Environment
Novell Open Enterprise Server 2 (OES 2) Linux Support Pack 3
Situation
The cluster resources would not show up in slptool findsrvs service:smdr.novell. The physical node would show up fine. The resources also did not show up in slptool findsrvs service:bindery.novell. The cluster resources were not being registered with slp. No errors were produced when the resources came online. They worked ok. Drives could be mapped to the resource volumes and the data accessed. The backup software could not select the cluster resource to be backed up because smdr could not see them. If an attempt was made to backup the cluster resource by its ip address, an error would be returned: "No TSAs Found". The physical node could be backed up fine. The physical node was being registered with slp correctly.
Resolution
The /etc/opt/novell/ncs/nodename file was corrupt. Someone had edited this file manually, and whatever editor they used had a put a carriage return on the end of the nodename. When the ncs resource comes online the nodename in this file is checked to see if it matches what is in edirectory. The carriage return caused an invalid node name to be passed into the slp registration process. It's best not to modify this file. The original name of the node can be reinserted into this file thus:
echo -n $HOSTNAME > /etc/opt/novell/ncs/nodename
This will make sure that no return is on the end of the nodename.
Ncs clustering then needs to be restarted on the node for the change to take affect. Rebooting the box would work too.
Ncs clustering then needs to be restarted on the node for the change to take affect. Rebooting the box would work too.