Cool Solutions

Alternative method of re-adding a node to a Cluster


October 12, 2010 11:23 am





There are scenarios where Adding a Node That Was Previously in the Cluster will fail when following the Online Documentation. The steps in this article provide a valid workaround to the issue.

Scenarios where this has been seen is clusters that have had nodes removed/added over time. The end result is that the /var/opt/novell/ncs/gipc.conf file is not sequential. For example:
nodeid 0
nodeid 3
nodeid 4
nodeid 5

Alternative method of reintroducing a server to a Cluster


1 – Server to be reintroduced will be using the same name and IP address as original

Scenarios to be used:

1 – Server has physically crashed and cannot be brought back up
2 – Server OS is to be re-installed/upgraded


1- Server to be reintroduced was on the same physical hardware as original Server


1 – If Server is up and operational (OS is to be reinstalled scenario, ie upgrade to 64 bit OS from 32 bit). Otherwise continue to step 2

Perform a cluster leave
Shutdown Server

2 – Remove objects from tree

DO NOT remove Cluster node object represented by this server
Remove NCP server object
Remove “*servername*” objects from container where NCP server object existed

3 – Install New OS onto server (OES2SP2)

Part of the install is to configure OES2 services (Do Not Configure NCS yet)
Make sure server name is the same as original server
Confirm IP Address is the same as original server

4 – Once server is configured and part of the tree configure NSS Devices

Confirm that NSSMU can see the SBD partition
Confirm NSSMU can see the other Cluster Devices
Confirm LUM operational

5 – Go into YAST2/OES Configuration and enable NCS configuration and configure

Put in appropriate information for the Cluster you will be rejoining
Uncheck “Start Clustering Services now” and Select Finish
Since the node already exists the attempt will error out. However most of the configuration required to be part of the cluster will have completed
Exit out of YAST

6 – Complete manual Configuration

Add the attribute of “NCS:NetWare Cluster” with a value of the cluster object’s FDN to the NCP server object of the cluster node.
Note. Even a 2-node OES2SP2 cluster will contain the above attribute. So keeping this consistent.

If you have enabled the local Firewall then you will need to make changes to it. During the normal YAST Cluster wizard for clustering, the /etc/sysconfig/SuSEfirewall2 would have received a modification as follows:

Adjust firewall conf file to ensure the following are set:
FW_ALLOW_FW_BROADCAST_EXT=”yes” <—this may not have been changed

7 – Confirm Node joins Cluster

Perform a rcadminfs start
Perform a rcnovell-ncs start
Node should successfully join cluster
Resources should successfully load on Node providing any specific Resource application software is installed properly on server

8 – Finalize NCS configuration

perform a “chkconfig adminfs on”
perform a “chkconfig novell-ncs on”

9 – Final testing

reboot server
confirm NCS software loads and server joins the cluster
Perform Fail overs of appropriate resources

Note. Success is very high when the pre-existing node sequence within /var/opt/novell/ncs/gipc.conf are sequential. For example:
nodeid 0
nodeid 1
nodeid 2
nodeid 3

1 vote, average: 5.00 out of 51 vote, average: 5.00 out of 51 vote, average: 5.00 out of 51 vote, average: 5.00 out of 51 vote, average: 5.00 out of 5 (1 votes, average: 5.00 out of 5)
You need to be a registered member to rate this post.

Categories: Uncategorized


Disclaimer: This content is not supported by Micro Focus. It was contributed by a community member and is published "as is." It seems to have worked for at least one person, and might work for you. But please be sure to test it thoroughly before using it in a production environment.