X1151A ports not coming online with optical cables

• X1151A Dual 100G Ethernet Controller IONIC card
• X65405 QSFP transceivers 
• X66200-<n> MPO optical cables

• The e3a/e3b ports take 15-20 minutes to come online when optical cables are used:

Wed Dec 13 11:56:02 +0800 [Node-02: nblade1: netif.linkInfo:info]: Ethernet adapter e3a: Link Resetting.
Wed Dec 13 11:56:05 +0800 [Node-02: nblade1: netif.linkInfo:info]: Ethernet adapter e3b: Link Resetting.
                           
Wed Dec 13 12:14:32 +0800 [Node-02: vifmgr: vifmgr.portup:notice]: A link up event was received on node Node-02, port e3b.
Wed Dec 13 12:14:32 +0800 [Node-02: vifmgr: vifmgr.portup:notice]: A link up event was received on node Node-02, port e3a.

• On some occasions, this can manifest as a "flapping" port issue, rather than ports not coming up.
• The automated ONTAP upgrade may be paused with the below error if this issue is encountered:

[Node-01: upgrademgr: upgrademgr.update.pausedErr:error]: The automated update of the cluster has been paused due to the following reason: Error: {Update timeout occured in State[Giveback]-- Tasks Pending in nodes  Task Name: do-giveback-job}, Action: {Ensure that nodes are healthy and try resuming the update operation after some time}
Wed Dec 13 12:10:51 +0800 [Node-01: notifyd: callhome.andu.pausederr:alert]: params: {'subject': 'AUTOMATED NDU PAUSED ON NODE: MUM', 'epoch': '6ddXXX0-bXX3-4XXf-b7c1-aXXXXXXXXXX67'}

• Manually setting the switch port speed to 100G does not resolve the issue.
• From switch CLI, ports may report errdisable due to flapping in the output of show interface brief command.
• The e3a/e3b ports come online immediately when copper cables are used.