- If you migrate a logical volume used as a namespace of an NVM subsystem when using NVMe-oF connection, use CCI to perform the operation. For details about the requirements of Volume Migration operations using CCI, see the Command Control Interface User and Reference Guide.
- When an ESE-VOL is used as a source or target volume, you can perform Volume Migration only when both of the following conditions are met:
- Both the source and target volumes are ESE-VOLs.
- When SOM 1256 is not set to ON, both the pool ID and CU number for the source and target volumes are even numbers or odd numbers.
- The system administrator can analyze the usage rate of drives, processors, and access paths for migrating volumes. However, you should keep in mind that Volume Migration operations can improve performance in one area while decreasing performance in another. Consider the following scenario:
- Parity Group A has an average usage value of 20%
- Parity Group B has an average usage value of 90%
Through Volume Migration, it is estimated that if one volume from Parity Group B is migrated to Parity Group A, the usage values of both parity groups become 55%. In this scenario, the I/O response time for Parity Group B will most likely decrease, while the I/O response time for Parity Group A will most likely increase. Overall, the throughput might increase or decrease.
Volume Migration must only be performed when you can expect a large improvement in storage system performance. Volume Migration operations may not provide significant improvement for cases in which parity group or volume usage varies only slightly, or for cases in which the overall MP usage is relatively high.
Note:A Dynamic Provisioning virtual volumes can be migrated to the following volumes:
- External volume
- Dynamic Provisioning virtual volume
When you migrate a DP-VOL to an external volume, you cannot estimate the usage rate of the volume after migration because you cannot measure the usage rate.
When an error condition exists in the storage system, resource usage can increase or become unbalanced. Do not use data collected during an error condition as the basis for planning Volume Migration operations.
- During firmware replacement, do not use a function that will not be supported after firmware replacement.
- If you run a command for migrating volumes or waiting for events after you perform delete operation of a Volume Migration pair, take time (10 seconds recommended) until deletion of the pair is complete.
If you perform the operation without waiting for a certain amount of time, the operation might end abnormally. To check whether deleting the Volume Migration pair is complete, verify that the VOL_ATTR displayed by the raidcom get ldev command is not ALUN.
- If you migrate a volume or cancel migration of a volume by using CCI, the command might be rejected with [EX_CMDRJE] displayed, depending on the conditions of the storage system.
- When you migrate a volume, the MP unit assigned to the LDEV used as the target volume is reassigned to the MP unit assigned to the source volume.
- When you start migrating a volume, the MP unit assigned to the target volume is changed. If the migration is complete or interrupted, the volume migration pair is deleted, and then the MP unit that was assigned to the target volume before migration is reassigned to the target volume.
However, in either of the following cases, the former MP unit is not reassigned to the target volume:
- A user changes the assignment of the MP unit to the source or target volume.
- When the volume migration is complete or interrupted and the Volume Migration pair is deleted, the cache write pending rate of the MP unit assigned to the target volume is 50% or more, or the cache write pending rate of the former MP unit is 50% or more.
Also, when the former MP unit is removed, another MP unit is assigned to the target volume.