i will get back to you as soon as possible.
# mv /etc/vx/dmppolicy.info /etc/vx/old_dmppolicy.info
NOTE: If EMC PowerPath is in use run powermt config and powermt save.
Once all the above items have been addressed perform a reconfiguration reboot.
# touch /reconfigure
# shutdown -i6 -y -g0
The article is based on Symantec technical support.
Thank you for reading this article.Please leave a comment if you have any doubt ,i will get back to you as soon as possible.
Leave a Reply