Setting mistakes can be prevented during storage migration and storage can be migrated efficiently by backing up and restoring object configuration information by using the sdxconfig command.
Migrate volume data by using the copy function per LUN (such as ACM CCM) of storage units in one of the following methods:
Copy to one migration destination case from one migration source case.
Copy to each migration destination case from each migration source case.
Figure H.3 Copying to one migration destination case from one migration source case
Figure H.4 Copying to each migration destination case from each migration source case
Information
For a single volume or a mirror volume in a mirror group which is not hierarchized, it can be migrated by using the copy function per slice (such as REC) of storage units. In this case, create a volume of the same size as the migration source on the migration destination storage, and then copy data between slices configuring the migration source and destination volumes.
The migration procedure is as follows:
1) Copy data to the migration destination storage from the migration source storage.
1-1) Copy data to the migration destination storage from the migration source storage per LUN by using the copy function of storage.
1-2) When copying of storage become synchronized, stop services where storage is used.
After that, once volumes are stopped, stop copying of storage.
2) Back up and delete the GDS configuration from the migration source storage.
2-1) Back up object configuration information. (sdxconfig Backup command)
2-2) Stop volumes. (sdxvolume -F command)
2-3) For a shared disk in the cluster system, convert the migration source class to a local class. (sdxattr -C command)
2-4) Remove the object configuration information. (sdxconfig Remove command)
2-5) For a shared disk in the cluster system, delete the migration source disk from the PRIMECLUSTER resource database. (cldeldevice command)
3) Remove the migration source storage from a server and connect the migration destination storage to a server.
4) Attach the disk label to the storage of migration destination on all the nodes of the cluster system. (format command)
5) Restore the GDS configuration to the migration destination storage.
5-1) For a shared disk in the cluster system, register the migration destination disk with the PRIMECLUSTER resource database. (clautoconfig command)
5-2) Convert the configuration files backed up in step 2-1) for the migration destination storage as below. (sdxconfig Convert command)
When copying to one migration destination case from one migration source case in step 1), remove disks from each mirror group so that only one disk will remain in each mirror group.
Change the physical disk names to names of the migration destination.
5-3) Restore object configuration information. (sdxconfig Restore command)
When copying data from each migration source case to each migration destination case, specify the -e skipsync option at the time of executing the sdxconfig Restore command.
5-4) For a shared disk in the cluster system, convert the classes restored in step 5-3) to shared classes. (sdxattr -C command)
6) Restart the services to use storage.
7) When copying to one migration destination case from one migration source case in step 1) and mirroring the migration destination storage, add a disk to the mirror group of the migration destination. (sdxdisk -M and sdxdisk -C commands)
See
For the method of using the sdxconfig command, see "D.13 sdxconfig - Object configuration operations" and "6.9 Backing Up and Restoring Object Configurations."
For the method of using the clautoconfig and cldeldevice commands, see the manual pages or "PRIMECLUSTER Cluster Foundation (CF) Configuration and Administration Guide."
