IBM TotalStorage SAN Volume Controller, Installation Manual

Page: 33 / 123

pool. You choose how each group is made up, and you can combine managed
disks from different manufacturers’ controllers in the same managed disk group if
required.
Note: Some operating systems cannot tolerate other operating systems in the
same zone.
You can remove one node in each I/O group from a cluster when hardware service
or maintenance is required. After you remove the node, you can replace the field
replaceable units (FRUs) in the node. All disk drive communication and
communication between nodes is performed through the SAN. All SAN Volume
Controller configuration and service commands are sent to the cluster through an
Ethernet network.
Each node contains its own vital product data (VPD). Each cluster contains VPD
that is common to all the nodes on the cluster, and any system connected to the
Ethernet network can access this VPD.
Cluster configuration information is stored on every node that is in the cluster to
allow concurrent replacement of FRUs. An example of this information might be
information that is displayed on the menu screen of the SAN Volume Controller.
When a new FRU is installed and when the node is added back into the cluster,
configuration information that is required by that node is read from other nodes in
the cluster.
SAN Volume Controller operating environment
v
Minimum of one pair of SAN Volume Controller nodes
v Minimum of two uninterruptible power supplies
Fibre Channel
Fabric
Host Host Host Host
SAN Volume
Controller
.
.
.
RAID RAID RAID RAID
. . .
Host zone
Disk zone
SAN Volume
Controller
SAN Volume
Controller
Figure 5. Example of a SAN Volume Controller in a fabric
Chapter 1. Overview of the SAN Volume Controller
3