Answer
In practical IBM i terms, HA is usually built on journaling and replication to a synchronized or near-synchronized secondary LPAR, often within the same data center or a short distance away, purpose-built to shrink the outage window for hardware failure, PTF problems, or human error. DR assumes that secondary path might not be available either, because the event took out the building, the region, or the staff who would normally execute the swap. That is why a serious DR design specifies an alternate location genuinely independent of the primary site, along with the access, communications, and staffing plan to operate from it.
Buyers evaluating either category should ask vendors to be specific about which problem their product actually solves. A replication tool that keeps data current is doing real HA or DR groundwork, but the declaration process, the validated startup sequence, and the proof that users and interfaces can reach the recovered environment are usually built through separate runbook discipline layered on top of the technology. Treating the software purchase as the whole solution, rather than one component of a broader recovery design, is the most common reason DR investments underperform when they are finally needed.