Questions to Determine

• What are the exact firmware versions on updated vs. un-updated hosts?
• Did the firmware update change SCSI behavior, ALUA handling, or FPIN (Fabric Performance Impact Notification) support?
• Are the updated hosts sending fabric notifications that confuse the older-firmware hosts?
• Is the FC fabric (Brocade/Cisco MDS) zoning configured so that old and new firmware hosts share the same target ports?
• Are the I/O errors correlated with specific SAN operations (snapshots, replication, failover tests)?
• Does the HBA firmware changelog mention breaking changes or required minimum firmware for mixed environments?
• Are multipath settings (multipathd) configured the same on all hosts?
• Is the lpfc driver version compatible with both firmware versions?
• Can the remaining 8 servers be updated immediately to resolve the mismatch?
• Are there any FC fabric-level errors visible in switch logs?