Cisco MDS 9000 Series IP Services Configuration Guide, Release 8.x

Overview of iSCSI

Cisco MDS 9000 Family IP Storage (IPS) services extend the reach of Fibre Channel SANs by using open-standard, IP-based technology. The iSCSI feature consists of routing iSCSI requests and responses between iSCSI hosts in an IP network and Fibre Channel storage devices in the Fibre Channel SAN that are accessible from any Fibre Channel interface of the Cisco MDS 9000 Family switch. Using the iSCSI protocol, the iSCSI driver allows an iSCSI host to transport SCSI requests and responses over an IP network. To use the iSCSI feature, you must explicitly enable iSCSI on the required switches in the fabric.

Note The iSCSI feature is not supported on the Cisco Fabric Switch for HP c-Class Bladesystem and Cisco Fabric Switch for IBM BladeCenter.

The iSCSI feature consists of routing iSCSI requests and responses between iSCSI hosts in an IP network and Fibre Channel storage devices in the Fibre Channel SAN that are accessible from any Fibre Channel interface of the Cisco MDS 9000 Family switch (see Figure 4-1).

Figure 4-1 Transporting iSCSI Requests and Responses for Transparent iSCSI Routing

Each iSCSI host that requires access to storage through the Fibre Channel module with IPS ports or MPS-14/2 module needs to have a compatible iSCSI driver installed. Using the iSCSI protocol, the iSCSI driver allows an iSCSI host to transport SCSI requests and responses over an IP network. From the host operating system perspective, the iSCSI driver appears to be an SCSI transport driver similar to a Fibre Channel driver in the host.

The Fibre Channel module with IPS ports or MPS-14/2 module provides transparent SCSI routing. IP hosts using the iSCSI protocol can transparently access targets on the Fibre Channel network. It (see Figure 4-1) provides an example of a typical configuration of iSCSI hosts connected to an Fibre Channel module with IPS ports or MPS-14/2 module through the IP network access Fibre Channel storage on the Fibre Channel SAN.

The Fibre Channel module with IPS ports or MPS-14/2 module create a separate iSCSI SAN view and Fibre Channel SAN view. For the iSCSI SAN view, the Fibre Channel module with IPS ports or MPS-14/2 module creates iSCSI virtual targets and then maps them to physical Fibre Channel targets available in the Fibre Channel SAN. They present the Fibre Channel targets to IP hosts as if the physical iSCSI targets were attached to the IP network (see Figure 4-2).

Figure 4-2 iSCSI SAN View—iSCSI Virtual Targets

For the Fibre Channel SAN view, the Fibre Channel module with IPS ports or MPS-14/2 module presents iSCSI hosts as a virtual Fibre Channel host. The storage devices communicate with the virtual Fibre Channel host similar to communications performed with real Fibre Channel hosts (see Figure 4-3).

Figure 4-3 Fibre Channel SAN View—iSCSHI Host as an HBA

The Fibre Channel module with IPS ports or MPS-14/2 modules transparently map the command between the iSCSI virtual target and the virtual Fibre Channel host (see Figure 4-4).

Figure 4-4 iSCSI to FCP (Fibre Channel) Routing

Routing SCSI from the IP host to the Fibre Channel storage device consists of the following main actions:

• The iSCSI requests and responses are transported over an IP network between the hosts and the Fibre Channel module with IPS ports or MPS-14/2 module.
• The SCSI requests and responses are routed between the hosts on an IP network and the Fibre Channel storage device (converting iSCSI to FCP and vice versa). The Fibre Channel module with IPS ports or MPS-14/2 module performs this conversion and routing.
• The FCP requests or responses are transported between the Fibre Channel module with IPS ports or MPS-14/2 module and the Fibre Channel storage devices.

Note FCP (the Fibre Channel equivalent of iSCSI) carries SCSI commands over a Fibre Channel SAN.
Refer to the IETF standards for IP storage at http://www.ietf.org for information on the iSCSI protocol.

iSCSI Configuration Limits

iSCSI configuration has the following limits:

• The maximum number of iSCSI and iSLB initiators supported in a fabric is 2000.
• The maximum number of iSCSI and iSLB initiators supported is 200 per port.
• The maximum number of iSCSI and iSLB sessions supported by an IPS port in either transparent or proxy initiator mode is 500.
• The maximum number of iSCSI and iSLB session support by switch is 5000.
• The maximum number of iSCSI and iSLB targets supported in a fabric is 6000.

Configuring iSCSI

This section describes how to configure iSCSI on the Cisco MDS 9000 Family switches.

This section includes the following sections:

• Enabling iSCSI
• Creating iSCSI Interfaces
• Using the iSCSI Wizard
• Presenting Fibre Channel Targets as iSCSI Targets
• Presenting iSCSI Hosts as Virtual Fibre Channel Hosts
• iSCSI Access Control
• iSCSI Session Authentication
• iSCSI Immediate Data and Unsolicited Data Features
• iSCSI Interface Advanced Features

Enabling iSCSI

To use the iSCSI feature, you must explicitly enable iSCSI on the required switches in the fabric. Alternatively, you can enable or disable the iSCSI feature directly on the required modules using Fabric Manager or Device Manager. By default, this feature is disabled in all switches in the Cisco MDS 9000 Family.

To enable iSCSI on any participating switch, follow these steps:

Caution When you disable this feature, all related configurations are automatically discarded.

---

To enable iSCSI on any switch using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane (see Figure 4-5).

Figure 4-5 iSCSI Tables in Fabric Manager

The Control tab is the default tab. You see the iSCSI enable status for all switches in the fabric that contain IPS ports.

Step 2 Choose enable from the Command column for each switch that you want to enable iSCSI on.

Step 3 Click the Apply Changes icon to save these changes.

To enable iSCSI on a module using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane.

Step 2 Click the Module Control tab.

You see the Module Control dialog box in the information pane (see Figure 4-6).

Figure 4-6 Module Control Dialog Box

Step 3 Check the Mode Admin check box to enable iSCSI for a specified port on the selected module.

Step 4 Click the Apply Changes icon to save these changes.

To enable iSCSI on a module using Device Manager, follow these steps:

Step 1 Choose IP > iSCSI

You see the iSCSI table (see Figure 4-7).

Figure 4-7 iSCSI Table

Step 2 Check the Mode Admin check box to enable iSCSI for the specified port on the selected module.

Step 3 Click Apply to save these changes.

Creating iSCSI Interfaces

Each physical Gigabit Ethernet interface on an Fibre Channel module with IPS ports, MPS-14/2 module or 1/10Gbps IPStorage port on a Cisco MDS 9250i Multiservice Fabric Switch can be used to translate and route iSCSI requests to Fibre Channel targets and responses in the opposite direction. To enable this capability, the corresponding iSCSI interface must be in an enabled state.

To enable iSCSI interfaces, follow these steps:

Step 1 Enable the required Gigabit Ethernet interface.

Step 2 Create the required iSCSI interface and enable the interface.

Note Use the tcp maximum-bandwidth-kbps and tcp maximum-bandwidth-mbps commands to configure the iSCSI speed and the switchport speed command to set the Physical IPStorage ports to 1Gbps or 10Gbps speed. The Cisco MDS switches do not limit the configuration of the iSCSI tcp maximum-bandwidth-kbps and maximum-bandwidth-mbps based on the speed of the underlying physical Gigabit Ethernet or IPStorage ports. Consequently, it is possible to configure iSCSI tcp maximum-bandwidth-kbps and tcp maximum-bandwidth-mbps commands to the equivalent of 10Gbps on a physical IPStorage port that is running at a 1Gbps speed. When configuring the tcp maximum bandwidth, ensure that it does not exceed the maximum speed of the physical IPStorage port.

Using the iSCSI Wizard

To use the iSCSI wizard in Fabric Manager, follow these steps:

Step 1 Click the iSCSI Setup Wizard icon.

You see the iSCSI Wizard Configure Initiator dialog box (see Figure 4-8).

Figure 4-8 iSCSI Wizard Configure Initiator Dialog Box

Step 2 Select an existing iSCSI initiator or add the iSCSI node name or IP address for a new iSCSI initiator.

Step 3 Select the switch for this iSCSI initiator if you are adding a new iSCSI initiator and click Next.

You see the iSCSI Wizard Select Targets dialog box (see Figure 4-9).

Figure 4-9 iSCSI Wizard Select Targets Dialog Box

Step 4 Select the VSAN and targets to associate with this iSCSI initiator and click Next.

Note The iSCSI wizard turns on the Dynamic Import FC Targets feature.

You see the iSCSI Wizard Select Zone dialog box (see Figure 4-10).

Figure 4-10 iSCSI Wizard Select Zone Dialog Box

Step 5 Set the zone name for this new iSCSI zone and check the ReadOnly check box if needed.

Step 6 Click Finish to create this iSCSI initiator.

If created, the target VSAN is added to the iSCSI host VSAN list.

Note iSCSI wizard automatically turns on the Dynamic FC target import.

Presenting Fibre Channel Targets as iSCSI Targets

The Fibre Channel module with IPS ports or MPS-14/2 module presents physical Fibre Channel targets as iSCSI virtual targets, allowing them to be accessed by iSCSI hosts. The module presents these targets in one of the two ways:

• Dynamic mapping—Automatically maps all the Fibre Channel target devices/ports as iSCSI devices. Use this mapping to create automatic iSCSI target names.
• Static mapping—Manually creates iSCSI target devices and maps them to the whole Fibre Channel target port or a subset of Fibre Channel LUNs. With this mapping, you must specify unique iSCSI target names.

Static mapping should be used when iSCSI hosts should be restricted to subsets of LUs in the Fibre Channel targets and/or iSCSI access control is needed (see the “iSCSI Access Control” section). Also, static mapping allows the configuration of transparent failover if the LUs of the Fibre Channel targets are reachable by redundant Fibre Channel ports (see the “Transparent Target Failover” section).

Note The Fibre Channel module with IPS ports or MPS-14/2 module does not import Fibre Channel targets to iSCSI by default. Either dynamic or static mapping must be configured before the Fibre Channel module with IPS ports or MPS-14/2 module makes Fibre Channel targets available to iSCSI initiators.

Dynamic Mapping

When you configure dynamic mapping the Fibre Channel module with IPS ports or MPS-14/2 module imports all Fibre Channel targets to the iSCSI domain and maps each physical Fibre Channel target port as one iSCSI target. That is, all LUs accessible through the physical storage target port are available as iSCSI LUs with the same LU number (LUN) as in the physical Fibre Channel target port.

The iSCSI target node name is created automatically using the iSCSI qualified name (IQN) format. The iSCSI qualified name is restricted to a maximum name length of 223 alphanumeric characters and a minimum length of 16 characters.

The Fibre Channel module with IPS ports or MPS-14/2 module creates an IQN formatted iSCSI target node name using the following conventions because the name must be unique in the SAN:

• IPS Gigabit Ethernet ports that are not part of a Virtual Router Redundancy Protocol (VRRP) group or port channel use this format:

• IPS ports that are part of a VRRP group use this format:

• Ports that are part of a port channel use this format:

Note If you have configured a switch name, then the switch name is used instead of the management IP address. If you have not configured a switch name, the management IP address is used.

With this convention, each IPS port in a Cisco MDS 9000 Family switch creates a unique iSCSI target node name for the same Fibre Channel target port in the SAN.

For example, if an iSCSI target was created for a Fibre Channel target port with pWWN 31:00:11:22:33:44:55:66 and that pWWN contains LUN 0, LUN 1, and LUN 2, those LUNs would become available to an IP host through the iSCSI target node name iqn.1987-05.com.cisco:05. MDS_switch_management_IP_address.01-01.3100112233445566 (see Figure 4-11).

Figure 4-11 Dynamic Target Mapping

Note Each iSCSI initiator may not have access to all targets depending on the configured access control mechanisms (see the “iSCSI Access Control” section).

To enable dynamic mapping of Fibre Channel targets into iSCSI, follow these steps:

To enable dynamic mapping of Fibre Channel targets into iSCSI using Device Manager, follow these steps:

Step 1 Choose IP > iSCSI.

You see the iSCSI configuration (see Figure 4-12).

Figure 4-12 iSCSI Configuration in Device Manager

Step 2 Click the Target tab to display a list of existing iSCSI targets (see Figure 4-13).

Figure 4-13 iSCSI Targets Tab

Step 3 Check the Dynamically Import FC Targets check box.

Step 4 Click Apply to save this change.

Static Mapping

You can manually (statically) create an iSCSI target by assigning a user-defined unique iSCSI node name to it. The iSCSI qualified name is restricted to a minimum length of 16 characters and a maximum of 223 characters. A statically mapped iSCSI target can either map the whole Fibre Channel target port (all LUNs in the target port mapped to the iSCSI target), or it can contain one or more LUs from a Fibre Channel target port (see Figure 4-14).

Figure 4-14 Statically Mapped iSCSI Targets

To create a static iSCSI virtual target for the entire Fibre Channel target port using Device Manager, follow these steps:

Step 1 Click IP > iSCSI.

You see the iSCSI configuration (see Figure 4-12).

Step 2 Click the Targets tab to display a list of existing iSCSI targets (see Figure 4-13).

Step 3 Click Create to create an iSCSI target.

You see the Create iSCSI Targets dialog box (See Figure 4-15).

Figure 4-15 Create iSCSI Targets Dialog Box

Step 4 Set the iSCSI target node name in the iSCSI Name field, in IQN format.

Step 5 Set the Port WWN field for the Fibre Channel target port you are mapping.

Step 6 Click the Select from List radio button and set the iSCSI initiator node names or IP addresses that you want this virtual iSCSI target to access, or click the All radio button to let the iSCSI target access all iSCSI initiators. Also see the “iSCSI Access Control” section.

Step 7 Click the Select from List radio button and check each interface you want to advertise the iSCSI targets on or click the All radio button to advertise all interfaces.

Step 8 Click Apply to save this change.

Tip An iSCSI target cannot contain more than one Fibre Channel target port. If you have already mapped the whole Fibre Channel target port, you cannot use the LUN mapping option.

Note See the “iSCSI-Based Access Control” section for more information on controlling access to statically mapped targets.

Advertising Static iSCSI Targets

You can limit the Gigabit Ethernet interfaces through which static iSCSI targets are advertised. By default iSCSI targets are advertised on all Gigabit Ethernet interfaces, subinterfaces, port channel interfaces, and port channel subinterfaces.

To configure a specific interface that should advertise the iSCSI virtual target using Device Manager, follow these steps:

Step 1 Select IP > iSCSI.

You see the iSCSI configuration (see Figure 4-12).

Step 2 Click the Targets tab to display a list of existing iSCSI targets (see Figure 4-13).

Step 3 Right-click the iSCSI target that you want to modify and click Edit Advertised.

You see the Advertised Interfaces dialog box.

Step 4 (Optional) Right-click an interface that you want to delete and click Delete.

Step 5 (Optional) Click Create to advertise on more interfaces.

You see the Create Advertised Interfaces dialog box.

To configure a specific interface that should advertise the iSCSI virtual target, follow these steps:

iSCSI Virtual Target Configuration Examples

This section provides three examples of iSCSI virtual target configurations.

Example 1

This example assigns the whole Fibre Channel target as an iSCSI virtual target. All LUNs that are part of the Fibre Channel target are available as part of the iSCSI target (see Figure 4-16).

Figure 4-16 Assigning iSCSI Node Names

Example 2

This example maps a subset of LUNs of a Fibre Channel target to three iSCSI virtual targets. Each iSCSI target only has one LUN (see Figure 4-17).

Figure 4-17 Mapping LUNs to an iSCSI Node Name

Example 3

This example maps three subsets of Fibre Channel LUN targets to three iSCSI virtual targets. Two iSCSI targets have one LUN and the third iSCSI target has two LUNs (see Figure 4-18).

Figure 4-18 Mapping LUNs to Multiple iSCSI Node Names

Presenting iSCSI Hosts as Virtual Fibre Channel Hosts

The Fibre Channel module with IPS ports or MPS-14/2 module connects to the Fibre Channel storage devices on behalf of the iSCSI host to send commands and transfer data to and from the storage devices. These modules use a virtual Fibre Channel N port to access the Fibre Channel storage devices on behalf of the iSCSI host. iSCSI hosts are identified by either iSCSI qualified name (IQN) or IP address.

Initiator Identification

iSCSI hosts can be identified by the Fibre Channel module with IPS ports or MPS-14/2 module using the following:

• iSCSI qualified name (IQN)

An iSCSI initiator is identified based on the iSCSI node name it provides in the iSCSI login. This mode can be useful if an iSCSI host has multiple IP addresses and you want to provide the same service independent of the IP address used by the host. An initiator with multiple IP addresses (multiple network interface cards—NICs) has one virtual N port on each IPS port to which it logs in.

• IP address

An iSCSI initiator is identified based on the IP address of the iSCSI host. This mode is useful if an iSCSI host has multiple IP addresses and you want to provide different service-based on the IP address used by the host. It is also easier to get the IP address of a host compared to getting the iSCSI node name. A virtual N port is created for each IP address it uses to log in to iSCSI targets. If the host using one IP address logs in to multiple IPS ports, each IPS port will create one virtual N port for that IP address.

You can configure the iSCSI initiator identification mode on each IPS port and all the iSCSI hosts terminating on the IPS port will be identified according to that configuration. The default mode is to identify the initiator by name.

To specify the initiator identification mode, follow these steps:

To specify the initiator identification mode using Fabric Manager, follow these steps:

Step 1 Choose Interfaces > FC Logical from the Physical Attributes pane.

You see the interfaces configuration in the Information pane.

Step 2 Click the iSCSI tab.

You see the iSCSI interfaces configuration.

Step 3 Right-click the Initiator ID Mode field for the iSCSI interface that you want to modify and select name or ipaddress from the drop-down menu.

Step 4 Click Apply Changes to save this change.

Initiator Presentation Modes

Two modes are available to present iSCSI hosts in the Fibre Channel fabric: transparent initiator mode and proxy initiator mode.

• In transparent initiator mode, each iSCSI host is presented as one virtual Fibre Channel host. The benefit of transparent mode is it allows a finer level of Fibre Channel access control configuration (similar to managing a “real” Fibre Channel host). Because of the one-to-one mapping from iSCSI to Fibre Channel, each host can have different zoning or LUN access control on the Fibre Channel storage device.
• In proxy initiator mode, there is only one virtual Fibre Channel host per one IPS port and all iSCSI hosts use that to access Fibre Channel targets. In a scenario where the Fibre Channel storage device requires explicit LUN access control for every host, the static configuration for each iSCSI initiator can be overwhelming. In this case, using the proxy initiator mode simplifies the configuration.

Caution Enabling proxy initiator mode of an iSCSI interface that is part of an iSLB VRRP group impacts load balancing on the interface. See the “Changing iSCSI Interface Parameters and the Impact on Load Balancing” section.

---

The Cisco MDS switches support the following iSCSI session limits:

• The maximum number of iSCSI sessions on a switch is 5000.
• The maximum number of iSCSI sessions per IPS port in transparent initiator mode is 500.
• The maximum number of iSCSI sessions per IPS port in proxy initiator mode is 500.
• The maximum number of concurrent sessions an IPS port can create is five (but the total number of sessions that can be supported is 500).

Note If more than five iSCSI sessions try to come up simultaneously on a port, the initiator receives a temporary error and later retries to create a session.

Transparent Initiator Mode

Each iSCSI host is presented as one virtual Fibre Channel host (that is, one Fibre Channel N port). The benefit of transparent mode is it allows a finer-level of Fibre Channel access control configuration. Because of the one-to-one mapping from iSCSI to Fibre Channel, each host can have different zoning or LUN access control on the Fibre Channel storage device.

When an iSCSI host connects to the Fibre Channel module with IPS ports or MPS-14/2 module, a virtual host N port (HBA port) is created for the host (see Figure 4-19). Every Fibre Channel N port requires a unique Node WWN and Port WWN.

Figure 4-19 Virtual Host HBA Port

After the virtual N port is created with the WWNs, a fabric login (FLOGI) is done through the virtual iSCSI interface of the IPS port. After the FLOGI is completed, the virtual N port is online in the Fibre Channel SAN and virtual N port is registered in the Fibre Channel name server. The Fibre Channel module with IPS ports or MPS-14/2 module registers the following entries in the Fibre Channel name server:

• IP address of the iSCSI host in the IP-address field on the name server
• IQN of the iSCSI host in the symbolic-node-name field of the name server
• SCSI_FCP in the FC-4 type field of the name server
• Initiator flag in the FC-4 feature of the name server
• Vendor-specific iSCSI GW flag in the FC-4 type field to identify the N-port device as an iSCSI gateway device in the name server.

When all the iSCSI sessions from the iSCSI host are terminated, the Fibre Channel module with IPS ports or MPS-14/2 modules perform an explicit Fabric logout (FLOGO) to remove the virtual N-port device from the Fibre Channel SAN (this indirectly de-registers the device from the Fibre Channel name server).

For every iSCSI session from the host to the iSCSI virtual target there is a corresponding Fibre Channel session to the real Fibre Channel target. There are three iSCSI hosts (see Figure 4-19), and all three of them connect to the same Fibre Channel target. There is one Fibre Channel session from each of the three virtual Fibre Channel hosts to the target.

iSCSI Initiator Idle Timeout

iSCSI initiator idle timeout specifies the time for which the virtual Fibre Channel N port is kept idle after the initiator logs out from its last iSCSI session. The default value for this timer is 300 seconds. This is useful to avoid N ports logging in to and logging off of the Fibre Channel SAN as transient failure occurs in the IP network. This helps reduce unnecessary RSCNs being generated in the Fibre Channel SAN.

To configure the initiator idle timeout, follow these steps:

To configure the initiator idle timeout using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane (see Figure 4-5).

Step 2 Click the Globals tab.

You see the iSCSI global configuration.

Step 3 Right-click on the InitiatorIdle Timeout field that you want to modify and enter the new timeout value.

Step 4 Click the Apply Changes icon to save these changes.

WWN Assignment for iSCSI Initiators

An iSCSI host is mapped to an N port’s WWNs by one of the following mechanisms:

• Dynamic mapping (default)
• Static mapping

Dynamic Mapping

With dynamic mapping, an iSCSI host is mapped to a dynamically generated port WWN (pWWN) and node WWN (nWWN). Each time the iSCSI host connects it might be mapped to a different WWN. Use this option if no access control is required on the Fibre Channel target device (because the target device access control is usually configured using the host WWN).

The WWNs are allocated from the MDS switch's WWN pool. The WWN mapping to the iSCSI host is maintained as long as the iSCSI host has at least one iSCSI session to the IPS port. When all iSCSI sessions from the host are terminated and the Fibre Channel module with IPS ports or MPS-14/2 module performs an FLOGO for the virtual N port of the host, the WWNs are released back to the switch's Fibre Channel WWN pool. These addresses are then available for assignment to other iSCSI hosts requiring access to the Fibre Channel Fabric.

The following are three dynamic initiator modes are supported:

• iSCSI—Dynamic initiators are treated as iSCSI initiators and can access dynamic virtual targets and configured iSCSI virtual targets.
• iSLB—Dynamic initiators are treated as iSLB initiators.
• Deny—Dynamic initiators are not allowed to log in to the MDS switch.

iSCSI dynamic mapping is the default mode of operation. This configuration is distributed using CFS.

Note Configuring dynamic initiator modes is supported only through the CLI, not through Device Manager or Fabric Manager.

To configure dynamic mapping (using the name option) for an iSCSI initiator, follow these steps:

Static Mapping

With static mapping, an iSCSI host is mapped to a specific pWWN and nWWN. This mapping is maintained in persistent storage and each time the iSCSI host connects, the same WWN mapping is used. This mode is required if you use access control on the target device.

You can implement static mapping in one of two ways:

• User assignment—You can specify your own unique WWN by providing them during the configuration process.
• System assignment—You can request that the switch provide a WWN from the switch’s Fibre Channel WWN pool and keep the mapping in its configuration.

Tip We recommend using the system-assign option. If you manually assign a WWN, you must ensure its uniqueness (see the Cisco Fabric Manager Fabric Configuration Guide and Cisco MDS 9000 Family NX-OS Fabric Configuration Guide for more information). You should not use any previously assigned WWNs.

To configure static mapping (using the name option) for an iSCSI initiator, follow these steps:

To configure static mapping for an iSCSI initiator using Device Manager, follow these steps:

Step 1 Select IP > iSCSI.

You see the iSCSI configuration (see Figure 4-12). The Initiators tab is the default.

Step 2 Click Create to create an iSCSI initiator.

You see the Create iSCSI Initiators dialog box (see Figure 4-20).

Figure 4-20 Create iSCSI Initiators Dialog Box

Step 3 Set the iSCSI node name or IP address and VSAN membership.

Step 4 In the Node WWN section, check the Persistent check box.

Step 5 Check the System Assigned check box if you want the switch to assign the nWWN or leave this unchecked and set the Static WWN field.

Step 6 In the Port WWN section, check the Persistent check box if you want to statically map pWWNs to the iSCSI initiator.

Step 7 If persistent, check the System Assigned check box and set the number of pWWNs to reserve for this iSCSI initiator if you want the switch to assign pWWNs. Alternately, you can leave this unchecked and set one or more pWWNs for this iSCSI initiator.

Step 8 (Optional) Set the AuthUser field if authentication is enabled. Also see the “iSCSI Session Authentication” section.

Step 9 Click Create to create this iSCSI initiator.

To configure static mapping (using the ip-address option) for an iSCSI initiator, follow these steps:

To assign the WWN for an iSCSI initiator, follow these steps:

Note If the system-assign option is used to configure WWNs for an iSCSI initiator, when the configuration is saved to an ASCII file the system-assigned WWNs are also saved. Subsequently if you perform a write erase, you must manually delete the WWN configuration from the ASCII file. Failing to do so can cause duplicate WWN assignments if the ASCII configuration file is reapplied on the switch.

Making the Dynamic iSCSI Initiator WWN Mapping Static

After a dynamic iSCSI initiator has already logged in, you may decide to permanently keep the automatically assigned nWWN/pWWN mapping so this initiator uses the same mapping the next time it logs in.

You can convert a dynamic iSCSI initiator to static iSCSI initiator and make its WWNs persistent (see the “Dynamic Mapping” section).

Note You cannot convert a dynamic iSCSI initiator to a static iSLB initiator or a dynamic iSLB initiator to a static iSCSI initiator.

Note Making the dynamic pWWNs static after the initiator is created is supported only through the CLI, not through Device Manager or Fabric Manager. In Fabric Manager or Device Manager, you must delete and then recreate this initiator to have the pWWNs static.

To permanently keep the automatically assigned nWWN/pWWN mapping, follow these steps:

Checking for WWN Conflicts

WWNs assigned to static iSCSI initiators by the system can be inadvertently returned to the system when an upgrade fails or you downgrade the system software (manually booting up an older Cisco MDS SAN-OS release without using the install all command). In these instances, the system can later assign those WWNs to other iSCSI initiators (dynamic or static) and cause conflicts.

You can address this problem by checking for and removing any configured WWNs that belong to the system whenever such scenarios occur.

To check for and remove WWN conflicts, follow these steps:

To permanently keep the automatically assigned nWWN mapping using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane (see Figure 4-5).

Step 2 Click the Initiators tab.

You see the iSCSI initiators configured.

Step 3 Check the Persistent Node WWN check box for the iSCSI initiators that you want to make static.

Step 4 Click the Apply Changes icon to save these changes.

Proxy Initiator Mode

In the event that the Fibre Channel storage device requires explicit LUN access control for every host use the transparent initiator mode (presenting one iSCSI host as one Fibre Channel host). Every iSCSI host has to be configured statically. This can mean several configuration tasks for each iSCSI host. If you do not need explicit LUN access control, using the proxy initiator mode simplifies the configuration.

In this mode, only one virtual host N port (HBA port) is created per IPS port. All the iSCSI hosts connecting to that IPS port will be multiplexed using the same virtual host N port (see Figure 4-21). This mode simplifies the task of statically binding WWNs. LUN mapping and assignment on the Fibre Channel storage array must be configured to allow access from the proxy virtual N port’s pWWN for all LUNs used by each iSCSI initiator that connects through this IPS port. The LUN is then assigned to each iSCSI initiator by configuring iSCSI virtual targets (see the “Static Mapping” section) with LUN mapping and iSCSI access control (see the “iSCSI Access Control” section).

Figure 4-21 Multiplexing IPS Ports

Proxy initiator mode can be configured on a per IPS port basis, in which case only iSCSI initiators terminating on that IPS port will be in this mode.

When an IPS port is configured in proxy-initiator mode, fabric login (FLOGI) is done through the virtual iSCSI interface of the IPS port. After the FLOGI is completed, the proxy-initiator virtual N port is online in the Fibre Channel fabric and virtual N port is registered in the Fibre Channel name server. The Fibre Channel module with IPS ports or MPS-14/2 module registers the following entries in the Fibre Channel name server:

• iSCSI interface name iSCSI slot /port is registered in the symbolic-node-name field of the name server
• SCSI_FCP in the FC-4 type field of the name server
• Initiator flag in the FC-4 feature of the name server
• Vendor specific flag (iscsi-gw) in the FC-4 type field to identify the N-port device as an iSCSI gateway device in the name server

Similar to transparent initiator mode, the user can provide a pWWN and nWWN or request a system assigned WWN for the proxy initiator N port.

Caution Enabling the proxy initiator mode of an iSCSI interface that is part of an iSLB VRRP group impacts load balancing on the interface. See the “Changing iSCSI Interface Parameters and the Impact on Load Balancing” section.

---

To configure the proxy initiator, follow these steps:

To configure the proxy initiator using Fabric Manager, follow these steps:

Step 1 Expand Switches, expand Interfaces, and then select FC Logical in the Physical Attributes pane.

You see the Interface tables in the Information pane (see Figure 4-22).

Figure 4-22 FC Logical Interface Tables

Step 2 In Device Manager, select Interface > Ethernet and iSCSI.

You see the Ethernet Interfaces and iSCSI dialog box (See in Figure 4-23).

Figure 4-23 Ethernet Interfaces and iSCSI Dialog Box

Step 3 Click the iSCSI tab in either FM or DM.

You see the iSCSI interface configuration table (see Figure 4-24).

Figure 4-24 iSCSI Tab in Device Manager

Step 4 Check the Proxy Mode Enable check box.

Step 5 Click the Apply Changes icon in Fabric Manager or click Apply in Device Manager to save these changes.

Note When an interface is in proxy initiator mode, you can only configure Fibre Channel access control (zoning) based on the iSCSI interface’s proxy N port attributes—the WWN pairs or the FC ID. You cannot configure zoning using iSCSI attributes such as IP address or IQN of the iSCSI initiator. To enforce initiator-based access control, use iSCSI based access control (see the “iSCSI Access Control” section).

VSAN Membership for iSCSI

VSAN membership can be configured for an iSCSI interface, called the port VSAN. All the iSCSI devices that connect to this interface automatically become members of this VSAN, if it is not explicitly configured in a VSAN. The default port VSAN of an iSCSI interface is VSAN 1. Similar to Fibre Channel devices, iSCSI devices have two mechanisms by which VSAN membership can be defined.

• iSCSI host—VSAN membership to iSCSI host. (This method takes precedent over the iSCSI interface).
• iSCSI interface—VSAN membership to iSCSI interface. (All iSCSI hosts connecting to this iSCSI interface inherit the interface VSAN membership if the host is not configured in any VSAN by the iSCSI host method).

Configuring VSAN Membership for iSCSI Hosts

Individual iSCSI hosts can be configured to be in a specific VSAN. The specified VSAN overrides the iSCSI interface VSAN membership.

To assign VSAN membership for iSCSI hosts, follow these steps:

To assign VSAN membership for iSCSI hosts using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane (see Figure 4-5).

Step 2 Click the Initiators tab.

You see the iSCSI initiators configured.

Step 3 Fill in the VSAN Membership field to assign a VSAN to the iSCSI hosts.

Step 4 Click the Apply Changes icon to save these changes.

Note When an initiator is configured in any other VSAN (other than VSAN 1), for example VSAN 2, the initiator is automatically removed from VSAN 1. If you also want it to be present in VSAN 1, you must explicitly configure the initiator in VSAN 1.

Configuring Default Port VSAN for iSCSI Interfaces

VSAN membership can be configured for an iSCSI interface, called the port VSAN. All the iSCSI devices that connect to this interface automatically become members of this VSAN, if it is not explicitly configured in a VSAN. In other words, the port VSAN of an iSCSI interface is the default VSAN for all dynamic iSCSI initiators. The default port VSAN of an iSCSI interface is VSAN 1.

Caution Changing the VSAN membership of an iSCSI interface that is part of an iSLB VRRP group impacts load balancing on the interface. See the “Changing iSCSI Interface Parameters and the Impact on Load Balancing” section.

---

To change the default port VSAN for an iSCSI interface, follow these steps:

To change the default port VSAN for an iSCSI interface using Device Manager, follow these steps:

Step 1 Choose Interface > Ethernet and iSCSI.

You see the Ethernet Interfaces and iSCSI dialog box (see Figure 4-23).

Step 2 Click the iSCSI tab.

You see the iSCSI interface configuration table (see Figure 4-24).

Step 3 Double-click the PortVSAN column and modify the default port VSAN.

Step 4 Click Apply to save these changes.

Example of VSAN Membership for iSCSI Devices

Figure 4-25 provides an example of VSAN membership for iSCSI devices:

• iSCSI interface 1/1 is a member of VSAN Y.
• iSCSI initiator host A has explicit VSAN membership to VSAN X.
• Three iSCSI initiators (host A, host B, and host C) C connect to iSCSI interface 1/1.

Figure 4-25 VSAN Membership for iSCSI Interfaces

Host A’s virtual Fibre Channel N port will be added to VSAN X because of explicit membership for the initiator. The virtual host-B and host-C N ports do not have any explicit membership configuration so they will inherit the iSCSI interface VSAN membership and be part of VSAN Y.

Advanced VSAN Membership for iSCSI Hosts

An iSCSI host can be a member of multiple VSANs. In this case, multiple virtual Fibre Channel hosts are created, one in each VSAN in which the iSCSI host is a member. This configuration is useful when certain resources such as Fibre Channel tape devices need to be shared among different VSANs.

iSCSI Access Control

Two methods of access control are available for iSCSI devices. Depending on the initiator mode used to present the iSCSI hosts in the Fibre Channel fabric, either or both of the access control methods can be used.

• Fiber Channel zoning-based access control—Fibre Channel zoning has been extended to support iSCSI devices, and this extension has the advantage of having a uniform, flexible access control mechanism across the whole SAN. In the case of iSCSI, multiple iSCSI devices may be connected behind an iSCSI interface. Interface-based zoning may not be useful because all iSCSI devices behind the interface will automatically be within the same zone.
• iSCSI ACL-based access control—iSCSI-based access control is applicable only if static iSCSI virtual targets are created. For a static iSCSI target, you can configure a list of iSCSI initiators that are allowed to access the targets. By default, static iSCSI virtual targets are not accessible to any iSCSI host.

Depending on the initiator mode used to present the iSCSI hosts in the Fibre Channel fabric, either or both the access control mechanisms can be used.

The following topics are included in this section:

• Fibre Channel Zoning-Based Access Control
• iSCSI-Based Access Control
• Enforcing Access Control

Fibre Channel Zoning-Based Access Control

Cisco SAN-OS Release 3.x and NX-OS Release 4.1(1b) VSAN and zoning concepts have been extended to cover both Fibre Channel devices and iSCSI devices. Zoning is the standard access control mechanism for Fibre Channel devices, which is applied within the context of a VSAN. Fibre Channel zoning has been extended to support iSCSI devices, and this extension has the advantage of having a uniform, flexible access control mechanism across the whole SAN.

Common mechanisms for identifying members of a Fibre Channel zone are the following:

• Fibre Channel device pWWN.
• Interface and switch WWN. Device connecting via that interface is within the zone.

See the Cisco Fabric Manager Fabric Configuration Guide and Cisco MDS 9000 Family NX-OS Fabric Configuration Guide for details on Fibre Channel zoning.

In the case of iSCSI, multiple iSCSI devices may be connected behind an iSCSI interface. Interface-based zoning may not be useful because all the iSCSI devices behind the interface will automatically be within the same zone.

In transparent initiator mode (where one Fibre Channel virtual N port is created for each iSCSI host as described in the “Transparent Initiator Mode” section), if an iSCSI host has static WWN mapping then the standard Fibre Channel device pWWN-based zoning membership mechanism can be used.

Zoning membership mechanism has been enhanced to add iSCSI devices to zones based on the following:

• IPv4 address/subnet mask
• IPv6 address/prefix length
• iSCSI qualified name (IQN)
• Symbolic-node-name (IQN)

For iSCSI hosts that do not have a static WWN mapping, the feature allows the IP address or iSCSI node name to be specified as zone members. Note that iSCSI hosts that have static WWN mapping can also use these features. IP address based zone membership allows multiple devices to be specified in one command by providing the subnet mask.

Note In proxy initiator mode, all iSCSI devices connecting to an IPS port gain access to the Fibre Channel fabric through a single virtual Fibre Channel N port. Zoning based on the iSCSI node name or IP address will not have any effect. If zoning based on pWWN is used, then all iSCSI devices connecting to that IPS port will be put in the same zone. To implement individual initiator access control in proxy initiator mode, configure an iSCSI ACL on the virtual target (see the “iSCSI-Based Access Control” section).

To add an iSCSI initiator to the zone database, follow these steps:

To add an iSCSI initiator to the zone database using Fabric Manager, follow these steps:

Step 1 Choose Zone > Edit Local Full Zone Database.

You see the Edit Local Zone Database dialog box (see Figure 4-26).

Figure 4-26 Edit Local Zone Database Dialog Box in Fabric Manager

Step 2 Select the VSAN you want to add the iSCSI host initiator to and click OK.

You see the available zones and zone sets for that VSAN (see Figure 4-27).

Figure 4-27 Available Zones and Zone Sets

Step 3 From the list of available devices with iSCSI host initiators, drag the initiators to add into the zone.

Step 4 Click Distribute to distribute the change.

iSCSI-Based Access Control

iSCSI-based access control is applicable only if static iSCSI virtual targets are created (see the “Static Mapping” section). For a static iSCSI target, you can configure a list of iSCSI initiators that are allowed to access the targets.

By default, static iSCSI virtual targets are not accessible to any iSCSI host. You must explicitly configure accessibility to allow an iSCSI virtual target to be accessed by all hosts. The initiator access list can contain one or more initiators. The iSCSI initiator can be identified by one of the following mechanisms:

• iSCSI node name
• IPv4 address and subnet
• IPv6 address

Note For a transparent mode iSCSI initiator, if both Fibre Channel zoning and iSCSI ACLs are used, then for every static iSCSI target that is accessible to the iSCSI host, the initiator’s virtual N port should be in the same Fibre Channel zone as the Fibre Channel target.

To configure access control in iSCSI follow these steps:

To configure access control in iSCSI using Device Manager, follow these steps:

Step 1 Select IP > iSCSI.

You see the iSCSI configuration (see Figure 4-12).

Step 2 Click the Targets tab.

You see the iSCSI virtual targets.

Step 3 Uncheck the Initiators Access All check box if checked.

Step 4 Click Edit Access.

You see the Initiators Access dialog box.

Step 5 Click Create to add more initiators to the Initiator Access list.

You see the Create Initiators Access dialog box.

Step 6 Add the name or IP address for the initiator that you want to permit for this virtual target.

Step 7 Click Create to add this initiator to the Initiator Access List.

Enforcing Access Control

Fibre Channel module with IPS ports and MPS-14/2 modules use both iSCSI and Fibre Channel zoning-based access control lists to enforce access control. Access control is enforced both during the iSCSI discovery phase and the iSCSI session creation phase. Access control enforcement is not required during the I/O phase because the Fibre Channel module with IPS ports or MPS-14/2 module is responsible for the routing of iSCSI traffic to Fibre Channel.

• iSCSI discovery phase—When an iSCSI host creates an iSCSI discovery session and queries for all iSCSI targets, the Fibre Channel module with IPS ports or MPS-14/2 module returns only the list of iSCSI targets this iSCSI host is allowed to access based on the access control policies discussed in the previous section. The Fibre Channel module with IPS ports or MPS-14/2 module does this by querying the Fibre Channel name server for all the devices in the same zone as the initiator in all VSANs. It then filters out the devices that are initiators by looking at the FC4-feature field of the FCNS entry. (If a device does not register as either initiator or target in the FC4-feature field, the Fibre Channel module with IPS ports or MPS-14/2 module will advertise it). It then responds to the iSCSI host with the list of targets. Each will have either a static iSCSI target name that you configure or a dynamic iSCSI target name that the Fibre Channel module with IPS ports or MPS-14/2 module creates for it (see the “Dynamic Mapping” section).
• iSCSI session creation—When an IP host initiates an iSCSI session, the Fibre Channel module with IPS ports or MPS-14/2 module verifies if the specified iSCSI target (in the session login request) is allowed by both the access control mechanisms described in the “iSCSI-Based Access Control” section.

If the iSCSI target is a static mapped target, the Fibre Channel module with IPS ports or MPS-14/2 module verifies if the iSCSI host is allowed within the access list of the iSCSI target. If the IP host does not have access, its login is rejected. If the iSCSI host is allowed, it validates if the virtual Fibre Channel N port used by the iSCSI host and the Fibre Channel target mapped to the static iSCSI virtual target are in the same Fibre Channel zone.

If the iSCSI target is an autogenerated iSCSI target, then the Fibre Channel module with IPS ports or MPS-14/2 module extracts the WWN of the Fibre Channel target from the iSCSI target name and verifies if the initiator and the Fibre Channel target is in the same Fibre Channel zone or not. If they are, then access is allowed.

The Fibre Channel module with IPS ports or MPS-14/2 module uses the Fibre Channel virtual N port of the iSCSI host and does a zone-enforced name server query for the Fibre Channel target WWN. If the FC ID is returned by the name server, then the iSCSI session is accepted. Otherwise, the login request is rejected.

iSCSI Session Authentication

The Fibre Channel module with IPS ports or MPS-14/2 module supports the iSCSI authentication mechanism to authenticate the iSCSI hosts that request access to the storage devices. By default, the Fibre Channel module with IPS ports or MPS-14/2 modules allow CHAP or None authentication of iSCSI initiators. If authentication is always used, you must configure the switch to allow only CHAP authentication.

For CHAP user name or secret validation, you can use any method supported and allowed by the Cisco MDS AAA infrastructure. AAA authentication supports a RADIUS, TACACS+, or local authentication device. See the Cisco Fabric Manager Security Configuration Guide.

The aaa authentication iscsi command enables AAA authentication for the iSCSI host and specifies the method to use. See Cisco MDS 9000 Family NX-OS Security Configuration Guide

To configure AAA authentication for an iSCSI user, follow these steps:

To configure AAA authentication for an iSCSI user using Fabric Manager, follow these steps:

Step 1 Choose Switches > Security > AAA in the Physical Attributes pane.

You see the AAA configuration in the Information pane.

Step 2 Click the Applications tab.

You see the AAA configuration per application (see Figure 4-28).

Figure 4-28 AAA per Application Configuration

Step 3 Right-click the ServerGroup Id List field for the iSCSI application and enter the server group that you want iSCSI to use.

Note You should use an existing server group or create a new server group before configuring it for iSCSI session authentication.

Step 4 Click the Apply Changes icon to save these changes.

The following topics are included in this section:

• Configuring Authentication Mechanism
• Configuring Local Authentication
• Restricting iSCSI Initiator Authentication
• Configuring Mutual CHAP Authentication
• Configuring an iSCSI RADIUS Server

Configuring Authentication Mechanism

You can configure iSCSI CHAP or None authentication at both the global level and at each interface level.

The authentication for a Gigabit Ethernet interface or subinterface overrides the authentication method configured at the global level.

If CHAP authentication is used, issue the iscsi authentication chap command at either the global level or at a per-interface level. If authentication should not be used at all, issue the iscsi authentication none command.

To configure the authentication mechanism for iSCSI, follow these steps:

To configure AAA authentication for an iSCSI user using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane (see Figure 4-5).

Step 2 Click the Globals tab.

You see the iSCSI authentication configuration table.

Step 3 Select chap or none from the authMethod column.

Step 4 Click the Apply Changes icon in Fabric Manager to save these changes.

To configure the authentication mechanism for iSCSI sessions to a particular interface, follow these steps:

To configure the authentication mechanism for iSCSI sessions to a particular interface using Fabric Manager, follow these steps:

Step 1 Choose Switches > Interfaces > Gigabit Ethernet in the Physical Attributes pane.

You see the Gigabit Ethernet configuration in the Information pane.

Step 2 Click the iSNS tab.

You see the iSCSI and iSNS configuration (see Figure 4-29).

Figure 4-29 Configuring iSCSI Authentication on an Interface

Step 3 Right-click on the IscsiAuthMethod field and select none or chap.

Step 4 Click the Apply Changes icon to save these changes.

Configuring Local Authentication

See the Cisco Fabric Manager Security Configuration Guide and Cisco MDS 9000 Family NX-OS Security Guide to create the local password database. To create users in the local password database for the iSCSI initiator, the iSCSI keyword is mandatory.

To configure iSCSI users for local authentication, follow these steps:

To configure iSCSI users for local authentication using Device Manager, follow these steps:

Step 1 Choose Security > iSCSI.

You see the iSCSI Security dialog box (see Figure 4-30).

Figure 4-30 iSCSI Security Dialog Box

Step 2 Complete the iSCSI User, Password, and Password Confirmation fields.

Step 3 Click Create to save this new user.

Restricting iSCSI Initiator Authentication

By default, the iSCSI initiator can use any user name in the RADIUS server or in the local database in authenticating itself to the Fibre Channel module with IPS ports or MPS-14/2 module (the CHAP user name is independent of the iSCSI initiator name). The Fibre Channel module with IPS ports or MPS-14/2 module allows the initiator to log in as long as it provides a correct response to the CHAP challenge sent by the switch. This can be a problem if one CHAP user name and password has been compromised.

To restrict an initiator to use a specific user name for CHAP authentication, follow these steps:

To restrict an initiator to use a specific user name for CHAP authentication using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane (see Figure 4-5).

Step 2 Right-click the AuthUser field and enter the user name to which you want to restrict the iSCSI initiator.

Step 3 Click the Apply Changes icon to save these changes.

Configuring Mutual CHAP Authentication

The Fibre Channel module with IPS ports or MPS-14/2 module supports a mechanism by which the iSCSI initiator can authenticate the Cisco MDS switch’s iSCSI target during the iSCSI login phase. This authentication is available in addition to the Fibre Channel module with IPS ports or MPS-14/2 module authentication of the iSCSI initiator.

In addition to the Fibre Channel module with IPS ports or MPS-14/2 module authentication of the iSCSI initiator, the Fibre Channel module with IPS ports or MPS-14/2 module also supports a mechanism for the iSCSI initiator to authenticate the Cisco MDS switch’s iSCSI target during the iSCSI login phase. This authentication requires the user to configure a user name and password for the switch to present to the iSCSI initiator. The provided password is used to calculate a CHAP response to a CHAP challenge sent to the IPS port by the initiator.

To configure a global iSCSI target user name and password to be used by the switch to authenticate itself to an initiator, follow these steps:

To configure a global iSCSI target user name and password to be used by the switch to authenticate itself to an initiator using Fabric Manager, follow these steps:

Step 1 Choose End Devices > iSCSI in the Physical Attributes pane.

You see the iSCSI tables in the Information pane (see Figure 4-5).

Step 2 Select the Globals tab.

You see the global iSCSI configuration.

Step 3 Fill in the Target UserName and Target Password fields.

Step 4 Click the Apply Changes icon to save these changes.

To configure a per-initiator iSCSI target’s user name and password used by the switch to authenticate itself to an initiator, follow these steps:

Use the show running-config and the show iscsi global commands to display the global configuration. Use the show running-config and the show iscsi initiator configured commands to display the initiator specific configuration.(See the “Displaying iSCSI Information” section for command output examples).

To configure a per-initiator iSCSI target’s user name and password used by the switch to authenticate itself to an initiator using Device Manager, follow these steps:

Step 1 Choose IP > iSCSI.

You see the iSCSI configuration (see Figure 4-12).

Step 2 Complete the Target UserName and Target Password fields for the initiator that you want to configure.

Step 3 Click Create to add this initiator to the Initiator Access List.

Configuring an iSCSI RADIUS Server

To configure an iSCSI RADIUS server, follow these steps:

Step 1 Configure the RADIUS server to allow access from the Cisco MDS switch's management Ethernet IP address.

Step 2 Configure the shared secret for the RADIUS server to authenticate the Cisco MDS switch.

Step 3 Configure the iSCSI users and passwords on the RADIUS server.

iSCSI Immediate Data and Unsolicited Data Features

Cisco MDS switches support the iSCSI immediate data and unsolicited data features if requested by the initiator during the login negotiation phase. Immediate data is iSCSI write data contained in the data segment of an iSCSI command protocol data unit (PDU), such as combining the write command and write data together in one PDU. Unsolicited data is iSCSI write data that an initiator sends to the iSCSI target, such as an MDS switch, in an iSCSI data-out PDU without having to receive an explicit ready to transfer (R2T) PDU from the target.

These two features help reduce I/O time for small write commands because it removes one round-trip between the initiator and the target for the R2T PDU. As an iSCSI target, the MDS switch allows up to 64 KB of unsolicited data per command. This is controlled by the FirstBurstLength parameter during iSCSI login negotiation phase.

If an iSCSI initiator supports immediate data and unsolicited data features, these features are automatically enabled on the MDS switch with no configuration required.

iSCSI Interface Advanced Features

Advanced configuration options are available for iSCSI interfaces on a per-IPS port basis. These configurations are similar to the advanced FCIP configurations and are already explained in that section (see Advanced FCIP Profile Configuration for more information).

To access these commands from the iSCSI interface, follow these steps:

Cisco MDS switches support the following advanced features for iSCSI interfaces:

• iSCSI Listener Port
• TCP Tuning Parameters
• Setting QoS Values
• iSCSI Routing Modes

iSCSI Listener Port

You can configure the TCP port number for the iSCSI interface that listens for new TCP connections. The default port number is 3260. Once you change the TCP port number, the iSCSI port only accepts TCP connections on the newly configured port.

TCP Tuning Parameters

You can configure the following TCP parameters:

• Minimum retransmit timeout (See the “Minimum Retransmit Timeout” section for more information).
• Keepalive timeout (See the“Keepalive Timeout” section for more information).
• Maximum retransmissions (See the“Maximum Retransmissions” section for more information).
• Path MTU (See the “Path MTUs” section for more information).
• SACK (SACK is enabled by default for iSCSI TCP configurations).
• Window management (The iSCSI defaults are max-bandwidth is 1 Gbps, min-available-bandwidth is 70 Mbps, and round-trip-time is 1 msec). (See the “Window Management” section for more information).
• Buffer size (The iSCSI default send buffer size is 4096 KB) (See the “Displaying FCIP Profile Information” section for more information).
• Window congestion monitoring (enabled by default and the default burst size is 50 KB) (See the “Monitoring Congestion” section for more information).
• Maximum delay jitter (enabled by default and the default time is 500 microseconds).

Setting QoS Values

To set the QoS values, follow these steps:

To set the QoS values using Fabric Manager, follow these steps:

Step 1 Expand Switches, expand Interfaces and then select FC Logical in the Physical Attributes pane.

You see the Interface tables in the Information pane (see Figure 4-22).

Step 2 In Device Manager, choose Interface > Ethernet and iSCSI.

You see the Ethernet Interfaces and iSCSI dialog box (see Figure 4-23).

Step 3 Click the iSCSI TCP tab in either Fabric Manager or Device Manager.

You see the iSCSI TCP configuration table.

Step 4 Set the QoS field from 1 to 6.

Step 5 Click the Apply Changes icon in Fabric Manager or click Apply in Device Manager to save these changes.

iSCSI Routing Modes

Cisco MDS 9000 Family switches support multiple iSCSI routing modes. Each mode negotiates different operational parameters, has different advantages and disadvantages, and is suitable for different usages.

• Pass-thru mode

In pass-thru mode, the port on the Fibre Channel module with IPS ports or MPS 14/2 module converts and forwards read data frames from the Fibre Channel target to the iSCSI host frame-by-frame without buffering. This means that one data-in frame received is immediately sent out as one iSCSI data-in PDU.

In the opposite direction, the port on the Fibre Channel module with IPS ports or MPS 14/2 module limits the maximum size of iSCSI write data-out PDU that the iSCSI host can send to the maximum data size that the Fibre Channel target specifies that it can receive. The result is one iSCSI data-out PDU received sent out as one Fibre Channel data frame to the Fibre Channel target.

The absence of buffering in both directions leads to an advantage of lower forwarding latency. However, a small maximum data segment length usually results in lower data transfer performance from the host because of a higher processing overhead by the host system. Another benefit of this mode is iSCSI data digest can be enabled. This helps protect the integrity of iSCSI data carried in the PDU over what TCP checksum offers.

• Store-and-forward mode (default)

In store-and-forward mode, the port on the Fibre Channel module with IPS ports or MPS 14/2 module assembles all the Fibre Channel data frames of an exchange to build one large iSCSI data-in PDU before forwarding it to the iSCSI client.

In the opposite direction, the port on the Fibre Channel module with IPS ports or MPS 14/2 module does not impose a small data segment size on the host so the iSCSI host can send an iSCSI data-out PDU of any size (up to 256 KB). The port then waits until the whole iSCSI data-out PDU is received before it converts, or splits, the PDU, and forwards Fibre Channel frames to the Fibre Channel target.

The advantage of this mode is higher data transfer performance from the host. The disadvantages are higher transfer latency and that the iSCSI data digest (CRC) cannot be used.

Note The store-and-forward mode is the default forwarding mode.

• Cut-through mode

Cut-through mode improves the read operation performance over store-and-forward mode. The port on the Fibre Channel module with IPS ports or MPS 14/2 module achieves this by forwarding each Fibre Channel data-in frame to the iSCSI host as it is received without waiting for the whole exchange complete. There is no difference for write data-out operations from store-and-forward mode.

Figure 4-31 compares the messages exchanged by the iSCSI routing modes.

Figure 4-31 iSCSI Routing Modes

Table 4-1 compares the advantages and disadvantages of the different iSCSI routing modes.

Caution Changing the forwarding mode of an iSCSI interface that is part of an iSLB VRRP group impacts load balancing on the interface. See the “Changing iSCSI Interface Parameters and the Impact on Load Balancing” section.

---

Displaying iSCSI Information

Use the show iscsi command to obtain detailed information about iSCSI configurations.

This section includes the following topics:

• Displaying iSCSI Interfaces
• Displaying iSCSI Statistics
• Displaying Proxy Initiator Information
• Displaying Global iSCSI Information
• Displaying iSCSI Sessions
• Displaying iSCSI Initiators
• Displaying iSCSI Virtual Targets
• Displaying iSCSI User Information

Displaying iSCSI Interfaces

Use the show iscsi interface command to view the summary, counter, description, and status of the iSCSI interface. Use the output to verify the administrative mode, the interface status, TCP parameters currently used, and brief statistics.

Example 4-1 Displaying the iSCSI Interface Information

Displaying iSCSI Statistics

Use the show iscsi stats command to view brief or detailed iSCSI statistics per iSCSI interface. See Example 4-2 and Example 4-3.

Example 4-2 displays iSCSI throughput on an IPS port in both inbound and outbound directions. It also displays the number of different types of iSCSI PDU received and transmitted by this IPS port.

Example 4-2 Displaying Brief iSCSI Statistics for an iSCSI Interface

Example 4-3displays detailed iSCSI statistics for an IPS port. Along with the traffic rate and the number of each iSCSI PDU type, it shows the number of FCP frames received and forwarded, the number of iSCSI login attempts, successes, and failures. It also shows the number of different types of iSCSI PDUs sent and received that are noncritical or occur less frequently, such as NOP in and out (NOP-In and NOP-Out), text request and response (Text-REQ and Text-RESP), and task management request and response (TMF-REQ and TMF-RESP).

Various types of errors and PDU or frame drop occurrences are also counted and displayed. For example, Bad header digest shows the number of iSCSI PDUs received that have a header digest that fails CRC verification. The iSCSI Drop section shows the number of PDUs that were dropped because of reasons such as target down, LUN mapping fail, Data CRC error, or unexpected Immediate or Unsolicited data. These statistics are helpful for debugging purposes when the feature is not working as expected.

The last section, Buffer Stats, gives statistics on the internal IPS packet buffer operation. This section is for debugging purposes only.

Example 4-3 Displaying Detailed iSCSI Statistics for the iSCSI Interface

Displaying Proxy Initiator Information

If the proxy initiator feature is enabled in the iSCSI interface, use the show interface iscsi command to display configured proxy initiator information (see Example 4-4 and Example 4-5).

Example 4-4 Displaying Proxy Initiator Information for the iSCSI Interface with System-Assigned WWNs

Example 4-5 Displaying Proxy Initiator Information for the iSCSI Interface with User-Assigned WWNs

Displaying Global iSCSI Information

Use the show iscsi global command to view the overall configuration and the iSCSI status. See Example 4-6.

Example 4-6 Displaying the Current Global iSCSI Configuration and State

Displaying iSCSI Sessions

Use the show iscsi session command to view details about the current iSCSI sessions in the switch. Without parameters, this command displays all sessions. The output can be filtered by specifying an initiator, a target, or both.

Example 4-7 displays one iSCSI initiator configured based on the IQN (iqn.1987-05.com.cisco:02.3021b0f2fda0.avanti12-w2k) and another based on its IPv4 address (10.10.100.199).

Example 4-7 Displaying Brief Information of All iSCSI Sessions

Example 4-8 and Example 4-9 display the iSCSI initiator configured based on its IPv4 address (10.10.100.199).

Example 4-8 Displaying Brief Information About the Specified iSCSI Session

Example 4-9 Displaying Detailed Information About the Specified iSCSI Session

Displaying iSCSI Initiators

Use the show iscsi initiator command to display information about all initiators connected to an iSCSI interface in the switch. The information can be filtered to display only the desired iSCSI initiator by specifying the initiator name. Detailed output of the iSCSI initiator can be obtained by specifying the detail option. The iscsi-session (and optionally detail) parameter displays only iSCSI session information. The fcp-session (and optionally detail) parameter displays only FCP session information. The output includes static and dynamic initiators. See Example 4-10 and Example 4-11.

Example 4-10 Displaying Information About Connected iSCSI Initiators

Example 4-11 Displaying Detailed Information About the iSCSI Initiator

Use the show fcns database (and optionally detail) to display the Fibre Channel name server entry for the Fibre Channel N port created for iSCSI initiators in the SAN. See Example 4-12 and Example 4-13.

Example 4-12 Displaying the FCNS Database Contents

Example 4-13 Displaying the FCNS Database in Detail

Use the show iscsi initiator configured to display information about all the configured iSCSI initiators. Specifying the name shows information about the desired initiator. See Example 4-14.

Example 4-14 Displaying Information About Configured Initiators

Displaying iSCSI Virtual Targets

Use the show iscsi virtual-target to display information about the Fibre Channel targets exported as iSCSI virtual targets to the iSCSI initiators. The output includes static as well as dynamic targets. See Example 4-15.

Example 4-15 Displaying Exported Targets

Displaying iSCSI User Information

The show user-account iscsi command displays all configured iSCSI user names. See Example 4-16.

Example 4-16 Displaying iSCSI User Names

Configuring iSLB

The iSCSI server load balancing (iSLB) feature provides a means to easily configure large scale iSCSI deployments containing hundreds or even thousands of initiators. iSLB provides the following features:

• The iSLB initiator configuration is simplified with support for initiator targets and auto-zones.
• Cisco Fabric Services (CFS) eliminates the need for manual configuration by distributing the iSLB initiator configuration among all MDS switches in the fabric.
• Dynamic load balancing of iSLB initiators is available using iSCSI login redirect and VRRP.

When not using iSLB, configuring iSCSI requires the following:

• You need to perform multiple configuration steps on the MDS switch, including the following:

– Initiator configuration using static pWWN and VSAN.

– Zoning configuration for initiators and targets.

– Optional create virtual target and give access to the initiator.

– Configuration of target LUN mapping and masking on the storage system for the initiator based on the static pWWN created for the initiator on the MDS switch.

• You need to duplicate the configuration manually on multiple MDS switches.
• There is no load balancing for IPS ports. For example:

– The Virtual Router Redundancy Protocol (VRRP) only supports active and backup, not load balancing.

– You must use multiple VRRP groups and configure hosts in different groups.

iSLB provides the following features:

• The iSLB initiator configuration is simplified with support for initiator targets and auto-zones.
• Cisco Fabric Services (CFS) eliminates the need for manual configuration by distributing the iSLB initiator configuration among all MDS switches in the fabric.

Note Only statically mapped iSLB initiator configuration is distributed throughout the fabric using CFS. Dynamically and statically mapped iSCSI initiator configurations are not distributed.

• Dynamic load balancing of iSLB initiators is available using iSCSI login redirect and VRRP.

This section covers the following topics:

• iSCSI Configuration Limits
• iSLB Configuration Prerequisites
• iSLB Initiators
• Configuring iSLB Using Device Manager
• Configuring iSLB Initiators
• Load Balancing Using VRRP
• Configuring Load Balancing Using VRRP
• iSLB Configuration Distribution Using CFS
• Distributing iSLB Configuration Using CFS

Note Before configuring iSLB, you must enable iSCSI (see the “Enabling iSCSI” section).

Note For iSLB, all switches in the fabric must be running Cisco MDS SAN-OS Release 2.1(1a) or later.

iSLB Configuration Limits

iSLB configuration has the following limits:

• The maximum number of iSLB and iSCSI initiators supported in a fabric is 2000.
• The maximum number of iSLB and iSCSI sessions supported by an IPS port in either transparent or proxy initiator mode is 500.
• The maximum number of iSLB initiators supported in a fabric is 2000.
• The maximum number of iSLB initiators and iSCSI sessions supported by a switch is 5000.
• The maximum number of iSLB sessions per IPS port in either transparent or proxy initiator mode is 500.
• The maximum number of iSLB and iSCSI targets supported in a fabric is 6000.
• The maximum number of switches in a fabric that can have iSLB with CFS distribution enabled is four.
• No more than 200 new iSLB initiators can be added to the pending configuration. Before adding more initiators, you must commit the configuration.
• You cannot disable iSCSI if you have more than 200 iSLB initiators in the running configuration. Reduce the number of iSLB initiators to fewer than 200 before disabling iSCSI.
• iSLB can be used without CFS distribution but if iSLB auto-zone feature is used, traffic is disrupted when any zoneset is activated.
• If IVR and iSLB features are enabled in the same fabric, you should have at least one switch in the fabric where both these features are enabled. Any zoning-related configuration and activation (for normal zones, IVR zones, or iSLB zones) must be performed on this switch. Otherwise, there may be traffic disruption in the fabric.

iSLB Configuration Prerequisites

Perform the following prerequisite actions prior to configuring iSLB:

• Enable iSCSI (see the “Enabling iSCSI” section for more information).
• Configure the Gigabit Ethernet interfaces (see the “Basic Gigabit Ethernet Configuration for IPv4” section).
• Configure the VRRP groups (see the “Configuring Load Balancing Using VRRP” section).
• Configure and activate a zone set (see the Cisco Fabric Manager Fabric Configuration Guide and Cisco MDS 9000 Family NX-OS Fabric Configuration Guide for more information).
• Enable CFS distribution for iSLB (see the “Enabling iSLB Configuration Distribution” section).

iSLB Initiators

iSLB initiators provide the following features in addition to those supported by iSCSI initiators:

• An iSLB initiator also supports iSLB virtual targets.
• Initiator targets–These targets are configured for a particular initiator.
• Load balancing using iSCSI login redirect and VRRP–If iSCSI login redirect is enabled, the IPS Manager redirects incoming sessions to the best interface based on the calculated load for each interface.
• Configuration distribution to other switches using CFS.

iSLB initiators provide the following features in addition to those supported by iSCSI initiators:

• An iSLB initiator also supports iSLB virtual targets. These targets are very similar to iSCSI virtual targets with the exception that they do not include the advertise interface option and as a result are distributable using CFS.
• Initiator targets–—These targets are configured for a particular initiator.
• Load balancing using iSCSI login redirect and VRRP—If load balancing is enabled, the IPS Manager redirects incoming sessions to the best interface based on the calculated load for each interface.
• Configuration distribution to other switches using CFS.

Configuring iSLB Using Device Manager

To configure iSLB using Device Manager, follow these steps:

Step 1 Choose IP > iSCSI iSLB.

You see the iSCSI iSLB dialog box (see Figure 4-32).

Figure 4-32 iSCSI iSLB Dialog Box

Step 2 Click Create to create a new iSCSI iSLB initiator.

You see the Create iSCSI iSLB Initiators dialog box (see Figure 4-33).

Figure 4-33 Create iSCSI iSLB Initiators Dialog Box

Step 3 Set the Name or IP Address field to the iSLB name or IP address.

Step 4 Set the VSAN Membership field to the VSAN that you want the iSLB initiator in.

Also see the “Assigning VSAN Membership for iSLB Initiators” section.

Step 5 Check the Persistent check box to convert a dynamic nWWN to static for the iSLB initiator.

Also see the “Making the Dynamic iSLB Initiator WWN Mapping Static” section.

Step 6 (Optional) Check the SystemAssigned check box to have the switch assign the nWWN.

Step 7 (Optional) Set the Static WWN field to manually assign the static nWWN. You must ensure uniqueness for this nWWN.

Step 8 (Optional) Check the Port WWN Mapping Persistent check box to convert dynamic pWWNs to static for the iSLB initiator.

See the “Making the Dynamic iSLB Initiator WWN Mapping Static” section.

Step 9 (Optional) Check the SystemAssigned check box and set the number of pWWNs you want to have the switch assign the PWWN.

Step 10 (Optional) Set the Static WWN(s) field to manually assign the static pWWNs.

You must ensure uniqueness for these pWWN.

Step 11 (Optional) Set the AuthUser field to the username that you want to restrict the iSLB initiator to for iSLB authentication.

Also see the “Restricting iSLB Initiator Authentication” section.

Step 12 Fill in the Username and Password fields to configure iSLB initiator target CHAP authentication.

Also see the “Configuring iSLB Session Authentication” section.

Step 13 In the Initiator Specific Target section, set the pWWN to configure an iSLB initiator target.

Step 14 (Optional) Set the Name field to a globally unique identifier (IQN).

Step 15 (Optional) Check the NoAutoZoneCreation check box to disable auto-zoning.

Step 16 (Optional) Check the TresspassMode check box.

Also see the “LUN Trespass for Storage Port Failover” section.

Step 17 (Optional) Check the RevertToPrimary check box to revert back to the primary port after an HA failover when the primary port comes back up.

Step 18 Set the PrimaryVsan to the VSAN for the iSLB initiator target.

Step 19 Click Create to create this iSLB initiator.

Step 20 If CFS is enabled, select commit from the CFS drop-down menu.

Configuring iSLB Initiators

This section includes the following topics:

• Configuring iSLB Initiator Names or IP Addresses
• Assigning WWNs to iSLB Initiators
• Making the Dynamic iSLB Initiator WWN Mapping Static
• Assigning VSAN Membership for iSLB Initiators
• Configuring Metrics for Load Balancing
• Verifying iSLB Initiator Configuration
• Verifying iSLB Authentication Configuration
• Configuring Load Balancing Using VRRP
• Configuring iSLB Session Authentication
• Verifying iSLB Authentication Configuration

Configuring iSLB Initiator Names or IP Addresses

You must specify the iSLB initiator name or IP address before configuring it.

Note Specifying the iSLB initiator name or IP address is the same as for an iSCSI initiator. See the “Static Mapping” section.

To enter iSLB initiator configuration submode using the name option for an iSLB initiator, follow these steps:

To enter iSLB initiator configuration submode using the ip-address option for an iSLB initiator, follow these steps:

Assigning WWNs to iSLB Initiators

An iSLB host is mapped to an N port’s WWNs by one of the following mechanisms:

• Dynamic mapping (default)
• Static mapping

Note Assigning WWNs for iSLB initiators is the same as for iSCSI initiators. For information on dynamic and static mapping, see the “WWN Assignment for iSCSI Initiators” section.

Tip We recommend using the SystemAssign system-assign option. If you manually assign a WWN, you must ensure its uniqueness (see the Cisco Fabric Manager Fabric Configuration Guide and Cisco MDS 9000 Family NX-OS Fabric Configuration Guide for more information). You should not use any previously assigned WWNs.

See the Configuring iSLB Using Device Manager.

Making the Dynamic iSLB Initiator WWN Mapping Static

After a dynamic iSLB initiator has logged in, you may decide to permanently keep the automatically assigned nWWN/pWWN mapping to allow this initiator to use the same mapping the next time it logs in (see the “Dynamic Mapping” section).

You can convert a dynamic iSLB initiator to a static iSLB initiator and make its WWNs persistent

Note You cannot convert a dynamic iSCSI initiator to a static iSLB initiator or a dynamic iSLB initiator to a static iSCSI initiator (see the “Dynamic Mapping” section on page 4-20).

Note Making the dynamic mapping for iSLB initiators static is the same as for iSCSI. See the “Making the Dynamic iSLB Initiator WWN Mapping Static” section“Making the Dynamic iSCSI Initiator WWN Mapping Static” section.

Note Only statically mapped iSLB initiator configuration is distributed throughout the fabric using CFS. Dynamically and statically configured iSCSI initiator configurations are not distributed.

See the Configuring iSLB Using Device Manager.

To permanently keep the automatically assigned nWWN/pWWN mapping, follow these steps:

Configuring iSLB Target Access Mapping

In iSLB, all fabric distributed configurations including iSCSI virtual target access are part of the iSCSI initiator configuration. Access is granted using the pWWN or the device alias. You can specify the one or more of the following optional parameters:

• Secondary pWWN
• Secondary device alias
• LUN mapping
• IQN

In addition, you can disable auto zoning.

If you configure an IQN for an initiator target then that name is used to identify the target. Otherwise, an unique IQN is generated for the initiator target.

To configure iSLB initiator access to an iSCSI virtual target, follow these steps:

To verify the iSLB target configuration, use the show islb initiator configured command.

Assigning VSAN Membership for iSLB Initiators

Individual iSLB hosts can be configured to be in a specific VSAN (similar to the DPVM feature for Fibre Channel. The specified VSAN overrides the iSCSI interface VSAN membership.

For more information, see the Cisco MDS 9000 Family NX-OS Fabric Manager Fabric Configuration Guide.

Note Specifying the iSLB initiator VSAN is the same as for an iSCSI initiator. See the VSAN Membership for iSCSI.

To assign VSAN membership for iSLB initiators, follow these steps:

Note When an iSLB initiator is configured in any other VSAN (other than VSAN 1, the default VSAN), for example VSAN 2, the initiator is automatically removed from VSAN 1. If you also want it to be present in VSAN 1, you must explicitly configure the initiator in VSAN 1.

See the Configuring iSLB Using Device Manager.

Configuring Metrics for Load Balancing

You can assign a load metric to each initiator for weighted load balancing. The load calculated is based on the number of initiators on a given iSCSI interface. This feature accommodates initiators with different bandwidth requirements. For example, you could assign a higher load metric to a database server than to a web server. Weighted load balancing also accommodates initiators with different link speeds.

Also, you can configure initiator targets using the device alias or the pWWN. If you configure an IQN for an initiator target, then that name is used to identify the initiator target. Otherwise, a unique IQN is generated for the initiator target.

For more information on load balancing, see the “Load Balancing Using VRRP” section.

Choose IP > iSCSI iSLB in Device Manager and set the LoadMetric field to change the load balancing metric for an iSLB initiator.

See the Configuring iSLB Using Device Manager.

To configure a weight for load balancing, follow these steps:

Verifying iSLB Initiator Configuration

To verify the iSLB initiator configuration, use the show islb initiator configured command.

Configuring iSLB Initiator Targets

You can configure initiator targets using the device alias or the pWWN. You can also optionally specify one or more of the following optional parameters:

• Secondary pWWN
• Secondary device alias
• LUN mapping
• IQN
• VSAN identifier

Note The VSAN identifier is optional if the target is online. If the target is not online, the VSAN identifier is required.

In addition, you can disable auto-zoning.

If you configure an IQN for an initiator target, then that name is used to identify the initiator target. Otherwise, a unique IQN is generated for the initiator target.

To configure iSLB initiator targets, follow these steps:

To configure additional iSLB initiator targets using Device Manager, follow these steps:

Step 1 Choose IP > iSCSI iSLB.

You see the iSCSI iSLB dialog box (see Figure 4-32).

Step 2 Click on the initiator you want to add targets to and click Edit Initiator Specific Targets.

You see the Initiator Specific Target dialog box.

Step 3 Click Create to create a new initiator target.

You see the Create Initiator Specific Target dialog box (see Figure 4-34).

Figure 4-34 Create Initiator Specific Target Dialog Box

Step 4 Fill in the pWWN field with the initiator target pWWN.

Step 5 (Optional) Set the Name field to a globally unique identifier (IQN).

Step 6 (Optional) Check the NoAutoZoneCreation check box to disable auto-zoning (see Figure 4-33).

Step 7 (Optional) Check the TresspassMode check box. See the “LUN Trespass for Storage Port Failover” section.

Step 8 (Optional) Check the RevertToPrimary check box to revert back to the primary port after an HA failover when the primary port comes back up.

Step 9 Set the PrimaryVsan to the VSAN for the iSLB initiator target.

Step 10 Click Create to create this iSLB initiator target.

Step 11 If CFS is enabled, select commit from the CFS drop-down menu.

Configuring and Activating Zones for iSLB Initiators and Initiator Targets

You can configure a zone name where the iSLB initiators and initiator targets are added. If you do not specify a zone name, the IPS manager creates one dynamically. iSLB zone sets have the following considerations:

• Auto-zoning of the initiator with the initiator targets is enabled by default.
• A zone set must be active in a VSAN for auto-zones to be created in that VSAN.
• iSLB zone set activation might fail if another zone set activation is in process or if the zoning database is locked. Retry the iSLB zone set activation if a failure occurs. To avoid this problem, only perform only one zoning related operation (normal zones, IVR zones, or iSLB zones) at a time.
• Auto-zones are created when the zone set is activated and there has been at least one change in the zoneset. The activation has no effect if only the auto-zones have changed.

Caution If IVR and iSLB are enabled in the same fabric, at least one switch in the fabric must have both features enabled. Any zoning related configuration or activation operation (for normal zones, IVR zones, or iSLB zones) must be performed on this switch. Otherwise, traffic might be disrupted in the fabric.

---

To configure the iSLB initiator optional auto-zone name and activate the zone set, follow these steps:

Choose IP > iSCSI iSLB in Device Manager and set the autoZoneName field to change the auto zone name for an iSLB initiator.

See the Configuring iSLB Using Device Manager.

Verifying iSLB Zoning Configuration

The following example shows the show zoneset active command output when the dynamically generated zone name is used:

The following example shows the show zoneset active command output when the configured zone name IslbZone is used:

Configuring iSLB Session Authentication

The Fibre Channel module with IPS ports and MPS-14/2 module support the iSLB authentication mechanism to authenticate iSLB hosts that request access to storage. By default, the Fibre Channel module with IPS ports and MPS-14/2 module allow CHAP or None authentication of iSCSI initiators. If authentication is always used, you must configure the switch to allow only CHAP authentication.

For CHAP user name or secret validation you can use any method supported and allowed by the Cisco MDS AAA infrastructure (see the Cisco Fabric Manager Security Configuration GuideCisco MDS 9000 Family NX-OS Security Configuration Guide for more information). AAA authentication supports RADIUS, TACACS+, or a local authentication device.

Note Specifying the iSLB session authentication is the same as for iSCSI. See the “iSCSI Session Authentication” section.

Restricting iSLB Initiator Authentication

By default, the iSLB initiator can use any user name in the RADIUS or local AAA database in authenticating itself to the Fibre Channel module with IPS ports or MPS-14/2 module (the CHAP user name is independent of the iSLB initiator name). The Fibre Channel module with IPS ports or MPS-14/2 module allows the initiator to log in as long as it provides a correct response to the CHAP challenge sent by the switch. This can be a problem if one CHAP user name and password have been compromised.

To restrict an initiator to use a specific user name for CHAP authentication, follow these steps:

Choose IP > iSCSI iSLB in Device Manager and set the AuthName field to restrict an initiator to use a specific user name for CHAP authentication.

See the Configuring iSLB Using Device Manager.

Mutual CHAP Authentication

In addition to the Fibre Channel module with IPS ports and MPS-14/2 module authentication of the iSLB initiator, the Fibre Channel module with IPS ports and MPS-14/2 module also support a mechanism for the iSLB initiator to authenticate the Cisco MDS switch’s initiator target during the iSCSI login phase. This authentication requires the user to configure a user name and password for the switch to present to the iSLB initiator. The provided password is used to calculate a CHAP response to a CHAP challenge sent to the IPS port by the initiator.

To configure a per-initiator user name and password used by the switch to authenticate itself to an initiator, follow these steps:

Choose IP > iSCSI iSLB in Device Manager and set the Target Username and Target Password fields to configure a per-initiator user name and password used by the switch to authenticate itself to an initiator.

See the Configuring iSLB Using Device Manager.

Verifying iSLB Authentication Configuration

Use the show running-config and the show iscsi global (see Example 4-6) commands to display the global configuration. Use the show running-config and the show islb initiator configured (see Example 4-14) commands to display the initiator specific configuration.

To verify the iSLB user name and mutual CHAP configuration, use the show islb initiator configured command:

Load Balancing Using VRRP

You can configure Virtual Router Redundancy Protocol (VRRP) load balancing for iSLB. The host is configured with a VRRP address as the portal address. When the VRRP master port receives the first iSCSI session from an initiator, it assigns a backup port to serve that particular host. The information is synchronized to all switches through CFS if recovery is needed when a master port fails. The initiator gets a temporary redirect iSCSI login response. The host then logs in to the backup port at its physical IP address. All iSCSI interfaces in a VRRP group that has load balancing enabled must have the same interface VSAN, authentication, proxy initiator mode, and forwarding mode.

You can configure Virtual Router Redundancy Protocol (VRRP) load balancing for iSLB. Figure 4-35 shows an example of load balancing using iSLB.

Figure 4-35 iSLB Initiator Load Balancing Example

The host is configured with a VRRP address as the portal address. When the VRRP master port receives the first iSCSI session from an initiator, it assigns a backup port to serve that particular host. This information is synchronized to all switches through CFS if recovery is needed when a master port fails. The initiator gets a temporary redirect iSCSI login response. The host then logs in to the backup port at its physical IP address. If the backup port goes down, the host will revert to the master port. The master port knows through CFS that the backup port has gone down and redirects the host to another backup port.

Note If an Ethernet port channel is configured between the Fibre Channel module with IPS ports and an Ethernet switch, the load balancing policy on the Ethernet switch must be based on source/destination IP address only, not port numbers, for load balancing with VRRP to operate correctly.

Note An initiator can also be redirected to the physical IP address of the master interface.

Tip iSLB VRRP load balancing is based on the number of iSLB initiators and not number of sessions. Any iSLB initiator that has more targets configured than the other iSLB initiators (resulting in more sessions) should be configured with a higher load metric. For example, you can increase the load metric of the iSLB initiator with more targets to 3000 from the default value of 1000.

Caution A Gigabit Ethernet interface configured for iSLB can only be in one VRRP group because redirected sessions do not carry information about the VRRP IP address or group. This restriction allows the slave backup port to uniquely identify the VRRP group to which it belongs.

---

Changing iSCSI Interface Parameters and the Impact on Load Balancing

All iSCSI interfaces in a VRRP group that has load balancing enabled must have the same interface VSAN, authentication, proxy initiator mode, and forwarding mode. When you need to change any of these parameters for the iSCSI interfaces in a VRRP group, you must do so one interface at a time. During the transition time when the parameter is changed on some interfaces in the VRRP group and not the others, the master port does not redirect new initiators and instead handles them locally.

Caution Changing the VSAN, proxy initiator, authentication, and forwarding mode for iSCSI interfaces in a VRRP group can cause sessions to go down multiple times.

---

VRRP Load Balancing Algorithm for Selecting Gigabit Ethernet Interfaces

When the VRRP master receives an iSCSI session request from an initiator, it first checks for an existing mapping to one of the interfaces in that VRRP group. If such a mapping exists, the VRRP master redirects the initiator to that interface. If no such mapping exists, the VRRP master selects the least loaded interface and updates the selected interface’s load with the initiator’s iSLB metric (weight).

Note The VRRP master interface is treated specially and it needs to take a lower load compared to the other interfaces. This is to account for the redirection work performed by the master interface for every session. A new initiator is assigned to the master interface only if the following is true for every other interface:

VRRP backup interface load > [2 * VRRP master interface load + 1]

Example 4-17 and Example 4-18 are based on the following configurations:

• GigabitEthernet2/1.441 is the VRRP master interface for Switch1.
• GigabitEthernet2/2.441 is the VRRP backup interface for Switch1.
• GigabitEthernet1/1.441 is the VRRP backup interface for Switch2.
• GigabitEthernet1/2.441 is the VRRP backup interface for Switch2.

Example 4-17 Load Distribution with the Default Metric

The follow example output shows the initial load distribution for three initiators with the default load metric value:

The following example output shows load distribution for four initiators. The interface load metric value for the master interface changed from 0 to 1000.

The following example output shows load distribution for nine initiators. The interface load metric values for the backup interfaces have changed.

Example 4-18 Load Distribution with the Metric Set to 3000 on One Initiator

The following example output shows the initial load distribution for three initiators with one initiator having load metric of 3000 and the remaining initiator with the default metric value:

The follow example output shows load distribution for four initiators. The interface load metric value for the master interface changed from 0 to 1000.

The following example output shows load distribution for nine initiators. The interface load metric values for the backup interfaces have changed.