12.7 Configuring an Unload Script for the Shared NSS Pool

A cluster resource unload script is automatically generated for the pool when you cluster-enable it. You can modify the script as needed to suit your needs by using iManager. For information about how to access the scripts for a cluster resource, see Section 10.6, Configuring an Unload Script for a Cluster Resource.

Newly created pool cluster resources use the following sequence for commands in an unload script. The order is reversed in a load script.

  • [CIFS]
  • NCP
  • Secondary IP address
  • [NFS] (If NFS is used, you manually add the command here.)
  • NSS (pool deactivation, which automatically dismounts the volumes)

If you modify the unload script, you must take the pool cluster resource offline, then bring it online for the changes to take effect.

12.7.1 Sample NSS Unload Script

The following sample values are used in the NSS unload script below:

Variable

Sample Value

Cluster resource’s virtual server name

NCS1-SHPOOL43-SERVER

Resource IP address

10.10.10.43

Pool name

SHPOOL43

Volume name

SHVOL43

Volume ID

252 (valid values are 0 to 254)

 
#!/bin/bash
. /opt/novell/ncs/lib/ncsfuncs

ignore_error ncpcon unbind --ncpservername=NCS1-SHPOOL43-SERVER --ipaddress=10.10.10.43

ignore_error del_secondary_ipaddress 10.10.10.43

ignore_error nss /pooldeact=SHPOOL43

exit 0

12.7.2 Using Volume Dismount Commands

By default, the pool deactivation automatically dismounts the volumes on the pool. The ncpcon dismount commands are not automatically added to the unload script. However, you can add ncpcon dismount commands for each volume before the pool deactivation line. For example:

ignore_error ncpcon dismount SHVOL43
ignore_error nss /pooldeact=SHPOOL43

Adding the ncpcon dismount command provides predictable and clean dismount behavior. It allows the volume dismount to occur and be logged independently of the pool deactivation, which can help with troubleshooting dismount issues. However, the ncpcon dismount commands are not automatically maintained and updated if you change a volume name or delete a volume. You must modify the unload script yourself.

12.7.3 Unloading OES CIFS as an Advertising Protocol

If OES CIFS is enabled as an advertising protocol for the pool, the following line is automatically added above the unbind command:

ignore_error novcifs 
   --remove '--vserver=".cn=CLUSTER-POOL1-SERVER.o=novell.t=AVALON_TREE."' 
   --ip-addr=10.10.10.43