FC Initiator Adapter offline in MetroCluster

Applies to

• ONTAP 9
• Fabric MetroCluster
• FAS/AFF Systems

Issue

• FC Adapter is offline in Fabric MetroCluster configuration:

cluster1::> system node run -node node1 -command sysconfig -a

slot 1: FC Host Adapter 1a (Emulex XE501 rev. 1, <OFFLINE (hard)>)
                 Board name:        111-03648
                 Serial Number:     FP12345678
                 Firmware rev:      11.4.277.0
                 Host Port Id:      N/A
                 FC Node Name:      2:000:00109b:ea9206
                 FC Port Name:      1:000:00109b:ea9206
                 Link Data Rate:    N/A

• The following alert is reported in the event logs:

[Node-01: nchmd: hm.alert.raised:alert]: Alert Id = StorageFCAdapterFault_Alert , Alerting Resource = 10000010XXXX206 raised by monitor node-connect

• The MetroCluster node health is degraded because of that:

cluster1::> system health subsystem show
Subsystem         Health
----------------- ------------------
SAS-connect       ok
Environment       ok
Memory            ok
Service-Processor ok
Switch-Health     ok
CIFS-NDO          ok
Motherboard       ok
IO                ok
MetroCluster      ok
MetroCluster_Node degraded
FHM-Switch        ok
FHM-Bridge        ok
SAS-connect_Cluster ok
13 entries were displayed.

• The node configuration may change to "Multi-Path" instead of "Quad-Path" configuration.

  Fri Jun 07 01:10:23 +0200 [ClusterA-01: start_asup_collector_thread: shelf.config.tompha:info]: System has transitioned to multi-path HA attached storage

• From the switch logs, it is observed that the switch port connected to the affected FC port reports No_Sync or No_Light:

MCC-sw2:admin> switchshow
switchName: MCC-sw2
switchType:  181.0
switchState: Online
switchMode:  Native

Index Port Address  Media Speed   State       Proto
==================================================
   8   8   070800   id    32G     No_Sync     FC

• The cable between the node FC port and the switch port is reseated.
• One or more disks reporting fewer than 8 paths (per cluster)/4 paths (per node):

  ::> storage show disk -fields path-count -path-count !8 -type !LUN