Migrating data with a smaller amount of free capacity

If the free capacity in the SAN Volume Controller cluster is smaller than the capacity of the data that is imported, you can still migrate data.

Scenario

Under one scenario, you could have one managed disk (MDisk) in the destination MDisk group. You add image mode logical units from an array on the storage system and migrate these logical units to the destination MDisk group. The logical units are then striped across the one managed-mode disk. Next, you add another logical unit to the destination MDisk group. The MDisk now contains two managed-mode disks, but all of the data is on the first managed-mode disk. As a result, some of the data must be migrated from the overloaded managed-mode disks to the underused managed-mode disks.

Attention: The migration causes an uneven distribution of data across the MDisks in the MDisk group. The impact of this depends on the number of MDisks that are initially in the MDisks group and how many of these have free capacity.

This task might require subsequent migration of data within the MDisk group in order to balance the distribution of data across the MDisks in the group.

Perform the following steps to migrate data:

Select an MDisk group that contains enough free capacity to migrate all of the logical disks on the first array that you want to migrate to the cluster.
Create an empty MDisk group that can temporarily contain the data that is imported.
Stop all I/O operations to the logical disks that you want to migrate first, and unmap these disks from their hosts.
Perform the following steps to create an image mode virtual disk (VDisk) from the first unmanaged-mode MDisk that contains the data that you want to import:
1. Map one logical disk from the storage system to the SAN Volume Controller ports.
2. Issue the svctask detectmdisk command-line interface (CLI) command on the cluster or use the SAN Volume Controller Console to discover MDisks. The new unmanaged-mode MDisk that is found corresponds with the logical disk that was mapped in the previous step.
3. Create an image mode VDisk from this unmanaged-mode MDisk and assign it to use the empty MDisk group just created.
4. Repeat steps 4.a through 4.c for all logical disks.
You can restart host applications by mapping the image mode VDisks to the hosts and restarting the hosts, using the SAN Volume Controller. Alternatively, you can wait until the migration is complete to map the VDisks to the hosts.
Perform the following steps to migrate the data into the MDisk groups that you created in step 1:
1. Select the first image mode VDisk that you want to migrate.
2. Using the svctask migratevdisk or svctask migratetoimage command, migrate this VDisk from its current MDisk group to one of the MDisk groups that you created in step 2. This migrates all of the data from the logical disk into the new free space.
3. Run the svcinfo lsmigrate command to verify that the migration has completed.
4. Select the next image mode VDisk and repeat the previous step when the migration completes.
Perform the following steps to reconfigure the RAID array that contains the logical disks and add it to the MDisk group that you selected in step 1:
1. Remove the MDisks from the temporary MDisk group.
2. At the storage system, unmap the logical disks that have been migrated from the SAN Volume Controller cluster and delete them from the array (if more than one existed).
3. Create a single logical disk that uses the entire array capacity.
4. Map this new logical disk to the SAN Volume Controller ports.
5. Issue the svctask detectmdisk CLI command on the cluster or use the SAN Volume Controller Console to discover MDisks. The new managed-mode MDisk that is found corresponds with the new logical disk that you created.
6. Add this managed-mode MDisk to the MDisk group that you selected in step 1.
Repeat steps 3 through 7 for the next array.