During a post-installation health check or right before initiating a Hardware Non-Disruptive Upgrade (HWNDU), verifying that frontend host multipathing is functioning correctly is arguably the most critical safety procedure an Implementation Engineer performs. If a host is single-pathed or improperly zoned, failing over a controller could trigger an All-Paths-Down (APD) event.
To validate that host traffic is redundant and healthy, the engineer must execute the purehost monitor -- balance command within the Purity CLI.
This powerful diagnostic tool instructs the FlashArray to actively sample incoming read and write SCSI
/NVMe commands across all connected host initiators for a brief period (defaulting to a 5-second interval). It calculates the exact distribution of traffic and displays it as a percentage ("I/O Relative to Max") for each initiator-target path. If a host has Round Robin MPIO configured properly, the output will show traffic flowing evenly across all active paths (e.g., all paths hovering near 90-100% relative to the busiest path). If the output reveals that 100% of the I/O is pinned to a single path while others sit at 0%, the engineer immediately knows there is a zoning or host-side MPIO misconfiguration to address.