raronson Super Contributor.
Super Contributor.

ncp volumes on a cluster resource with posix-permissions

I've been trying to create an ncp share on an ext3 cluster resource with the Inherit_POSIX_permissions=on switch with no success. It's making a total mess of my cluster while I try and work it out. This is my resource load script. I've tried a bunch of different approaches that fail horribly. It appears that the only way it works is if the volume is created without a volume id on my server, then dismounted and remounted. For a time I thought the volume needed to be defined in /etc/opt/novell/ncpserv.conf but its more like it needs to be defined through nrm or something like that. I can't put my finger on waht makes it work consistently.

Any ideas?

. /opt/novell/ncs/lib/ncsfuncs

# define the IP address
# define the file system type
#define the container name
# define the device
# define the mount point
# define the ncp volume name

#activate the container
exit_on_error activate_evms_container $container_name $MOUNT_DEV $NCS_TIMEOUT

# mount the file system
exit_on_error mount_fs $MOUNT_DEV $MOUNT_POINT $MOUNT_FS

# add the IP address
exit_on_error add_secondary_ipaddress $RESOURCE_IP

# run the iPrint relocate script
exit_on_error /opt/novell/iprint/bin/iprint_relocate $MOUNT_POINT

# start the iPrint driver store & printer manager
exit_on_error /opt/novell/iprint/bin/ncs_iprint_svc_start

# Start up NFS server
/etc/init.d/nfsserver start

# Create the ncp volume name
exit_on_error ncpcon mount $NCP_VOLUME=248,PATH=/media/iprint

# Mount with posix permissions

ncpcon dismount nfs1
ncpcon change volume nfs1 Inherit_POXIS_Permissions=on
mount nfs1

# Create the NCP server and bind to IP address
exit_on_error ncpcon bind --ncpservername=HVMC1_NFS1_SERVER --ipaddress=$RESOURCE_IP

# return status
exit 0
Labels (1)
2 Replies
raronson Super Contributor.
Super Contributor.

Re: ncp volumes on a cluster resource with posix-permissions

I may have found the solution. It
appears to work as long as I don't assign a volume ID. It likes to default so far to a volume id of 2.

That worries me a lot. If I use it this way, when my resource fails over it may migrate to another host with a volume already mounted with that ID. It would be easy for someone to to setup an ncpshare without knowing what it will break. Then my users will not be able to reconnect. This is such a pain.
Anonymous_User Absent Member.
Absent Member.

Re: ncp volumes on a cluster resource with posix-permissions

Hi Raronson:

I have a question for you...when you create an NCP share, it creates a volume object in e-directory. I have been successful in that, and have also been successful in failing the volume over, starting stopping, etc in the cluster.

In addition, I've been successful in creating a virtual server object and binding the IP of the resource to it. When I fail over the resource, I can access the contents of the volume through the virtual server.

HOWEVER...I have NOT been able to bind the resource to the volume object which is created when the NCP share is created. When I try, (exit_on_error ncpcon bind --ncpservername=$ncp_volume --ipaddress=$resource_ip) the resource goes comotose. Novells Doc suggests this is possible (page 91 of the NCP Server for Linux Administration Guide), and that when you fail a resource to another node, you should be able to access the contents of the volume directly through the volume object in eDirectory no matter what node it's sitting on.

Can you tell me what your experience has been in this? I may have to settle for the NCP Virtual Server Object, but would like to have a useable volume object if possible.


The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.