Veritas cluster migrating to a new array.

[sreeks]

We have veritas cluster version -5 running on solaris 10. We need to migrate the storage to a new array. Mirrioring volumes is not possible as there are many diskgroups and volumes. The plan is to sync all the LUN's to new array and siwtch the cables, deport and import disk groups. Can someone help me on the steps to perform this migration.

[mikebounds · Accepted solution]

You can do online using volume manager mirroring at volume level and in terms of I/O traffic this is no different to mirroring at a LUN level using the array.   If you use array mirroring then deport and import of disk groups should work fine.  If you are concerned about the amount of work mirroring at a volume level,then you can use VOM - see extract from http://origin-symwisedownload.symantec.com/resources/sites/SYMWISE/content/live/DOCUMENTATION/7000/D...

Enterprises frequently need to move (migrate) storage volumes to new disk
enclosures. Volume migration can be time-consuming and complex. Storage
Provisioning and Enclosure Migration Add-on simplifies the complexities so that
you can easily migrate volumes between disk enclosures.
Using the Management Server console, you can migrate volumes in the following
ways.
See “Migrating volumes by host” on page 216.
See “Migrating volumes by enclosure” on page 219.
See “Migrating volumes by disk group” on page 221.

Mike

[sreeks]

Thanks Mike for your reply. Can you please send the steps to perform deport and import of diskgroups. and what care needs to be taken. This will be of great help for me.

[KNR]

Graceful OS shutdown should deport the disk groups and VCS shutdown deports the DGs part of service groups. You may comment out SAN file systems in vfstab and disable VCS auto startup until you verify all the disks in all the DGs are available for import after the post replication reboot. You have to perform additional steps if you are not using DMP for multipathing. Brief summary of steps to follow. 1. Unmount the filesystems, stop volumes, deport DG, comment out in vfstab. (For DGs which are not part of VCS service group) Umount mount-point Vxvol -g DG stopall Vxdg deport DG Vi /etc/vfstab Perform the same on other node as well if any local DGs. 2. Stop VCS, disable VCS auto startup. Hastop -all Mv rc2.d/SVCS scripyT 3. Verify all DGs are deported on all the nodes. Vxdisk -o alldgs list. 4. Let the SAN replication complete and present the new storage Luns. 5. Solaris 10 with auto discover Luns enabled system should see the new Luns. However, reboot helps to clean the stale devices and to rebuild the DMP. Reconfigure reboot may also be considered. Cfgadm -c configure controller helps you to scan the new Luns if required. 6. After the reboot, verify all disks are available. And import the local disk group, start volumes, mount and uncomment in fstab. And start VCS, and rename the VCS startup script to its original name. If you are not able to imporT DGs because of UDID mismatch, you should find tech note for the same with instructions to clear the flag. HTH, KNR