I have created a traditional linux reiser volume, and a corresponding NCP share in a SLES 10/OES2 clustered environment.

SO...I've got everything migrating/failing over properly including the NCP share. I have created a virtual NCP server object in the container, bound the IP of the resource to it and it's functioning just fine (meaning I can browse to that object, double click it, and no matter what node the volume is mounted on, that object displays the contents correctly.

HOWEVER, the volume object that is created in the eDirectory container when you initially create the NCP share is NOT linked to the NCP Volume, unless the volume happens to be sitting on the node it was created.

I have followed Novell's doc (page 91 of the NCP Server Administration Guide for Linux), and it suggests that you should be able to bind the volume to the IP of the resource to accomplish this (exit_on_error ncpcon bind --ncpservername=$NCP_VOLUME --ipaddress=$RESOURCE_IP in the load script). This does NOT work. What I mean by that is that it puts the resource into comotose state no matter which node it's sitting on.

What I want to know is, was this simply a typo on Novell's part...and is it even possible to have an NCP Volume object in eDirectory display the contents of an NCP share no matter which node the volume is sitting on, or am I going to have to just settle for the NCP Virtual Server object?

Any help would be appreciated...I've been banging my head off the wall over this long enough.