Planning Volume 2, Control Workstation and Software Environment
Some configuration options that can make your control workstation more
available are separate from the high availability control workstation
program. They include disk mirroring, uninterruptible power supplies,
and dual disk controllers both internal and external. You must also be
aware of any frame supervisor changes, HACWS limits and restrictions, and
complete the related HACMP planning worksheets.
An uninterruptable power supply can supply electricity to a device to keep
it running when main power is interrupted or is unreliable. Usually an
uninterruptible power supply is not the sole source of power. Rather,
it is typically used to smooth a fluctuating source or to provide enough power
to enable a device to shut down gracefully. You can use an
uninterruptible power supply in conjunction with all other means of assuring
control workstation reliability. See the RS/6000 General Services
Document Site and Hardware Planning Information for the power
consumption requirements of your control workstation.
Each control workstation should be attached to a different electrical power
source or breaker panel if possible. They should at least be on
separate circuits so that maintenance or failures in main power can affect
only one control workstation.
The process of mirroring occurs when each block of data written to one disk
is also written to another disk. You always have a copy of your data in
case one disk or disk adapter fails. As a middle ground to availability
you can decide to have a single control workstation and mirror the root volume
group to provide better availability of the control workstation. This
involves two or three times the number of disks needed for the root volume
group, depending on whether you want one or two mirrors of the
original. See the discussion on mirroring in the book AIX System
Management Guide: Operating System and Devices. That book
describes the AIX support for mirroring root volume groups. See the
book PSSP: Administration Guide for additional SP-specific
support of mirroring root volume groups.
You can cable spare SP Ethernet adapters into the existing Ethernet LAN
segments for the SP and leave them in a defined but unavailable configuration
state. When an Ethernet adapter fails, you can unconfigure the failing
adapter and configure the spare Ethernet adapter for that LAN segment.
You can use the spare adapter until the failed one is repaired or
replaced. Note that the spare Ethernet adapter still counts as one of
the stations in the 30 total stations you can have on an Ethernet LAN
segment.
Check with your IBM representative for information about ordering the
necessary hardware.
The high availability control workstation support has the following
limitations and restrictions:
- DCE restriction:
If you plan to have DCE authentication enabled, you cannot use
HACWS. If you already use HACWS, do not enable DCE
authentication.
- |There are no problems in the authentication of both systems on the
|nodes with the restricted root access option enabled. However, you have
|to copy the authorization files, /.rhosts or
|/.klogin, to the backup control workstation. Also, it
|is important that you activate restricted root access from the active primary
|control workstation, since it is the Kerberos V4 Master.
- IPv6 restriction:
HACWS does not tolerate IPv6 aliases for IPv4 addresses.
- You cannot split the load across a primary and backup control
workstation. Only one or the other provides all the function at any
given time.
- |The primary and backup control workstations must each be a supported
|and independent control workstation. You cannot use an SP or clustered
|enterprise server node as a backup control workstation.
- The backup control workstation cannot be used as the control workstation
for another SP system.
- The backup control workstation cannot be a shared backup of two primary
control workstations.
There is a one-to-one relationship of primary to backup control
workstations; a single primary and backup control workstation combination
can be used to control only one SP system.
- If your primary control workstation is a PSSP Kerberos V4 authentication
server, the backup control workstation must be a secondary authentication
server.
|HACWS is not supported in any configuration of clustered enterprise
|servers. HACWS is also not supported in SP systems with SP-attached
|servers that use the CSP or HMC hardware protocol. Generally, it is not
|a good idea to run HACWS on systems with SP-attached servers.
|However, if you already have HACWS running on your SP system, have
|reasoned that you cannot do without HACWS, accept all the limitations, have
|specialized experience, and can implement your own manual intervention
|procedures for fail over, you might be able to make HACWS continue to work in
|your SP system with SP-attached servers that use the SAMI hardware
|protocol. IBM 
pSeries 680 and RS/6000 Enterprise Servers
|S70, S7A, and S80 connect to the control workstation by two serial
|connections, making them SP-attached servers in an SP system. One
|connection is for hardware monitoring and control and the other is for serial
|terminal support. These servers use the SAMI hardware protocol for
|those connections. Only one control workstation at a time can be
|connected to each server, so there cannot be automatic physical failover done
|by the HACWS software. When the primary control workstation fails over
|to the backup control workstation, hardware control and monitoring support and
|serial terminal support are not available for these servers.
The following apply if you use high availability control workstation
support with |SAMI protocol RS/6000 or pSeries servers in an SP-attached or clustered
enterprise server configuration:
- |Each server is directly attached to the control workstation.
|Depending on the type of server, you might only be able to connect to one
|control workstation at a time. Using HACWS might require manual
|intervention or might not be an option at all. Check the control
|workstation hardware specifications carefully.
- When a control workstation fails or scheduled downtime occurs, and the
backup control workstation becomes active, you will lose hardware monitoring
and control and serial terminal support for the servers:
- The servers will have the SP Ethernet connection from the backup control
workstation, so PSSP components requiring this connection will still work
correctly. This includes components such as the availability
subsystems, user management, logging, authentication, the SDR, file
collections, accounting and others.
- You cannot use PSSP to make configuration changes related to the
servers. For example, you cannot add new servers.
- You cannot use PSSP to reboot a server.
- You can use PSSP to shutdown or restart the SP system but it will not
effect SP-attached servers.
- You cannot use PSSP to shutdown or restart a system of clustered
enterprise servers.
|For a list of the monitoring and control functions that might be
|lost depending on your server hardware, see the discussion of SP-attached and
|clustered enterprise servers in the HACWS chapter of the book PSSP:
|Administration Guide.
You will have to complete the following worksheets in the HACMP
documentation:
- Shared Volume Group/File System Worksheet (Non-Concurrent)
- Defining Shared LVM Components for Non-Concurrent Access
As you complete the HACMP planning and installation steps, take the
Non-Concurrent option whenever you are given the choice.
See the book HACMP: Planning Guide, for complete planning
information.
[ Top of Page | Previous Page | Next Page | Table of Contents | Index ]