Hosts

A host system is a computer that is connected to the system through supported connection protocols.

The system supports host connections through Fibre Channel, NVM Express (NVMe) over Fibre Channel (FC-NVMe), NVM Express (NVMe) over RDMA, NVM Express (NVMe) over TCP, or an IP network. See the System Overview page on the relevant Hardware Guide to understand which supported adapters can be used for each host attach protocol.

A host object is a logical object that represents a list of worldwide port names (WWPNs), NVMe qualified names (NQNs), or iSCSI names that identify the interfaces that the host system uses to communicate with the system. Fibre Channel connections or Fibre Channel over Ethernet use WWPNs to identify host interfaces to the system. iSCSI names can be either iSCSI qualified names (IQNs) or extended unique identifiers (EUIs). NQNs are used to identify hosts that use NVMe connections.

A typical configuration has one host object for each host system that is attached to the system. If a cluster of hosts accesses the same storage, you can add the hosts into a Host Cluster object to make configuration simpler.

When you create a new Fibre Channel host object, the system presents a list of candidate WWPNs that have been logged into the system but are not yet configured in host objects. You must enter the WWPNs for such hosts manually or use the -fdminame attribute in the mkhost command to create a host object. There is no equivalent list of candidate IQNs available when creating iSCSI hosts. All iSCSI host port IQNs must be entered manually.

Note: A new FC-NVMe host can be mapped to a user-defined Fibre Channel portset only. Pre-existing FC-NVMe host objects which are part of the default Fibre Channel portset must be moved from the default portset to user-defined Fibre Channel portset. For specific information about FC-NVMe enhanced scaling and system requirements, see Configuration limits.

The system can detect only WWPNs that have connected to the system through the Fibre Channel network or through any IP network. Some Fibre Channel HBA device drivers do not leave ports logged in if no disks are detected on the fabric or IP network. The configuration interface provides a method to manually type the port names.

A WWPN or iSCSI name can be added to only one host object or Host Cluster. Do not include WWPNs or IQNs that are assigned to nodes to host objects.

The system does not automatically present volumes to the host system. You must map each volume to a particular host object or host group to enable the volume to be accessed through the WWPNs or iSCSI names that are associated with the host object or Host Cluster.

For Fibre Channel, FC-NVMe, NVMe over RDMA, and NVMe over TCP-attached hosts, the system reports the node login count, which is the number of nodes that can detect each WWPN or NQN. If the count is less than expected for the current configuration, you might have a connectivity problem. To view potential connectivity issues for Fiber Channel-attached hosts in the management GUI, select Settings > Network > Fibre Channel Connectivity or use the lsfabric command. To view connectivity issues for FC-NVMe, NVMe over RDMA, and NVMe over TCP-attached hosts in the management GUI, select Settings > Network > NVMe Connectivity or use the lsnvmefabric command.

All unconfigured host ports that are sorted by fdmi name get listed in Hosts > Hosts > Unconfigured hosts in the management GUI.

For Ethernet-attached hosts, the number of logged-in nodes refers to sessions that are created between hosts and nodes. The number of logged-in nodes can be greater than the current number of nodes on the system.

You can display the login information for each host in the management GUI to determine which ports for the logged-in nodes are connected to the Ethernet-attached hosts. To display this information, select Hosts > Hosts and right-click the host and select IP Login Information. This information can be used to detect connectivity issues between the system and hosts and to improve the configuration of hosts to improve performance.
To view login information in the command-line interface, use the lshostiplogin to list login session details for a specified host that is logged in to a configured Ethernet port on the system. This information can be used to detect connectivity issues between the system and hosts and to improve the configuration of hosts to improve performance.
You can also set a status policy and site status to determine how host statuses are displayed. For example, if you select Complete, the system displays an Online status only when all host ports are logged in to all nodes and hosts can access all mapped volumes. If a single host or node loses connectivity, the status for the host displays as Degraded. If you select Redundant for the Status Policy, an Online status is displayed when a sufficient number host ports to node logins exist to retain access to all its mapped volumes, even when a host port or single node is removed or lost.
If hosts are part of a multiple site configuration, each host needs to be assigned to a specific site and you can specify how host statuses are displayed for these hosts. In the management GUI, you can specify All for Site Status to indicate that host connectivity to the nodes on all sites determine the status of the host. If you select Local for Site Status, the host status relates only to connectivity to nodes on the local site. Host status for other nodes mapped to different sites is not included.

Hosts can be assigned to an ownership group. For details, see Ownership groups.

Ownership can be defined explicitly or it can be inherited from the user, user group, or from other parent resources, depending on the type of resource. Hosts that are not part of a Host Cluster can be owned if they are assigned an ownership group explicitly or by inheritance from the user creating them. The following rules apply to hosts that are not part of a Host Cluster that is defined in ownership groups:

If a host is not part of a Host Cluster, an ownership group can be associated with the host when the host object is created or changed.

Hosts that are a part of a Host Cluster inherit ownership from the Host Cluster. You cannot change the ownership group of the host, but can change the ownership group of the Host Cluster. The following rules apply to hosts that are in a Host Cluster:

The host inherits the ownership group of the Host Cluster that it belongs to.
If a host is removed from a Host Cluster within an ownership group, the host inherits the ownership group of the Host Cluster it used to belong to.
If a host is removed from a Host Cluster not within an ownership group, the host inherits no ownership group.
Hosts can be added to a Host Cluster if the host and Host Cluster have the same ownership group.
Changing the ownership group of a Host Cluster automatically changes the ownership group of all the hosts inside the Host Cluster.

Multiple target ports

When you create a host mapping to a host, the host ports that are associated with the host object can view the LUN that represents the volume on many ports. Nodes follow the American National Standards Institute (ANSI) Fibre Channel (FC) standards for SCSI LUs that are accessed through multiple node ports. All nodes within a single I/O group present a consistent set of SCSI LUs across all ports on those nodes.