Configuring IBM FAStT Storage Processor to Use Multipathing with HBA and Storage Port Failover Along with Clustering

To configure an IBM FAStT storage array to use clustering as well as multipathing with HBA and storage port failover, make sure the following are true:

• All servers must use the same preferred and active paths.
• All servers must use the same alternate path to logical units (LUNs).
• The storage processor owning the LUNs must be on the preferred and active path connection.

The output lists the World Wide Port Names (WWPNs) of HBAs on your ESX Server and storage processors on the FAStT. Compare these to those listed in the FAStT profile. The target numbers and WWPNs must be identical as seen by each ESX Server connected to the same FAStT.

Resetting Persistent Bindings

In certain configurations, persistent bindings may have been saved while the servers were connected to SANs with different configurations, which may cause incorrect persistent bindings to be saved. If this occurs, you can reset and update the persistent bindings to use the correct paths to the SAN LUNs. To reset the persistent bindings settings through the VMware Service Console:

1. Log on to the service console as root.
2. Delete /etc/vmware/pbindings:
   rm /etc/vmware/pbindings
3. Use a text editor on the service console to edit /etc/init.d/vmware.
4. Find the following line:
   save_san_persistent_bindings

   Add a comment (#) to the beginning of the line:
   # save_san_persistent_bindings

5. Save the file.
6. Reboot ESX Server.
7. Log on to the service console as root.
   
8. Use a text editor on the service console to edit /etc/init.d/vmware.
9. Find the following line:
   # save_san_persistent_bindings

   Remove the comment (#) from the line:
   save_san_persistent_bindings

10. Save the file.
11. Enter the following command:
    wwpn.pl -v

    Note: Run wwpn.pl on all ESX Servers connected to the same FAStT Storage Server and compare the output. If the target numbers and storage processor ports WWPNs do not match, then you may need to investigate the SAN cabling scheme. Refer to Knowlege Base article 1301, at http://www.vmware.com/support/kb/enduser/std_adp.php?p_faqid=1301, for instructions on setting connections for SAN failover using two storage processors.

Configuring SAN Failover for FAStT Storage Servers Using Two or More Storage Processors

To configure a proper highly available SAN failover configuration with FAStT storage models equipped with two storage processors, each ESX Server must have:

• Two Fibre Channel host bus adapters (HBAs), such as QLogic or Emulex.
• Two fibre switches.
• Two storage processors (SPs): SPA and SPB. Each storage processor must have at least two ports connected to the SAN.

• HBA0 and HBA1 on each ESX Server connected to switches 1 and 2, respectively.
• Switch 1 connected to port 1 on SPA and port 2 on SPB. Switch 2 connected to port 2 on SPA and port 1 on SPB. Make sure to connect SPA ports to the lower switch port number compared to SPB ports. This allows SPA ports to be on the first paths seen by the host bus adapters in ESX Servers.

This configuration provides two paths from each HBA. The order of the paths in the above configuration provides HBA and switch failover without the need to trigger Auto-Volume Transfer (AVT). LUNs must be owned by the storage processor to which the preferred and active paths are connected (in the above configuration, they should be owned by SPA). Only a storage processor failure triggers AVT on the FAStT Storage Server. Please note that the above assumes that the switches are not connected via Inter-Switch Link (ISL) in one fabric.

Viewing Multipath Information

To view your current multipathing configuration, log on to your ESX Server machine as root and enter:
vmkmultipath -q

The output lists all known paths to each SAN LUN. The asterisk (*) indicates that the path is the current, active path, and the pound sign (#) indicates that this is the preferred path from the server to the LUN. When connected to IBM FAStT SANs, the default policy in ESX Server 2.0.1 is fixed and the default policy in ESX Server 2.1 is MRU.

For more information on configuring multipathing, see "Using Multipathing in ESX Server" at http://www.vmware.com/support/esx2/doc/esx20admin_multipath_disks.html.