A server with two HBAs is connected to two identical fabrics and configured as active-active, load balanced with automatic failover. Under normal conditions, traffic is moving across both fabrics equa

The problem indicates that Fabric A is intentionally offline for maintenance, and the issue lies with Fabric B where the switch port connected to HBA B shows no light, and the server reports all paths lost. Since the HBA is the server's interface to the SAN, checking HBA B's status on the server is the most logical first step to determine if it is operational or has a physical fault.

Removing the server from production is a drastic measure and not the first troubleshooting step; the immediate goal is to identify and resolve the connectivity problem.

Replacing the cable on Fabric A is irrelevant because Fabric A is purposely offline for maintenance, and the reported issue is with Fabric B's connection.

Replacing the switch in Fabric B is premature; a lack of light on a single port could indicate a problem with the cable, HBA, or port configuration, and replacing an entire switch should be considered only after more focused troubleshooting.