Cisco SN 5428-2 Storage Router Software Configuration Guide, Release 3.2
Chapter 9 - Configuring a High Availability Cluster
Downloads: This chapterpdf (PDF - 213.0KB) The complete bookPDF (PDF - 7.29MB) | Feedback

Configuring a High Availability Cluster

Table Of Contents

Configuring a High Availability Cluster

Prerequisite Tasks

Guidelines for Configuring SCSI Routing Instances

Creating a Cluster

Adding an Unconfigured SN 5428-2 Storage Router

Adding a Minimally Configured SN 5428-2 Storage Router

Joining Stand-alone Storage Routers in a Cluster

Changing Clusters


Configuring a High Availability Cluster


This chapter explains how to configure SN 5428-2 Storage Routers in a cluster to allow the storage routers to back each other up in case of failure. The following tasks are covered:

Prerequisite Tasks

Creating a Cluster

Joining Stand-alone Storage Routers in a Cluster

Changing Clusters

High availability clusters can be configured using CLI commands, as described in this chapter, or via the web-based GUI. To access the web-based GUI, point your browser to the storage router's management interface IP address. After logging on, click the Help link to access online help for the GUI.


Note SN 5428-2 Storage Routers that are deployed for transparent SCSI routing cannot participate in a high availability cluster.


Prerequisite Tasks

All storage routers that will participate in a cluster must have connectivity to the same hosts and the same storage systems, and must be connected to each other through their management and HA interfaces.


Note At least of the interface connections must be live; you cannot connect storage routers in a cluster using cross-over cables.


Guidelines for Configuring SCSI Routing Instances

When you configure SCSI routing instances to run in a high availability cluster, the following operational guidelines apply:

A cluster supports up to 12 active SCSI routing instances.

If you map targets using WWPN, be sure to specify both the primary WWPN (the WWPN associated with the storage resource as known to the primary node in the cluster) and the secondary WWPN (the WWPN associated with the storage resource as known to the second node in the cluster).

Each storage router in a cluster maintains and exchanges information about available resources. Failover by eligibility is enabled by default; HA bases the decision to automatically fail over a SCSI routing instance to another storage router in a cluster based on the Fibre Channel and other resources available to that SCSI routing instance.

Failover occurs when:

All mapped targets are unavailable or a critical resource for the SCSI routing instance is unavailable, and some or all mapped targets would be available from another storage router in the cluster. A critical resource can be a configured Gigabit Ethernet interface, a required Fibre Channel interface, or an internal resource needed to run the SCSI routing instance.

Some mapped targets are unavailable and all mapped targets are available on another storage router in the cluster.

All mapped targets are available, but another storage router in the cluster also has all targets available and is designated at the primary for the SCSI routing instance.

The storage router stops receiving heartbeats from another node within the cluster.


Note If you need more manual control over where a SCSI routing instance runs, you can turn off failover by eligibility on a storage router. If a SCSI routing instance fails over to a storage router that is configured with failover by eligibility turned off, it will continue running on that storage router unless there are no mapped targets available or a critical resource is unavailable. Normal failover resumes when failover by eligibility is turned back on for the storage router where the SCSI routing instance is running.


Creating a Cluster

A high availability cluster is composed of two SN 5428-2 Storage Routers (or one SN 5428-2 and one SN 5428) that back each other up in case of failure. Storage routers in a cluster have connectivity to the same hosts and storage systems, and are connected to each other through their management and HA interfaces. Storage routers in a cluster must be running the same version of software.

The following configuration settings are shared cluster-wide, and when configured on the first storage router in the cluster, will be shared with the other storage router that joins the cluster.

AAA authentication

Access lists

Administrator mode and Monitor mode passwords

Cluster name

SCSI routing instances

VLAN information (VID, VTP mode, domain name, etc.)

To create a cluster, you typically configure a principal storage router (including all cluster-wide settings), and then add a new, unconfigured node or a minimally configured node to the cluster.


Note A minimally configured storage router is one in which the management IP address, system name, and optional network management interfaces have been configured. Other system information, such as HA IP address, DNS, and NTP server may also have been configured. A minimally configured storage router does not have any cluster-wide settings configured.


To create a cluster, perform the following steps:


Step 1 Respond to the prompts from the initial system configuration script, as described in "First-Time Configuration." This script configures the following settings:

Management IP address

System name

HA configuration mode

Cluster name

HA IP address

When prompted to select HA configuration mode, choose clustered. When prompted for cluster name, enter the name of the new cluster. At the end of the initial configuration script, the storage router automatically reboots.

Step 2 When the storage router restarts, complete the system configuration using the setup wizard or other CLI commands, as described in "First-Time Configuration," or "Configuring System Parameters," or the web-based GUI.

Step 3 (Optional) If you are participating in Fibre Channel switched fabric zoning with the storage router, complete the configuration for zoning using the procedures described in "Configuring Fibre Channel Interfaces."

Step 4 Configure all desired SCSI routing instances and access lists, as described in "Configuring SCSI Routing."

Step 5 Add another storage router to the cluster. To add a new, unconfigured SN 5428-2 to the cluster, follow the additional steps in the "Adding an Unconfigured SN 5428-2 Storage Router" section.

To add a minimally configured SN 5428-2 to the cluster, follow the additional steps in the "Adding a Minimally Configured SN 5428-2 Storage Router" section.


Adding an Unconfigured SN 5428-2 Storage Router

To add a new, unconfigured SN 5428-2 to the existing cluster, perform the following steps:


Step 1 Respond to the prompts from the SN 5428-2 initial system configuration script. When prompted to select HA configuration mode, choose clustered. When prompted for cluster name, enter the name of the existing cluster. At the end of the initial system configuration script, the storage router automatically reboots.

Step 2 When the storage router restarts, it communicates with the other member of the cluster to obtain current cluster configuration information. Once the storage router is completely restarted, verify the new cluster configuration. Issue the show cluster command to verify the cluster name and confirm that the SN 5428-2 is exchanging heartbeats with the other member of the cluster.

Step 3 To verify that both storage routers in the cluster include the same cluster configuration elements, issue the following commands from the principal storage router in the cluster:

show aaa

show accesslist all

show scsirouter all from bootconfig

show vlan

show vtp

Issue the same commands from the SN 5428-2 just added to the cluster. The displays should be the same.

Step 4 Use the setup configuration wizard, other CLI commands, or the GUI to complete SN 5428-2 configuration. If you are participating in Fibre Channel switched fabric zoning with the storage router, configure for zoning using the procedures described in "Configuring Fibre Channel Interfaces."

Step 5 (Optional) Save any changes made to the configuration by issuing the appropriate save command with the bootconfig keyword, which updates the bootable configuration for the SN 5428-2 and notifies all storage routers in the cluster of the configuration changes.

Step 6 (Optional) To divide the workload between the storage routers in the cluster, you can manually fail over selected SCSI routing instances using the failover scsirouter command. See "Maintaining and Managing the SN 5428-2 Storage Router," for more information about failing over SCSI routing instances.


Adding a Minimally Configured SN 5428-2 Storage Router

To add a minimally configured SN 5428-2 to an existing cluster, perform the following steps:


Step 1 Run the setup cluster configuration wizard.

When prompted to select HA configuration mode, choose clustered.

When prompted for cluster name, enter the name of the existing cluster.

When prompted, enter the HA IP address for the SN 5428-2. The HA interface for each storage router in a cluster must be on the same IP subnet.

When prompted to retain or delete "scsirouter" instances, enter delete. Deleting means that any existing SCSI routing instances will be deleted from this SN 5428-2. (Since this is a minimally configured SN 5428-2, there should be no SCSI routing instances to delete.)

Enter yes to confirm your changes. The storage router automatically reboots.

Step 2 When the storage router restarts, it communicates with other member of the cluster to obtain current cluster configuration information. Once the storage router is completely restarted, verify the new cluster configuration. Issue the show cluster command to verify the cluster name and confirm that the SN 5428-2 is exchanging heartbeats with the other member of the cluster.

Step 3 To verify that both storage routers in the cluster include the same cluster configuration elements, issue the following commands from the principal storage router in the cluster:

show aaa

show accesslist all

show scsirouter all from bootconfig

show vlan

show vtp

Issue the same commands from the SN 5428-2 just added to the cluster. The displays should be the same.

Step 4 Complete additional system configuration of the SN 5428-2 just added to the cluster, as needed. For example:

Use the setup netmgmt configuration wizard to configure the storage router for network management via SNMP.

Use the setup time configuration wizard to configure the storage router date and time, and optional NTP server information.

Use the CLI or GUI to configure CDP and logging.

If you are participating in Fibre Channel switched fabric zoning with the storage router, configure for zoning using the procedures described in "Configuring Fibre Channel Interfaces."

Step 5 Save any changes to the configuration by issuing the appropriate save command with the bootconfig keyword, which updates the bootable configuration for the SN 5428-2 and notifies all storage routers in the cluster of the configuration changes.

Step 6 (Optional) To divide the workload between the storage routers in the cluster, you can manually fail over selected SCSI routing instances using the failover scsirouter command. See "Maintaining and Managing the SN 5428-2 Storage Router," for more information about failing over SCSI routing instances.


Joining Stand-alone Storage Routers in a Cluster

In some cases you may prefer to completely configure both storage routers (including SCSI routing instances and access lists) as stand-alone systems before joining them into a cluster.

The following example explains the steps required to create a cluster named Cluster1, composed of two SN 5428-2s named Sys1 and Sys2. This example assumes that both SN 5428-2s are fully configured with SCSI routing instances and access lists. Use the scsirouter primary command to assign a preferred storage router to any or all of the SCSI routing instances, if desired. See "Configuring SCSI Routing," for more information about configuring SCSI routing instances.

To create a cluster from fully configured SN 5428-2s, perform the following steps:


Step 1 Use the setup cluster configuration wizard to define Sys1 as a member of the cluster Cluster1. When prompted, enter retain to keep the access list and SCSI routing instance information already defined.

Step 2 Use the show cluster command to verify the cluster name after Sys1 reboots. Verify that all instances and access lists are still available, using show scsirouter and show accesslist commands.

Step 3 (Optional) When Sys2 joins Cluster1, all cluster elements (access lists, VLANs, passwords and AAA settings) currently configured on Sys2 will be deleted. If you want to make any of the cluster elements currently configured on Sys2 available to the cluster, use the appropriate save command to save the elements to a file. For example, to make any of the access lists currently configured on Sys2 available to the cluster, use the save accesslist command to save the access lists to a file. The following command saves all access lists to a file named Sys2_AccessLists:

[Sys2] save accesslist all Sys2_AccessLists

Step 4 (Optional) Because cluster elements can only be manipulated from the first storage router in a cluster, the saved configuration file(s) from Sys2 must be made available to Sys1. See "Maintaining and Managing the SN 5428-2 Storage Router," for more information about managing saved configuration files using either the copy savedconfig command or FTP.

Step 5 Join Sys2 to the new cluster named Cluster1, using the setup cluster configuration wizard. When prompted, enter retain to share the existing SCSI routing instances across the cluster.

Step 6 Use the show cluster command to verify the cluster name after Sys2 reboots. Verify that the defined SCSI routing instances were retained, using show scsirouter command.

Step 7 (Optional) Restore the cluster elements (such as access lists) saved in Step 3 using the appropriate restore from command. Cluster elements can only be manipulated from the first storage router in a cluster, so these commands must be issued from the system Sys1. For example, to restore all access lists from the configuration file named Sys2_AccessLists saved in Step 3:

[Sys1] restore accesslist all from Sys2_AccessLists


Note Restoring AAA or VLAN information overwrites any existing information. You may prefer to make the configuration modifications to the storage router currently in the cluster before adding the new member.


Step 8 (Optional) Save all configuration information on system Sys1 by issuing a save all bootconfig command, which updates the bootable configuration of all storage routers in the cluster with the saved cluster configuration elements.

Step 9 Verify that all SCSI routing instances are active using the show scsirouter stats command on both storage routers.


Changing Clusters

In some situations, you may need to move a storage router from one cluster to another cluster. Moving a fully configured storage router from one cluster to another is more complex than simply adding it to a cluster. Advanced planning is required.

To successfully move a storage router from one cluster to another, perform the following steps:


Step 1 Verify that the storage router to be moved has the same hardware configuration as the other storage router in the cluster you are planning to join. Each node in the cluster must have connectivity to the same IP hosts and Fibre Channel storage. All management interfaces and all HA interfaces for the storage routers within a cluster must be on the same IP subnet; however, the management interfaces must be on a different IP network than the HA interfaces.

Step 2 Decide if you need to retain any of the SCSI routing instances that are configured on the storage router joining the cluster. Retaining data means all SCSI routing instances existing on the storage router joining the cluster will be added to those already configured for the cluster. If the existing instances are not retained, they are deleted.

Step 3 If you are going to retain data, determine if you have any duplicate SCSI routing instance names. When a storage router is added to the cluster, the data in the cluster will overwrite the existing data. You may prefer to change the configuration in the storage router before it joins the cluster to prevent this situation.

Step 4 If you are going to retain data, determine if you need to save existing access list information, or other cluster elements (such as VLANs or AAA configuration). Cluster elements are not retained. Any access lists, VLANs or AAA configuration on the storage router will be discarded when it joins the new cluster. The storage router will also learn Administrator mode and Monitor mode passwords from the cluster.

You can save cluster elements, such as access list information, and then restore them to the cluster. Cluster element information can be restored before or after the storage router joins the cluster by transferring the saved configuration file to the first storage router in the cluster and performing the appropriate restore functions.


Note Restoring AAA or VLAN information overwrites any existing information. You may prefer to simply make the appropriate configuration modifications to the storage router currently in the cluster before adding the new member.


Step 5 Use the setup cluster configuration wizard to join the new cluster. Respond to the prompts to retain or delete configuration as required. The storage router automatically reboots at the end of the configuration wizard.

Step 6 Perform any additional configuration that may be needed. For example, you can fail over SCSI routing instances to this new cluster member to balance traffic load between all storage routers in the cluster.

Step 7 After making configuration changes, use the save all command with the bootconfig keyword to copy and save the storage router configuration, thereby updating the cluster.