- ONTAP 9.x
- Data ONTAP 8.x
- Data ONTAP completes reconstructs in the least impactful way by design, but still requires some system resources
- For systems close to a resource bottleneck either at CPU or the disk layer, this can lead to an increase in overall latency
- Potential mitigation options:
- Reduce the overall workload on the node if possible
- Add disks which will spread workload across more spindles to increase the I/O bandwidth.
- Move to SSDs which have more I/O bandwidth with decreased I/O latency.
Note: The process of rebuilding the data from parity from a failed drive is more impactful than performing a sick disk copy