Novell D7171A - NetServer - LPr, LC2000r - NetServer - 128 MB RAM, LH3000r - NetServer - 128 MB RAM, LH4r - NetServer - 256 MB RAM Installation Manual

Final

disclaimer

Novell, Inc. makes no representations or warranties with respect to the
contents or use of this documentation, and specifically disclaims any
express or implied warranties of merchantability or fitness for any
parti cula r purpos e. Furt her, Nove ll, Inc. re serves the rig ht to revi se this
publication and to make ch anges to its conte nt, at any time, without
obligation to noti fy any person or e ntity of such re visions or changes.

Further, Novell, Inc. makes no represe ntation s or warranti e s with
respect to any software, and specifically disclaims any express or
implied warranties of me rchantab ility or fitn e ss f or any particular
purpose. Furthe r, Nov ell, In c. res erves th e right to make c hanges to a ny
and all pa rts of Nov e ll softwar e , at any ti me, without any obli gation to
notif y any per son or e n tity of su c h changes.

export notice

trademarks

This product may require export authorization from the U.S.
Department of Commerce prior to exporting from the U.S. or Canada.

Novell and NetWare and GroupWise are registered trademarks of
Novell, Inc. in the United States and other countries. ConsoleOne, NDS,
NetWare Cluster Services, Novell Client, and Novell Storage Services
are trademarks of Novell, Inc. Novell Authorized Reseller is a service
mark of Novell, Inc.

Windows and Wi ndows NT are registered trademar ks of Microsoft
Cor p or a ti o n . Ne tscape E nt e rp rise S e rv e r is a trade m a rk of Ne tsca p e
Communications Co rp oration. Oracle is a registered trademark of
Oracle Corporation.

Copyright © 1999 Novell, Inc. All rights reserved. No part of this
publication m ay be reproduced, photocopi ed, stored on a retrieval
system, or tra nsm itted without the express written consent of the
publisher.

U.S. Patent Nos. 5,157,663; 5,349,642; and 5,455,932. U.S. Patent
Application No. 5,572,528. U.S. and Foreign Pat ents Pending.

Novell, Inc.
122 East 1700 South
Provo, UT 84606
U.S.A.

Overview and Installation
8/99
100-004511- 001 A

ontents

C

Introduction to Cluster Services

Introduction to Cluster Services. . . . . . . . . . . . . . . . . . . . . . . . . . . v

1Overview

Product Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Product Benefits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Cluster Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Cluster Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Inst allat ion and Setu p

Hardware Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Software Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Shared Disk System Requirements. . . . . . . . . . . . . . . . . . . . . . . . . 8

Installing NetWare Cluster Services . . . . . . . . . . . . . . . . . . . . . . . . 9

NetWare Cluster Services Licensing . . . . . . . . . . . . . . . . . . . . . 9

Begin the Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Add Servers to the Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Setting Up NetWare Cluster Services. . . . . . . . . . . . . . . . . . . . . . . . 14

Create Cluster Volumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Cluster-Enable Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Create Cluster Resource Templates . . . . . . . . . . . . . . . . . . . . . 16

Create Cluster Resources. . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Configure Load Scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Configure Unload Scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Set Failover and Failback Modes . . . . . . . . . . . . . . . . . . . . . . . 19

Assign Nodes to a Resource . . . . . . . . . . . . . . . . . . . . . . . . . 20

Configuration Settings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Client Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Edit Quorum Membership and Timeout Properties . . . . . . . . . . . . . . 21

Membership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Contents

iii

Cluster Protocol Properties. . . . . . . . . . . . . . . . . . . . . . . . . . 22

Heartbeat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Tolerance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Master Watchdog. . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Slave Watchdog . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Max Retransmits . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Cluster Port Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Cluster Node Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Number+IP Address . . . . . . . . . . . . . . . . . . . . . . . . . . 24

NCP Server. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Migrate Resources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Identify Cluster and Resource States . . . . . . . . . . . . . . . . . . . . . . . 25

NetWare Cluster Services Console Commands . . . . . . . . . . . . . . . . . . 27

iv

Overview and Installation

Introduction to Cluster Services

Novell Cluster Services is a server clustering system that ensures high
availability and manageability of critical network resources including data
(volumes), applications, server licenses, and services. It is a multinode, NDS­enabled clustering product for NetWare 5 that support s failover , failback, and
migration (load balancing) of individually managed cluster resources.

Introduction to Cluster Services

v

vi

Overview and Installation

chapter

1

Overview

NetWare® Cluster ServicesTM is a server clustering syst em tha t ensur es high
availability and manageability of critical network resources including data
(volumes), applications, server licenses, and services. It is a multinode
clustering product for Ne tWare 5 that is enabled for NDS
failove r, failback, and migration (loa d balancing) of indi vidually managed
cluster resources.

Product Features

NetWare Cluster Services includes se veral importan t features to help you
ensure and manag e the availability of your network resources. These include:

• Support for shared dis k configurations or local disk configurations.

TM

and supports

• Multinode al l-act ive c luste r (u p to 32 no des). Any Ne tWare server i n the
cluster can resta rt resources (applications, services, IP addresses, and
volumes) from a failed server in the cluster.

• Single point of administration through a Java-based ConsoleOne
cluster con figuration and monitoring GUI.

• The abil ity to tailor a cluster to the specific applications and hardwa re
infrastructure that fit your organization.

• Dynamic assignment and reassignment of server storage on an as­needed basis.

TM

Chapter 1: Overview

1

Product Benefits

NetWare Cluster Services allows you to configure up to 32 NetWare 5 serve r s
into a high-availability cluster, where resources can be dynamically switched
or moved to any serv er in the cluster. Resources can be configured to
automatic ally switch or be moved in the event of a server failure, or ca n be
moved manually to troubleshoot ha rdware or balance the workload.

NetWare Cluster Services provides high availability from commodity
components. Lower costs are obtained through the consolidation of
applicati ons and operation s onto a cluster . Th e ability to manage a cluster from
a single point of control and to adjust res ources to meet changi ng workload
requiremen ts ( thus , manually “load balance” the cluster) are also important
benefits of NetWare Cluster Services .

An equall y imp ortant benefi t of i mplemen ting NetWare C luster Se rvice s is t hat
you can reduce unpl anned service outages and reduce planned outages for
software and hardware maintenance and upgrades.

Reasons you would want to implement NetWare Cluster Services inc lude:

• Increased availability

• Improved performance

• Low cost of operation

• Scalabi lity

• Disaster recovery

• Dat a p r ot ec ti o n

• Shared resources

Shared disk fault tol erance can be obtained by implemen ting RAID Level 5 on
the shared dis k subsystem.

An example of the benefits NetWare Cluster Services provides can be better
understood through the following scenario.

Suppose you have conf igured a three-server cluster, with a Web s erver installed
on e ach of the three ser vers in t he cl uster . Eac h of t he s ervers in the cluste r ho sts

2

Overview and Installation

two Web sites. All t he dat a, gra phi cs, and e-mai l messa ges fo r each Web site is
stored on a shared disk subsystem connected to each of the servers in the
cluster. The following figure depicts how this setup might look.

Web Server 1

Web Site A
Web Site B

Web Server 2

Web Site C
Web Site D

Fiber Channel Switch

Web Site E
Web Site F

Shared Disk

System

Web Server 3

During normal clu ster operation, each server is in constant comm unication
with the other se rvers in the cluster and performs periodic poll ing of all
registered resources to detect failure.

Suppose Web Server 1 experiences hardware or soft ware pr oblems and the
users depending on Web Server 1 for Internet access, e-mail, and information
lose their connections. The following figure shows how resources are moved
when Web Server 1 fails.

Web Server 1

Web Server 2

Web Site A
Web Site C
Web Site D

Fiber Channel Switch

Shared Disk

System

Web Server 3
Web Site B
Web Site E
Web Site F

Chapter 1: Overview

3

W e b Sit e A moves t o Web Server 2 and Web Site B moves to Web Server 3. IP
addresses and applicabl e licenses a lso move to Web Server 2 and Web Server 3.

When you configured the cluster, you decided where the Web sites hosted on
each Web server would go should a fai lure oc cur. In the previous example you
configure d Web Site A to move to Web Server 2 and Web Site B to move to
W e b Se rver 3. This way, the workloa d once ha ndled b y Web Server 1 i s evenl y
distribute d.

When Web Server 1 failed, NetWare Cluster Services software

• Detected a failure.

• Restarted applications (th at were running on Web Server 1) on Web
Server 2 and Web Server 3 as specified.

• Transferred IPaddresses to Web Server 2 and Web Server 3 as specified.

• Remounted the shared data volumes (that were formerly mounted on
Web server 1) on Web Server 2 and We b Server 3 as specified.

In this example, the failover proces s ha ppened quickly and users regained
access to the Inter net, Web site information, a nd e-mail within second s, and in
most cases, without having to log in again.

Now suppose the problems with Web Server 1 are resolved, and Web Server 1
is returned to a nor mal operating state. Web Site A and Web Site B will
automatic ally failback, or be moved back to Web Server 1, and Web Server
operation will return back to the way it was before Web Server 1 failed.

NetWare Cluster Ser v ices also pro v ides resource mi gr ation capabilities. You
can move applications, Web sites, etc. to other se rvers in your cluster without
waiting for a server to fail.

For example, you coul d have m anual ly move d Web Site A or Web Site B from
W e b Server 1 to eith er of t he other se rvers i n the clust er. Y ou migh t want to do
this to up grade or perform schedule d maintenance on Web Server 1, or just to
increase perform an c e or accessi bilit y of the Web sites.

4

Overview and Installation

Cluster Configuration

Typical cluster configurations normally include a shared disk subsystem
connected to all servers in the cluster. The shared disk subsystem can be
connected via high-speed fiber cha nnel cards, cables, and switches, or be
configured t o use sha red SCSI. If a server fails, another des ignated s erver in the
cluster automatically mounts the shared subsyst em volumes previously
mounted on t he faile d server . This giv es network us ers continu ous acc ess to the
volumes on the shared disk subsystem.

Typical resources m ight include data (volumes), applications, server licenses,
and services. The following figure shows how a typical fiber channel cluster
configuration might look.

Network Hub

Network

Interface

Card(s)

Server 1 Server 2 Server 3 Server 4 Server 5 Server 6

Sys Sys Sys Sys Sys Sys

Fiber Channel Switch

Shared Disk

System

Although fi ber channel is the recommended configuration, you can configure
your cluster to use shared SCSI. The following figure shows how a typical
shared SCSI cluster configuration might look.

Fiber
Channel
Card(s)

Chapter 1: Overview

5

Network Hub

Network

Interface

Card

Server 1 Server 2

Sys Sys

Cluster Components

The following components make up a NetWare Cluster Services cl uster:

• From 2 to 32 NetWare 5 servers configured to use IP, each cont aining at
least o n e local disk device (used for a loca l SYS: volume).

SCSI
Adapter

Shared Disk

System

Network

Interface

Card

SCSI
Adapter

• NetW a re Clus ter Serv ices s oftware runni ng on each Net Ware 5 server in
the cluster.

• A shared d isk s ub system conne cted to all s ervers in the c luster ( opti ona l,
but recommended for most co nfigurations).

• High-speed fiber channel cards, cables, and switch or SCSI cards and
cables used to connect the servers to the shared disk subsystem .

6

Overview and Installation

chapter

2

Installation and Setup

Hardware Requirements

The following list specifies hardware requirements for installing NetWare®
Cluster Services
configura tion. Addi tional h ardware might be necess ar y depe nding o n ho w you
intend to use Net Ware Cluster Services.

A minimum of t w o NetWare 5 server s . A maximum of eight servers per



cluster is recommended for this release.

At least 64 MB of memory on all serve r s in the cluster (128 MB



recommended for fa iling multiple applications to the same server).

At least one local dis k device (not shared) for volume SYS: on each



server.

TM

. These requirements represent the minimum hardware

It is recommended the cli ent machine used to manage the cluster be at



least 300 MHz or above and have at least 90 MB of memory. Although
you can use slower machines with less memory, faster processor
machines will greatly increase the performance of ConsoleOne, so
choose the highest performance works tation available.

Software Requirements

You must have the NetWare 5 Support Pack 3 or la ter running on each cluster
server. Y ou ca n install Support Pack 3 for Net Ware 5.0 from the NetWare
Cluster Services product CD. Ensur e the following requirements are met:

All servers in the cluste r configur ed with the IP protoc ol and on the same



IP subnet.

All serve r s in the cluster in the same NDSTM tree.



Chapter 2: Installation and Setup

7

The Novell® ClientTM included o n t he Net Ware Cluster Servi ces product



CD inst alled on the workst ations u sed to m anage and a ccess y our c lust er.

ConsoleOneTM installed (from the NetWare Cluster Se rvices product



CD) on the workstati on used to manage your cluster.

Shared Disk System Requirements

A shared dis k system is required for eac h cluster if you want data to be highly
available. If a shared disk subsystem is used, ensure the following:

At least 10 MB of free disk space on the shared disk sy stem for creating



a spec ia l cl u s te r partiti o n.

The NetWare Cluster Services installation automatically allocates one
cylin d er on one drive of the shared disk system fo r the spec ial cluster
partition. If the drive where the cluste r partition is to be created on the
shared dis k system is la rger than 10 GB, you may need more free space.

The sha red di sk s yst em i s pr oper ly s et up a nd functi onal ac cording t o the



manufacturer’s instructions.

Prior t o installation, verify that all the drives in your share d disk system
are recogniz ed by Net Ware by running the List Devices command on
each server you intend to add to your cluster. If any of the drives in the
shared disk system do not show up in the list, consult your NetWare 5
documentat ion or the shared disk sys tem documentation for
troubleshooting information.

All sh ared disk system vo lu mes are configured to use Novell Storage



Services

The disks contained in the shared storage subsystem are configured in a



mirroring or RAID 5 configuration to add faul t tolerance to the sh ared
disk subsystem.

Warning I f the disks in the shared disk system are not configure d to use mirroring or R AID

5, a single device error can cause a system failure. NetWare Cluster Services
softwa re will not protect against such faults.

TM

.

8

Overview and Installation

Installing NetWare Cluster Services

The NetWare Cluster Services installation

• Creates a new Cluster object in NDS.

• Installs NetWare Cluster Services software on the servers you specify to
be part of your cluster.

• Updates th e AUTOEXEC.NCF file on ea ch of the servers i n your cluste r .
Note changes to AUTOEXEC.NCF:

• LOAD TCPIP.NLM is replaced with

LOAD TCPIP . NLM FORWARD=NO
This command d isables I P forwarding . If IP for warding i s enabl ed,

client reconnect will not function, and users won’t automatically
reconnect to different servers in the cluster after failover.

• NSS is replaced with NSS/AUTODEACTIVATE

VOLUME=ALL
NetWare Cluster Services autom atically mounts clus ter volumes

on specified servers in the cluster. This command ensures that
cluster vol umes don’t accid entally get mount ed on multiple se rvers
in the cluster, which would cause data corruption. Without this
command, the MOUNT ALL command could be run from any
server console and that server would tr y to mo unt a volume that
might alrea dy be mounted on another server in the cluster.

Once you have created a cluster, you will need to run the NetWare Cluster
Services ins tallation agai n ea ch time you want to add servers to your cl us ter.

NetWare Cluster Services Licensing

NetWare Cluster Services changes the licensing scheme currently provided
with NetWare 5. The license changes performed by the NetWare Cluster
Services i nstallati on depend on whe ther you have a Master Li cense Agreement
(MLA). Changes to NetWare 5 non-MLA licensing include:

• The server base licenses for each cluster server are deleted and are
replaced with a server base license upgrade. This server base license
upgrad e works specifically with NetWare Cluster Services.

Chapter 2: Installation and Setup

9

The server base license upgrade is insta lled when you install NetWare
Cluster Services and is placed in the same context a s its Server object.

• All server connect ion licenses for ea ch cluster server ar e deleted and are
all replaced with one or more Cluster User Access Licenses (CUAL).
Y ou need one or more CUALs to equal the number of conne ctions fo r the
server c onne ction li cens es that are del et ed. The CUAL is an o bje ct i n t he
NDS database and acts as a holding container for connecti ons . This
means that client s connecting to any clus ter server can draw connec tions
from the CUAL as needed, as long as there are still conne ctions
available.

The CUAL is installed when you install NetWare Cluster Servi ces and
by default is placed in the lowest common co n text for all servers in t h e
cluster. In order for users to get connections to servers in the cluster,
CUALs need to be accessible to User objects. This means individual
CUALs need to be placed at or above the user contex t in the NDS tree so
they are accessible.

• The Cluster Serv er License is a new licens e type and i s required on ever y
server you have in your cluster. The Cluster Ser ver L icense allows a
server to join a cluster.

The Cluster Server License is installed on a server when you install
NetWare Cluster Services on that server. The Cluster Server License
object is cr eated in the same context as the Cluster object.

In most cases, the Clust er Server License is the only license required by
NetWare Clust er Services if an MLA l icense is present i n th e D irectory tree.
The Net W are Cluste r Serv ices i nstalla tion wi ll aut omatica lly dete ct whe ther a n
MLA exists in the Direct or y tr ee and if so, ins tall only the Cluster S erver
License.

For more informat ion on Novell Licensing Services go to http://

www.novell.com/documentation/lg/nls501/docui/index.html.

10

Overview and Installation

Begin the Installation

T o begin the installation, insert the NetWare Cluster Services CD and complete
the following steps:

1. From the initial splash screen, launch the NetWare Cluster Services

2. Select Create a New Cluster or Edit an Existing Cluster, then click

3. Do one of the followi n g:

installation.

Continue through the ins tallation screens until you get to the screen that
prompts you to cre ate a new cluster or edit an existing cluster.

Next.

• (If creating) Enter the name for the new Cluster object you are
creating and sp ecify the Directory tree and context where you want
it created. Then click Next.

• (If editing) Specify the Directory tree, context, and name of the
cluster you will add servers to. If you don’t know a cluster name,
browse and selec t one from the list. Then click Next.

4. Enter the name of the server you want to add to the cluster, or
browse and select one from the list, and then click Add to Cluster.
Repeat this step for every serve r you want to add to the cluster . Then
click Next.

You can also remove servers you ju st added to the cluster by selecting
them from the NetWare Servers in Cluster list and clicking Remove.

When you add a server to a cluster, NetWare Cluster Services
automatic ally det ects the serve r’s IP add ress. I f the server you a re adding
has more than one IP address, you will be prompted to select the IP
address you want Net Ware Cluster Services to use.

5. (Condi tional ) If you a re cr eati ng a n ew cl uster, specif y whe ther your
cluster has a shared disk system and if so, select the drive where you
want the small cluster partition created. Then click Next.

NetWare Cluster Services requires a small cluster partition on the shared
disk system . You are also given the option of mirroring the partition for
greater fault tolerance.

Chapter 2: Installation and Setup

11

Imp or tant You must have at least 10 MB of free space that is not part of an NSS partition

on one of the s hared disk drives to create the cluster partition. If no free space
is available, the shared disk drives can’t be used by NetWare Cluster Services.

6. Choose whether or not you want the servers you are adding to your
cluster to reboo t a f ter instal l ati o n , t hen cl ic k N ex t.

The inst allation program inst alls NetWa re Cluster Se rvices software on
the serve rs in your clust er. It does not a utomati cally loa d t he s oftware on
the servers in your cluster. An easy way to automatically load the
software is to reboot the servers you are adding to your cluster after the
installation is complete.

7. Specify the location of the license files and click Add > Next.

License files are located on the license diskettes included in the NetWare
Cluster Services product box. The Cl uster Server License diskette
contains t wo Cluster Server Licenses, two S erver Base License Upgrade s
and one 10-user CUAL. Cluster Server licenses are the only licenses
required for most MLA customer s. Additional CUALs, Server Ba se
Licen se U p grades, and Cl u s t er S er v er Li censes ca n b e obtained fro m
your Novell Authorized Resell er

SM

.

If you do not have a Master License Agree ment, you will be prompted
for an activation key for each node you add to a cluste r. Obtaining
activation keys will require you to provide the serial numbers for the
license s you ar e upgra ding. Activat ion keys can be obtained at http://

www.novell.com/products/activation.

You are given the option of installing with or wit hout licenses. If you
choose to install without licenses, you will need to manually install the
necessary licenses at a later date using NetWare Administrator. NetWare
Cluster Services will not function without the proper li censes in place.

8. Continue through the final installation screen.

The install ation progr am will t hen creat e a new clus ter for you and inst all
NetW are Clust er Services software on th e servers you specifie d to be part
of your clust er. If you are editing an exi sting cluster, the installation
progra m will just install NetW ar e Cluster Serv ices software on the
servers you ar e add ing to your cluster.

12

Overview and Installation

Add Servers to the Cluster

After creating your cluste r, you will need to run the NetWare Cluster Services
installation again each time you want to add additional servers to your cluster.

If you plan to add new nodes to an exi sting cluster in the future, you should
reserve those nodes before bringing up your cluster. To reserve future nodes
before bring ing up your cluster:

1. In ConsoleOne, right-click the Cluster object and select Properties.

This will bring up the Cluster Object prope rty page.

2. Selec t th e Protocol tab and then Internals.

3. Add the IP address and node number of your future node to the
Node number + IP address mapping table.

For example, if the IP address of your new node is 123.45.67.89 and the
next unused cl uster node number is 6, add the following line.

nodeid 123.45.67.89 6

If you don’t reserve future nodes, each time you want to add a new node you
will nee d to run the i nstallat ion program aga in, and al so temporari ly remove al l
existing s ervers from t he cluster a nd then have th em join the cluster ag ain. This
is necessary for the exist ing servers in the clu ster to re cognize a ny new serv ers
you have added to the cluster.

To add additional servers to the cluster after creating the cluster, complete the
following steps.

1. Complete the installation procedure in “Begin the Installation” on
page 11.

Follow the instructions for editing an existing cluster. When you get to
the step to reboot new cluster servers after installation, choose not to
reboot.

2. Type CLUSTER DOWN at the server console on any server in t he
cluster.

This wi ll ca use al l s er vers i n t he cl uste r to le ave t he c lu ster at ex ac tly the
same time. All loaded resour ces will be automaticall y unloaded. If all
servers don’t leave the clus ter at exactly the same ti me, resources will

Chapter 2: Installation and Setup

13

begin to f ailove r from se rvers t hat ha ve lef t the c luster t o ser vers t hat ar e
stil l in th e cl u s ter.

3. Type ULDNCS at the server console on every server in the cluster.

This will unload NetWare Cluster Services software on each server.

4. On a server in your NDS tree, Run DSRepair > Advanced Options
menu > Replica and Partition Operations > Synchronize the Replica
on All Servers.

This is necessary to resynchronize the NDS tree. If possible, DSRepair
should be run on a server that contains a maste r replica of the root
partiti on.

5. Reboot the new server or servers you have just added to the cluster.

6. Type LDNCS at the server console on each server in the cluster
except the new server or servers you have just added.

This rel o ads NetWare Cluster Services software on exi sting cluster
server s and cau ses them t o rejoin t h e cluster. The existing ser vers in the
cluster will now recognize the servers you have just added.

Setting Up NetWare Cluster Ser v ices

If you created a new clus ter, you now nee d to create and configure cluster
resources, and if necessary, create cluster volumes. You also should consider
cluster-e nabling NetWare volumes and creati ng cluster resource templ ates.

While ConsoleOne is required for most cluster-related tasks, you can use
NetW a re Administrator t o modify trus tee assignme nts on files, dir ectori es, and
cluster-enabled volumes. To use NetWare Administrator to modify trustee
assignments, you will first need to establish a connection to the server or
cluster-enabled volume.

14

Overview and Installation

Create Cluster Volumes

If you plan on using a shared disk system in your cluster and need to cre ate new
NetWare volumes, those volu mes must be created with only one server up in
the cluster. The serve r used to create the volumes should already have NSS
installed and running.

After creati ng the volumes , bring up the ot her servers i n the cluster. Depending
on your c onfi guratio n, the ne w volum es will eit her m ount auto matic ally when
resources that require them start or will have to be mounted manually on
individua l servers after they are up.

If you have a volume that is not being managed by NetWare Cluster Services
and is no t share d storag e, you ne ed to m anually ac tivate t hat volum e before yo u
can access it. Otherwise, the volume will be deactivated by the NSS/
AUTODEACTIVATE VOLUME=ALL command in the AUTOEXEC.NCF
file. To manually activa te a non-sha red volume, ty pe NSS/ACTIVATE=VOL
at the server console prompt. Replace VOL wit h the volume name. You will
then need to manually mount the volume by typing MOUNT followed by the
volume name.

Cluster-Enable Volumes

If you have a shared disk system that is part of your cluster and want the
volumes on the shared disk system to be highly available to NetWare clients,
you will need to cluster-enable those volumes. Cluster-enabling a volume
allows it to be moved or mounted on different servers in the cluster during
failure s or when migration is necessary.

Some server applications don’t require NetWare client access to volumes, so
cluster-e nabling volumes might not be necessary.

1. In ConsoleOne, browse and select the Cluster object.

2. Click File > New > Cluster > Cluster Volume.

3. Browse and select a volume on the shared disk system to cluster­enable.

4. Enter an IP address for the volume.

Each volume you clust er-enable requires it s own IP address.

Chapter 2: Installation and Setup

15

5. Click Create, t hen continue with “Set Failover and Failback Modes”
on page 19.

To complete the process for cluste r-enab lin g a volume, you now ne ed to
set failove r and fa ilback modes and if necessary, change the node
assignments for the volume (See “Assign Nodes to a Resource” on
page 20).

Create Cluster Resource Templates

Templates simpl if y the process of creating similar or identical cluster
resources. For example, templates are helpf ul when you want to create multiple
instances of the same resourc e on differen t serve rs. Yo u can cre ate temp lat es
for any server application or resource you want to add to your cluster.

NetWare Cluster Ser v ices currently provide s te mplates for GroupWise

®

,
Oracle*, and Netsca pe E nterprise Server*, as well as a generic IP SERVICE
template . The gene ric IP SERVICE template can be used when configuring
certain server applications to run on your cluster. You can edit and customize
any of the templates for your specific need s.

1. In ConsoleOne, browse and select the Cluster object where you want
to create a cluster resource template.

2. From the menu bar, click File > New > Cluster > Cluster Resource.

3. Enter a name for the new cluster resource template.

4. Check the Create Cluster Resource Template check box.

This optio n lets you create a cluste r resource te mplate i nstead of a c luster
resource.

5. Check the Define Add itional Properti es check box, then continue
with “Configure Load Scripts” on page 17.

T o finis h creating a clu ster resourc e template, you need to configu re load
and unload scri pts, set failover and failback modes and, if necess ary,
change the node assignments for the resource template.

16

Overview and Installation

Create Cluster Resources

Cluster resour ce s must be crea ted for ev ery res ource or ap plica tion you run on
servers in your cluster . Clust er resources can inc lude W eb site s, e-mail serv ers,
databases, and any other server-based applications or services you want to
make availabl e to u sers at all times.

1. In ConsoleOn e, browse and select the Cluster object you want to
create res ou rces fo r.

2. Click File > New > Cluster > Cluster Resource.

3. En t er a name for the n ew cluster resource .

4. If a template exists for the resource you are creating, enter the
template name in the Inherit From Template field, or browse and
select it from the list. If a template does not exist, check the Define
Additional Properties check box.

5. Click Create.

6. (Conditional) If you are not using a template for the resources,

Configure Load Scripts

A load scri pt is required for each resource, service, or volume in your cluster.
The load s cript specifies the comma nds to start the r es our ce or service on a
server, or to mount the volum e on a server.

You can use any commands in the load script that would be used in an .NCF
file run from the serve r console. If you don’t know which commands to add to
your load script, consult the documentation for the application or resource.

continue with “Configure Load Scripts” on page 17.

If you are not using a template, you must complete the process for
creating the cluster resource by configuring load and unload scripts,
setting failover and failback modes, and if necessary change the node
assignments for the resource.

If you are using a template for this resource, additional resource
configuration is performed automatically by the template.

1. Select the Load Script tab on the property page.

Chapter 2: Installation and Setup

17

2. Edit or add the necessary commands to the script to load the
resource on th e server.

Some commands m ay require command line input. You ca n add << to a
command to indica te command-line input. For example, a script
command might read:

LOAD SLPDA <<Y
This means that when SLPDA is lo ad ed, it will receive a Y at the

command-line , presumabl y to a quest ion tha t needs a yes answer . If t here
are more inputs required, they can be continued on subsequent lines, as
follows:

LOAD SLPDA <<Y
<<Y
<<N
The string can be up to 32 chara cters.

3. Specify a timeout value.

The default is 600 seconds, or 10 minutes. The timeout valu e determines
how much ti me the script is given to complete. If the script does not
com pl et e w i th in th e spec i fi e d time, the res ou r ce beco m e s id l e .

Configure Unload Scripts

Depending on you r cluster application or resource, you can add an unl oad
script to specify how the application or resource should terminate. An unload
script is not required by all resources or applications, but it can ensure that
during a fail bac k or ma nual migration, a resou rce unloads before it loads on
another node. Consul t your appl icat ion vendor or documenta tion to determ ine
if you should add com mands to unload the resour ce.

1. Select the Unload Script tab on the property page.

2. Edit or add the necessary commands to the script to unload the
intended resource on the server

You can use any commands used in an .NCF file run from the server
console. If you don’t know which commands to add, consult the
documentation for the application or resource you want to unload.

18

Overview and Installation

3. Specify a timeout value.

The default is 600 seconds, or 10 minutes. The timeout valu e determines
how much ti me the script is given to complete. If the script does not
com pl et e w i th in th e spec i fi e d time, the res ou r ce beco m e s id l e .

Set Failover an d Failback Modes

You can configure failover and failback of cluster resources to happe n
manually or aut om atically. If you want applications or resources to
automatically move to specified nodes in the event of hardware and software
failures, set the failover mode to Auto ma tic.

Setting the failback mode to autom atic ensures applications or res ources
automatically move back to their preferred node when the preferred node is
brought back online. The preferred node is the first server in the nodes list on
the Nodes property page. The resource will not automaticall y failback to any
other servers in the cluster except the preferred node.

Set the failover mode to Manual i f you want to intervene a fter a failu re occurs
and before the re sourc e is move d to anothe r node. Sett ing the fa ilover mode to
Manual gives you time to bring up failed node s or migrate resources on other
nodes before allowing the resource to move.

Manual failback works much the same as manual failover . Use Manual
failback to pre vent a re sourc e from moving ba ck to its pre fer red node af ter the
preferred node is brought back online .

1. Select the Policies tab on the property page.

2. Check the Ignore Quorum check box if you don’t want the cluster­wide timeout period and node number limit enforced.

The quorum default valu es were set when you installed NetWare Cluster
Services. You can change the quorum default values by ac ce ssing the
properties page for the Cluster object.

3. Choose the failover and failback modes for this resource.

The default for both failover and failback is Auto.

Chapter 2: Installation and Setup

19

Assign Nodes to a Resource

When you create a resource on a c lus ter or cluster-enable a volum e, the nodes
in the cluster are automatically assigned to the resource or volume. The order
of assignmen t is the order the nodes appear in the resou rce list . You can assign
or unassign nodes to the resource or volume, or change the failover order.

1. Select the Nodes tab on the property page.

2. From the li s t of u na s si gned no des, select the ser v er yo u wa n t t h e
resource assigned to and click the right arrow button to move the
selected server to the Assigned Nodes list.

Repeat this st ep for all servers you want as s igned to the resource. You
can also us e the left arrow button to unassign servers from the resource.

3. Click the up and down arro w buttons to change t he failover order of
the servers assigned to the resource or volume.

Configuration Settings

Depending on your nee ds and clus ter setup , some ad ditional configur ation may
be required for you to effective ly use NetW are Cluste r Services. This ad ditional
configura tion consis t s of changing the network sett ing s on the clien ts that will
access the cluster. You might also need to change the values on some of the
properties for the Cluster object and the Cluster node objects.

Client Configuration

Certain advanced settings are required for Windows 95/98 and W indows NT
clients to transparently reconnect or access cluster enabled volumes in
NetW are Cl uster Servi ces. Use the f ollowing intr uctions to ens ure workstat ions
are properly configured.

Note Autoreconnect only works for vol u m es that are cluster-enabled.

For Windows* 95/98 clients:

1. In the Control Panel, select Network, then select Novell NetWare
Client.

2. Click Properties.

20

Overview and Installation

3. Select the Advanced Settings tab.

4. Verify or change the settings to the following:

Auto Reconnect Level=3
Auto Reconnect Timeout=default
Handle Net Errors=ON
Name Cache Level=0
Net Status Timeout=60
Netware Protoc ol=NDS

For Wi ndows NT* clients:

1. In the C on tro l Pane l, se l ec t Net w o r k , th en sel ec t Se rv i ce s.

2. Select Novell Client for NT and verify or change the following
settings.

On the Advanced Settings tab:

Auto Reconnect=ON

On the Protocol Preferences tab:

Preferr ed Net work Protocol=IP
Protocol Com ponent Settings=NDS

Edit Quorum Membership and Timeout Properties

To edit Quorum Membership and Timeout properties, in ConsoleOne, right
click the Cluster object and select Properties. This will bring up the Cluster
Object Prope rty page. Select the Quoru m tab.

Membership

The Quorum Membership is the number of nodes that must be run ning in the
cluster bef ore resources will start to load. When you first bring up servers in
your cluster, NetWare Cluster Services reads the number specified in the
Membershi p field and waits until that number of servers is up and running in
the cluster before it starts loading resources. Set the Mem b er ship valu e to a
number gr eater t han 1 so tha t all res ources do n’ t a utomati call y l oad on the fir st
server that is brought up in the cluster . For example , if you set the Membership

Chapter 2: Installation and Setup

21

value to 4, the re must be four ser vers up i n the cluster be fore any resourc e will
load and start.

Timeout

Ti meout s pecifi es the amount of ti me to wait for t he num ber of se rve rs def ined
in the Membership field to be up and running. If the timeout period elapses
before the quorum member sh ip reaches its specifi ed number, resources will
automatically start loading on the servers that are currently up and running in
the clust er. For example, if y ou s peci fy a Membe rs hip val ue of 4 a nd a tim eout
value equal to 30 seconds, and after 30 seconds, only two servers are up and
running in the cluster, resources will begin to load on the two servers that are
up and running in the cluster.

Cluster Protocol Properties

You can use the Cluster Prot ocol property pages to vi ew or edit the transmit
frequency and tolerance settings for all nodes in the cluster, including the
master nod e. The master node is generally the first node brought online in the
cluster, but if that node fails, any of the other nodes in the cluster can become
master.

Heartbeat

Tolerance

To view or edit Clus ter Protocol properties, in ConsoleOne , right click the
Cluster object and select P roperties. This will bring up the Cl uster Object
property page. Select the Pr otocol tab. This ta b has two pages, Settings and
Internal s. The Internals page lets you view the sc ript used to configure the
cluster protocol settings, but not change it. Use the Settings page to make
changes to cl uster protocol properties.

Heartbeat specifies the amount of time between transmits for all nodes in the
cluster e xcept t he mast er. For example, if you s et thi s val ue to one , non-m ast er
nodes in the cluster will send a signal that they are alive to the master node
every second.

T oleran ce specifie s the amount of time the master node give s all other nodes in
the cluster to signal that they are alive. For example, setting this value to four
means that if the master node does not receive an "I’m alive" signal from a node
in the cluster within four seconds, that node will be removed from the cluster.

22

Overview and Installation

Master Watchdog

Slave Watchdog

Max Retransmits

Master Watchdog specifies the amount of time between transmits for the
master nod e in the clus ter. For example, if you s et this va lue to one, the mast er
node in the cluster will transmit an "I’m alive" signal to al l the other node s in
the cluste r every second.

Slave W a tchdog speci fies the amoun t of time the mas ter node has to s ignal that
it is alive. For example, setting this value to five means that i f th e non-master
nodes in t he cl uste r do not rec eive an "I’m a liv e" si gnal fro m the master wi thi n
5 seconds, the maste r node will be removed from the cluster and one of the
other nodes will become the master node.

This option is not currently use d with NetWare Cluster Services but will be
used for future versions.

Cluste r P o rt Pr operties

The default cluster port number is 7023, and is automatically assigned when
the cluster is created. The cluster port number does not need to be changed
unless ther e is a co nflict created by another resource using the s ame port
number . If there is a port number confli ct , cha nge the Port number to any othe r
value that doesn’t cause a conf lict.

T o view or edit the Cluste r Port property, in ConsoleOne, righ t click the cluste r
object and select Propertie s. Thi s will bring up the Cluster Object property
page. Select the Port tab.

Cluster Node Properties

You can use the Cluster Node property page to view or edit the cluster node
number or IP address of th e selected node or view the cont ext for the NetWare
Server object.

To view or edit Clus ter Node properties, in ConsoleOne, selec t the Cluster
object, right clic k the desired cluster no de on the right side of the ConsoleOne
display screen, and select Properties. This will bring up the Cluster Node
property page. Select the Node tab.

Chapter 2: Installation and Setup

23

Number+IP Address

Number+IP Address specifies the cluster nod e number a nd IP address for the
selected nod e. If the clus ter node numbe r or IP address c hanges for the sele cted
node, the new information is not automatically updated in NDS. Edit the
information and click Apply to update the information in NDS.

NCP Server

The NCP Se rver field is used to v iew the context for the Ne tWare S er v er
object. Thi s field cannot be edited.

Migrate Resources

You can migrate resources to different servers in your cluster without waiting
for a failure to occur. You might want to migrate resources to lessen the load
on a specific server , to free up a server so it can be brought down for sc heduled
maintena nce, or to increase the performance of the resource or application by
putting it on a faster machine.

Migrating res ources allows you to balan ce the load and evenly distribute
applications among the servers in your cluster.

1. In ConsoleOne, browse and select the Cluster object that contains
the resource you want to migrate.

Resources must be in a running state to be migrat ed.

2. Ensure the right half of ConsoleOne displays the Cluster View State
by selecting VIEW > CLUSTER ST ATE from the menu at the top of
the screen .

3. In the Cluster Resource List, select the resource you want to migrate.

The Cluster Re source Manager s creen a ppears, displ ayi ng the serv er t he
selected re sourc e is cu rrently runni ng o n, and a list of p ossible serv ers t o
which you can migrate resources.

4. Select a server from the list and click Migrate to move the resource
to the se le ct ed server.

Note If y ou sel ect a r eso urce and cl ic k Of fline , t he res our ce wil l be un loaded fr om the

server. It will not load on any other se rver s in the cluster and will remain

24

Overview and Installation

unloaded u ntil you lo ad it again. This opt ion i s usefu l f o r edit ing r esourc es s ince
resources can’t be edited while loaded or running on a server.

Identify Cluster and Resource States

The Cluste r S tate view in ConsoleOne gives you important information abo ut
the status of serve rs and resources in your cluster.

Cluster servers and resources display in different colors, depending on their
operating state. When servers and resources are green, they are in a normal
operati ng condition. When a server that has been part of the c luster is red with
a break in the icon, it has failed. When a resource is red, it is waiting for
administrat or inter vention . When a serve r is red with no bre ak in the icon, tha t
server is not part of the cluster , or its state is unknown. When a resource is
blank, o r has no co lored ico n, it is unas signed, of fli ne, i dle , or i n th e p roces s of
loading or unl oading.

The yellow ball i n th e mid dle of t he server i con desi gnates t he ma ster server i n
the cluster. The master server is initially the first s erver in the cluster, but
another server can becom e master if the first server fails.

The Epoch number in dicates the number of times the cluster state has changed.
The clu ster state will change every time a server joins or leaves the cluster.

Clicking on the Event Log tab gives you a detailed history of your cluster.
Every time the cluster state changes, a new event is added to the Event Log.
You can sort the events in the log by clicking on the column headings of the
table. You can reverse the sort order by pressing the Shift key while clicking
on a column heading.

Clicking on the HTML Report tab in the Cluster Sta te view screen launches a
more detail ed report of the state of your cluster. You can view this report or
save it to an HTML file for printing or viewing with a browser.

The following table identifies the diff erent resource states and gives
descriptions and possible actions for each state.

Chapter 2: Installation and Setup

25

Resource State Description Possible Actions

Alert Either the Start, Failover,

or Failback mode for the
resource has been set to
Manual. The resource is
waiting to start, failover, or
failback on the specified
server.

Comatose The resource is not

running properly and
requi re s administra to r
interv e n tio n .

Unloading The resource is in the

process of unloading from
the server it was running
on.

Running The resour ce is in a normal

runni ng stat e.

Clic k th e A le r t sta tus
indicator and depending
on the resource state, you
will be prompted to Start,
Failover, or Failback the
resource.

Clic k th e C omatos e st a tu s
indicator and offline the
resource. Once resource
problems have been
resolved, the resource can
be put back online
(returned to the running
state) .

None.

Click the Running status
indicator and choose to
either migrat e the resource
to a dif ferent server in your
cluster, or unload (offline)
the resource.

26

Overview and Installation

Loading The re source is in the

process of loading on a
server.

Unassigned There isn’t an assigned

node up that the resource
can be loaded on.

None.

Clic k th e U nassi gn ed
status indicator and if
desired, Offline the
reso urce. Offlinin g th e
resource will prevent it
from running on any of its
preferr ed nodes should a ny
of them join th e cl uster.

Resource State Description Possible Actions

NDS_Sync The properties of the

resour ce have change d and
the changes are still being
synchronized in NDS.

Offlin e Offli ne st atus indi cate s the

resource is shutdown or is
in a dormant or inactive
state . It do es not indicat e a
server f ailu re has occ urr ed
or the resource is waiting
to failover.

Quorum Wai t The res ource is wai ting for

quorum to be established
so it can begin loading.

None.

Clic k th e O ffline st atus
indicator and if des ired,
click the online button to
load the resource on the
best node possible, given
the current state of the
cluster and its preferred
nodes list.

None.

NetWar e Cluster Services Console Commands

Two .NCF script files that can be run from the server console are created by
NetWare Cluster Services. These commands can be useful for updating your
cluster softwa re or troubleshooting cluster problems.

ULDNCS.NCF unloads NetWare Cluster Services software from the server.
LDNCS.NCF reloads NetWare Cluster Services software on the server.

TRUSTOOL is a server cons ole command with three command-line options.

• TRUSTOOL<volname>DUMP lists to the screen an d to
SYS:ETC\TRUSTDMP.TXT all Dis tinguished Names and the state of
the migration.

• TRUSTOOL<volname>FIX forces TRUSTMIG to finish a partial
migrate back to the previous host serve r.

• TRUSTOOL<volname>PURGE is used to delete bad Dis tinguished
Names (DNs) in the TRUSTMIG.FIL file. You must first manually
mount the volume before using TRUSTOOL with the PURGE option.
This will rename TRUSTMIG.FIL to TRUSTMIG.BAK and make a new
TRUSTMIG.FIL in t he _NETWARE directory. Bad DNs occur because

Chapter 2: Installation and Setup

27

a user or group wa s deleted without removing tr ustee r ights to file s on the
volume, o r becaus e NDS didn’ t sync hronize s oon enough after new users
with tr u s te e as signm ents w er e adde d to th e tree.

Occasional problems occur if a volume fails over immediately after a
user is deleted who had trustee rights to that volume. If you experience
errors when a volume fails over and its truste es migrate to the new server,
you will need to run TRUST OOL with the PURGE option to delet e the
user from the trus t ee databa se. You should re view the errors and re move
only those usernames that correspond with deleted users.

NetW a re Cluster Se rvices pr ovides oth er server co nsole commands t o help you
perform certain cluster related tasks. Type HELP CLUSTER at the console
prompt to get information on the commands and their functions.

28

Overview and Installation