Configuration requirements

This topic describes steps you must perform before you configure the SAN Volume Controller.

Steps

Perform the following steps:

  1. Your IBM service representative must have installed the SAN Volume Controller.
  2. Install and configure your disk controller systems and create the RAID resources that you intend to virtualize. To prevent loss of data, virtualize only those RAID that provide some kind of redundancy, that is, RAID 1, RAID 10, RAID 0+1, or RAID 5. Do not use RAID 0 because a single physical disk failure might cause the failure of many virtual disks. RAID 0, like other types of RAID offers cost-effective performance by using available capacity through data striping. However, RAID 0 does not provide a parity disk drive for redundancy (RAID 5) or mirroring (RAID 10).

    When creating RAID with parity protection (for example, RAID 5), consider how many component disks to use in each array. The more disks you use, the fewer disks you need to provide availability for the same total capacity (one per array). However, if you use more disks, it will take longer to rebuild a replacement disk after a disk failure. If a second disk failure occurs during the rebuild period, all data on the array is lost. More data is affected by a disk failure for a larger number of member disks resulting in reduced performance while rebuilding onto a hot spare and more data being exposed if a second disk fails before the rebuild has completed. The smaller the number of disks, the more likely it is that write operations span an entire stripe (strip size x number of members minus 1). In this case, write performance is improved because the disk write operations do not have to be preceded by disk reads. The number of disk drives required to provide availability might be unacceptable if the arrays are too small.

    When in doubt, create arrays with between six and eight member disks.

    If reasonably small RAID arrays are used, it is easier to extend an MDisk group by adding a new RAID array of the same type. Construct multiple RAID devices of the same type, when possible.

    When creating RAID with mirroring, the number of component disks in each array does not affect redundancy or performance.

    Most back-end disk controller systems enable RAID to be divided up into more than one SCSI logical unit (LU). When configuring new storage for use with the SAN Volume Controller, you do not need to divide up the array. New storage should be presented as one SCSI LU. This will give a one-to-one relationship between MDisks and RAID.

    Attention: Attention: Losing an array in an MDisk group can result in the loss of access to all MDisks in that group.
  3. Install and configure your switches to create the zones that the SAN Volume Controller needs. One zone must contain all the disk controller systems and the SAN Volume Controller nodes. For hosts, use switch zoning to ensure that each host fibre-channel port is zoned to exactly one fibre-channel port of each SAN Volume Controller node in the cluster. The SAN Volume Controller and the master console exist in both zones.
    Note:Note: The SAN Volume Controller and the master console are defined in each zone.
  4. If you want the SAN Volume Controller to export redundant paths to disks, you must install the Subsystem Device Driver (SDD) on all of your hosts that are connected to the SAN Volume Controller. Otherwise, you will not be able to use the redundancy inherent in the configuration. Install the SDD from the following Web site:

    http://www-1.ibm.com/server/storage/support/software/sdd.html

    Be sure to install version 1.4.x.x or higher.

  5. Install and configure the SAN Volume Controller master console. The communication between the master console and the SAN Volume Controller runs under a client-server network application called Secure Shell (SSH). Each SAN Volume Controller cluster is equipped with SSH Server software and the master console comes to you equipped with the SSH Client software called PuTTY. You will need to configure the SSH client key pair using PuTTY on the master console. Once you have installed your master console, you can configure and administer the SAN Volume Controller using a graphical interface or a command-line interface.
    • You can configure the SAN Volume Controller using the SAN Volume Controller Console, Web-based application that is preinstalled on the master console.
      Note:Note: You can also install the master console on another machine (which you provide) using the a CD-ROM provided with the master console.
    • You can configure the SAN Volume Controller using the command-line interface (CLI) commands.
    • You can install an SSH client if you only want to use the CLI commands. If you want to use the CLI from a host other than the master console, ensure that the host has an SSH client installed on it.
      Notes:
      Notes:
      1. AIX comes with an installed SSH client.
      2. Linux comes with an installed SSH client.
      3. PuTTY is recommended for Windows.
Result

When you and the IBM service representative have completed the initial preparation steps, you must:

  1. Add nodes to the cluster and set up the cluster properties.
  2. Create managed disk groups from the managed disks to make pools of storage from which you can create virtual disks.
  3. Create host objects from the HBA fibre-channel ports to which you can map virtual disks.
  4. Create virtual disks from the capacity that is available in your managed disk groups.
  5. Map the virtual disks to the host objects to make the disks available to the hosts, as required.
  6. Optionally, create Copy Services (FlashCopy and Remote Copy) objects as required.
Related topics

(C) Copyright International Business Machines Corporation 2003, 2004