Technical data

There will be some cluster down time. During the procedure, you can
maintain cluster operations except for the time it takes to shut down the
second system and boot the first system as a single-node cluster.
______________________ Note _______________________
If you are not using a quorum disk, the first member you shut
down must have zero votes for the cluster to survive its shutdown.
Use the
clu_quorum command to adjust quorum votes. See the
clu_quorum
(8) reference page and the Cluster Administration
manual for more information.
To upgrade from a virtual hub configuration to a standard hub configuration,
follow the steps in Table 5–5. In this procedure, system1 is the member
system that will be shut down first. Member system system2 will be shut
down last. The procedure is written with the assumption that you have
dual-rail failover-pair Memory Channel adapter modules.
Table 5–5: Upgrading from a Virtual Hub Configuration to a Standard Hub
Configuration
Step
Action Refer to:
1
Install the Memory Channel hubs at an appropriate
distance from the member systems.
Section 5.4
If you are adding fiber optics, for each system you
will have in the cluster you need to:
—
Set the hub linecard J2 and J3 jumpers to
enable fiber optics.
Section 5.1.2
Install the optical converters in the hub, ensuring
that you connect the optical cable to the optical
converter when it is installed.
Section 5.5.2.4
Connect the fiber-optic module in the hub to the linecard
with a 1-meter (3.3-foot) BN39B-01 link cable.
Section 5.5.2.4
2
Manually relocate all applications from system1 to
system2. Use the cluster application availability
(CAA) caa_relocate command.
caa_relo-
cate
(8) refer-
ence page and
Cluster Admin-
istration
3
On system1 log in as the root user and execute the
shutdown -h command to halt the system.
Tru64 UNIX
System
Administration
Setting Up the Memory Channel Cluster Interconnect 5–27