|
|
Advanced Novell Network Management: NetWare 6
Chapter 10: Configure and Test High Availability File Access
Objectives:
This chapter discusses setting up high availability data access. The
objectives are:
- Configure NCS for High Availability File Access
- Manage Resources in an NCS Cluster
Concepts:
Configure NCS for High Availability File Access
The chapter begins with instructions to build a shared storage area.
It adds some advice not offered before:
- Running ConsoleOne from a workstation is faster than running it from
a server.
- The path to ConsoleOne from a workstation is presented:
SYS:PUBLIC\MGMT\CONSOLEONE\1.2\BIN\CONSOLEONE.EXE.
- In ConsoleOne, select the cluster object or the object for a server
in the cluster. Right-click it, select Properties, and select
the Devices tab.
- Select the device to create the shared partition on. Place
a check in the box for Sharable for Clustering if it is not already
there.
- Choose the Media tab, Partitions, and New.
- Select the same device as above. Enter the partition size,
choose NSS as the type, and enable Hot Fix and Mirror
options.
- Choose Create New Mirror Group, and click OK.
A variation is presented for creating a volume and a pool at the same
time. If you begin by creating a new volume, you will be prompted to create
a pool. You will need to provide an IP address for the pool. You will
also need to select protocols for server advertising. Choose NCP
for Novell clients, CIFS for Microsoft clients, and AFP
for Macintosh clients.
As described in another chapter, you should turn on the Flush Files
Immediately option to improve data integrity (less lag time for disk
writes), but you should be aware that it may slow network performance.
It is possible to cluster-enable an existing NSS volume (and its
pool). When you do so, a new volume object replaces the old one.
The new volume object is associated with a new pool object, which
is associated in turn with a new virtual server object. The virtual
server provides access to the volume in case of any device's failure.
Manage Resources in an NCS Cluster
The second objective covers the concepts of migrating and troubleshooting
resources.
As in an earlier chapter, the definition of migration here is
that of relocating a resource from one server to another. This discussion
covers moving the resource as a management choice rather than as
a response to device failure. You might choose to do this to balance
loads on servers, to free a device for maintenance, or to improve
a service by migrating it to newer devices.
ConsoleOne and NetWare Remote Manager can both be used
to migrate resources. Resources must be running to be migrated.
Verify that they are running with the Cluster State View in ConsoleOne,
and from Cluster Management, Cluster Status View in Remote
Manager.
In order to change the settings for a resource, take it off line first.
This is also suggested as the first option in troubleshooting a
problem: take the resource off line, then bring it back on
line. Your next option is to examine the state of the cluster:
- Alert - a resource is waiting for a manual start, failover,
or failback. These are the actions you should choose from.
- Comatose - a resource is not running, and should be taken off
line. Take the resource off line and correct any problems.
- Unloading - shown while a resource is being unloaded. No action
to take.
- Running - normal state, in which a resource can be migrated.
Migrate, take off line, or leave it alone.
- Loading - shown while a resource is being loaded. No action
to take.
- Unassigned - shown when no node is available for the resource
to run on. Take it off line to prevent it being loaded, or bring up
a device to run it.
- NDS_Sync - eDirectory has not synchronized changes yet. Wait
for it to do so.
- Offline - a resource is not on online. It can be brought on
line from this state.
- Quorum Wait - a resource is waiting for a quorum of servers
before it can be run. No action to take.
|
