Home > Storage > Unity XT > Storage Admin > Dell Unity Dynamic Pools > Wear balancing within a dynamic pool
Within Dell Unity storage systems, managing wear within flash drives is crucial to extending the lifespan of the drives. Due to data access patterns and end user applications, certain areas within a LUN, file system, VMware datastore, or other storage resource may have hotspots of activity. If this activity consists of writes, certain areas of particular drives may be targeted, which may increase the wear of the drives. Internally, all flash drives used within the storage systems have spare cells to manage wear. Dell Unity systems include additional wear mitigation techniques to help extend the lifespan of flash drives.
As outlined in the previous sections, dynamic pools have been designed to spread out workloads and wear as evenly as possible. First, dynamic pools do not require hot spares, like traditional pools. These drives are engaged and utilized within the pool. When creating a dynamic pool, user data is spread across all drives as much as possible. When expanding a dynamic pool, user data again is spread out as evenly as possible to distribute customer workloads and wear as much as possible.
Wear is also managed and mitigated across drive partnership groups within a dynamic pool. If a certain drive partnership group is wearing faster than another, 256 MB slices can be moved from one drive partnership group to another. This migration of data then moves the workload targeted at that data within the 256 MB slice to a new location. Wear balancing is not performed within a drive partnership group, since moving data within a drive partnership group actually causes more wear, as the data is being moved within the same set of drives.
If a drive is close to the end of its usable life based on wear, a proactive copy operation may be started by the system to replace the worn drive. This proactive copy does not copy drive extents within the drive to spare space extents within the pool, but rather to a free drive within the system if available. After the proactive copy operation completes, the drive is failed, an alert is generated, a call home is initiated, and the drive should be replaced. The proactive copy to spare space extents within the drive partnership group is not done because it would only increase the wear on the drives in the group, which may have the same amount of wear as the drive being replaced. If a valid spare is not available, no proactive copy operation is completed.
In Dell UnityOS version 4.2 and later, system alert messages are sent at 180, 90, and 30 days before a flash drive is expected to reach 100% wear. At the 60-day mark, Dell support is notified by a call home that the drive may reach 100% wear, and a replacement is required. These alerts are generated for each individual drive.