Planning for recovery group space and scale up

IBM Storage Scale Erasure Code Edition manages all disks in the recovery group. The recovery group puts different type of disks into the declustered array. The declustered array space is the base for creating vdisks.

IBM Storage Scale Erasure Code Edition supports the scale-up of declustered array space by adding new disk in the server. When you plan for IBM Storage Scale Erasure Code Edition recovery group, it is needed to consider the initial disk type and the disk number on each server node. The method for expanding disk slots needs to be planned.

When you add new disks into the declustered array of recovery group, ensure to meet the following requirements:
  • The new disks that are inserted into the empty slots of the server for expanding declustered array space must be the same type of the existing disks in the declustered array.
  • All nodes in the recovery group must have the same number of new disks that are added for the declustered array.
  • New disks might be added into multiple declustered arrays at the same time.
  • It is recommended to add the same number of disks as the existing disks in the declustered array to double the declustered array space.

    For example, start from 1/2 populated (20 disks per node) for high storage density servers (40 disk slots per node), then add another 1/2 populated in scale up.

    You can start from 1/4 populated (10 disks per node), then add another 1/4 populated in each of the scale up until the server is fully populated.