Usage level rebalancing among parity groups

If multiple parity groups that contain LDEVs used as pool volumes exist, pages are allocated to rebalance the page usage rate among parity groups by reallocating pages. If a parity group contains multiple pool-VOLs, rebalancing is performed as if each parity group were a single pool-VOL. Therefore, after rebalancing, the usage rate among pool-VOLs in each parity group might not be balanced. If all data in the page to be migrated to rebalance is zero, the zero data page is reclaimed, and then the corresponding page is released.

When the pool capacity is expanded, data is moved to the added space on a per-page basis. When data is migrated, the usage is rebalanced across the parity groups in which the pool-VOLs are defined.

This operation might change the host I/O performance because the existing data is migrated.

If you do not want to automate rebalancing of the usage levels of pool-VOLs, contact customer support for assistance.

The page usage rate among parity groups is automatically rebalanced under the following conditions:

• Pool capacity expansion
• Zero data page reclamation

Rebalancing automatically stops if the cache memory is not redundant.

The pool-VOLs contained in a pool can be added or deleted. Removing a pool-VOL does not delete the pool or any related V-VOLs. You must delete all V-VOLs related to the pool before the pool can be deleted. When the pool is deleted, all data in the pool is also deleted.

The total throughputs of page movements due to rebalancing and pool shrinking are limited on a per-storage system basis according to the average processor utilization on the storage system to mitigate the impact on I/O performance. The following table shows the total throughput limits of page movements according to the average processor utilization.