PowerStore: Direct Attach Fibre Channel Host Initiator not Detected in PowerStoreOS 2.0

Administrators may notice a combination of the following symptoms:

• The initiator WWPN is not seen when adding an FC host initiator on PowerStore Manager for the first time.
• The initiator WWPN is no longer seen and recabling to the original ports appears to resolve the issue when connecting cables from the host to other FC ports on the PowerStore.
• The host HBA LED status indicates that no link is present.
• The operating system host HBA driver reports no link present.

Restarting each node within an appliance in sequence while the hosts are online and connected allows the initiators to be seen in PowerStore Manager.

Wait for the first node you rebooted to be online and initiator that is detected before restarting the other node in the appliance.

For instructions, see Rebooting procedures for a PowerStore node within the Dell PowerStore Power Down and Reboot Procedures Guide.

Dell has confirmed this issue is present in all versions of PowerStoreOS 2.0.x.

To address this issue, upgrade to PowerStoreOS 2.1.0.0 or later. There are several fixes in the release relating to the PowerStore FC driver and firmware including MDT-324698. The 2.1.0.0 release notes are linked in KB article 175213, PowerStore: PowerStoreOS Matrix

For more information about PowerStore updates, see the following KB articles:

• KB 182455: PowerStore: How to obtain PowerStoreOS software update packages.
• KB 183630: PowerStore: How to prepare for a PowerStore Non-Disruptive Upgrade (NDU).

If the PowerStoreOS update does not resolve the issue, consider that there can be other potential causes that show the same symptoms. Check the following items:

• Links speeds of 32 Gbps, 16 Gbps, and 8 Gbps are supported. Ensure that the host HBA and SFP transceivers are able to supporting these speeds, autonegotiate setting is fine.
• Incorrect specification, badly secured, contaminated, or faulty SFP transceivers check that the SFPs used are appropriate for FC transmission.
• Damaged or contaminated fiber optic cables, dust, or scratches on fiber optic LC connectors are a common cause of poor performance and CRC errors. See KB article 58192: Connectrix: Connection issue or port problem, - Cleaning Cables
• Host HBA firmware requires updating. Follow the HBA vendors guidelines to update the firmware.
• Incorrect Host HBA adapter settings, in general, set the HBA to default settings using the HBA BIOS or HBA control software.
• The host HBA device driver requires updating. Follow the HBA or Operating System vendors instructions to update the device driver.

Addendum: 
In October 2021, VMware released ESX 7 update 3, and after the release, Dell technical support observed direct attach hosts may lose access to the PowerStore target volumes when running some Broadcom Emulex HBA models that now enable NVMe support.

Per ESXi 7 update 3 release notes:

Merging the lpfc and brcmnvmefc drivers: Starting with vSphere 7.0 Update 3, the brcmnvmefc driver is no longer available. The NVMe over Fibre Channel functionality previously delivered with the brcmnvmefc driver is now in the lpfc driver.

Workaround for Emulex HBAs on ESXi 7 update3:
PowerStoreOS 2.0 does not support direct attach NVMe (N2N) over fiber channel. An FC switch must be used if FC-NVMe protocol support is required. For direct-attached FC configurations, Dell Technologies recommends disabling FC-NVMe on the lpfc driver by changing the following driver parameter lpfc_enable_fc4_type.

• To enable only SCSI support protocol, set lpfc_enable_fc4_type=1.
• To enable both SCSI and NVMe support protocol, set lpfc_enable_fc4_type=3.

1. If this is a production system, put the ESX host in maintenance mode as a host reboot is required.
2. Enable SSH access to the ESX host and connect to the ESXi host as root.
3. Use the following command to disable FC-NVMe support on the lpfc driver: 

esxcli system module parameters set -m lpfc -p lpfc_enable_fc4_type=1

4. Reboot the ESXi host to complete the change.