IBM Storage Scale Erasure Code Edition limitations

This topic describes the known limitations of IBM Storage Scale Erasure Code Edition.

General limitations
  • The installation toolkit does not support installing mixed Elastic Storage Server (ESS) and IBM Storage Scale Erasure Code Edition in the same cluster. If you need this configuration, see Incorporating IBM Storage Scale Erasure Code Edition in an IBM Elastic Storage Server (ESS) cluster and Incorporating IBM Elastic Storage Server (ESS) building block in an IBM Storage Scale Erasure Code Edition cluster. For information about installation toolkit limitations in an IBM Storage Scale Erasure Code Edition environment, see Installation toolkit-related limitations.
  • Rack-level fault tolerance is not fully supported. This fault tolerance can be achieved by spreading servers evenly between the racks. However, it is recommended that there must be no more than two storage servers per rack for an N+3P erasure code, and one storage server per rack for N+2P Erasure Code.
  • When you use NVMe drives that are hot-swappable, you must create an EDF file to specify the drives and their slot numbers. For more information, see Mapping NVMe disk slot location. However, users are responsible for defining the correct mapping.
  • Users must check the physical SAS disk location with IBM Storage Scale Erasure Code Edition commands, and remap the disk slot location when the command output mismatches with the physical locations. For more information, see Mapping LMR disk location. However, users are responsible for checking and setting the correct slot locations.
  • Configuration with single server per compute chassis is supported. Configurations with 2 or more servers that are packaged together in the same physical unit are not supported.
  • Attaching storage that is not associated with IBM Storage Scale Erasure Code Edition (for example, SAN-attached disks) to recovery group server nodes is not supported.

Configuration limitations

All limitations of the IBM Storage Scale apply, notably:
  • All nodes in the RG must be configured the same (memory, drives, CPU, and network).
  • Supported erasure codes are 4+2P, 4+3P, 8+2P, 8+3P, 3WayReplication, and 4WayReplication.
  • All nodes/HBAs/drives in an RG must have consistent firmware levels, and be at a level that is supported by the hardware provided. For more information, see Hardware checklist.

For information about the hardware requirements, see Minimum hardware requirements and precheck.