NSS supports moves of devices containing NSS volumes from a NetWare 6.0 server to an OES SP1 Linux or later server. For information, see Section 6.5.4, Guidelines for Moving Devices Cross-Platform.
To preserve the NSS pool and volumes on the device when you move it, you must modify the volumes’ Storage objects in eDirectory. You “decommission” the volume by removing its related Storage object from eDirectory for the original server. You “recommission” the volume by creating a new Storage object in eDirectory for the destination server. When moving clustered devices cross-platform, such as in a mixed cluster configuration, Novell Cluster Services automatically manages the Storage object updates to eDirectory.
NOTE:The “decommision” and “recommission” terminology is used only to illustrate the process; it does not represent a particular technology or tool.
IMPORTANT:Similar handling is necessary when moving devices with NSS pools between any two OES servers (NetWare - Linux, NetWare - NetWare, or Linux - Linux).
This section describes how to move devices cross platform from a NetWare 6.0 server to an OES SP1 Linux or later server:
The prerequisites in this section apply to moving multiple devices from a NetWare 6.0 server to an OES SP1 Linux server.
IMPORTANT:When moving a non-clustered device, you must also move any other devices that contribute segments to the NSS pools on the device you are moving.
Before you begin, make sure you understand the Section 6.5, Compatibility Issues for Using NSS Cross-Platform.
NSS must be installed on an EVMS-managed device on the destination server. For information, see
Installing Linux with EVMS as the Volume Manager of the System Device
in the
OES Linux Installation Guide.
Before or after you move an NSS volume from NetWare 6.0 to OES SP1 Linux, you need to set up file access for users on the OES Linux server.
The original server and the destination server can be in the same or different eDirectory trees.
If the destination server is in the same tree as the original server, the file system trustees and trustee rights continue to work after the move.
If the destination server is in a different tree, use eDirectory to enable or reassign affected users for access in the destination tree. For information, see the Novell eDirectory 8.7.3 Administration Guide .
To provide access for users on the OES Linux server, do one or more of the following, depending on your network environment:
NCP Server and Services: Install and configure NCP Server to allow the users to access the volume with the Novell Client or other NCP services. For information, see the NCP Server for Linux Administration Guide .
Other Protocols and Services: Install and configure other protocols, such as Samba (CIFS), NFS, or AFP, and enable users with Linux User Management to allow the users to access the volume with the non-NCP protocols and to execute Linux commands and services on the volume.
In this case, if you want to preserve file ownership, such as for configuring and enforcing user space quotas after the move, you must perform this configuration before you move the devices. For information about enabling users for Linux, see the Novell Linux User Management Technology Guide .
For guidelines about users and access, see Section 6.3, Access Control Issues for NSS on OES Linux.
For each NSS pool, decommission the pool and its volumes from the original server.
On the NetWare 6.0 server, if the device you want to move contains any mirrored partitions, you must first break the mirror (remove all but 1 segment) before upgrading.
If you use non-NCP protocols or Linux services for user access on the destination OES Linux server, you must Linux-enable the current users of the volumes before you move the devices if you want file ownership information to be usable after the move, such as for enforcing user space restrictions on a volume.
IMPORTANT:If you use only NCP Server and NCP services for user access, this step is not necessary.
Use one of the following methods to Linux-enable users of the volumes on the device where you want to preserve file ownership information:
To enable multiple users at once, use the nambulkadd command.
User IDs are automatically refreshed after the enabling process ends.
To enable a single user at a time, use iManager.
For information, see the Novell Linux User Management Technology Guide .
Deactivate the pool: At a server console, enter
nss /pooldeactivate=pool
Remove the eDirectory Storage objects for the NSS pool and each of its volumes.
In iManager, click
.
Click .
Specify the name and context of the object or objects you want to delete.
Click .
Repeat Step 2 and Step 3 for each pool on the devices you plan to move.
If you are using DFS in the tree where the original server is located, run the vldb repair command.
On the primary VLDB server, at the command prompt, enter
vldb repair
The VLDB Repair removes a GUID entry for each of the decommissioned volumes from the VLDB.
Remove or reallocate the devices from the original server. Depending on your storage configuration, this might require a server shutdown.
For each NSS pool, recommission the pool and its volumes on the destination server.
Relocate or reassign the devices to the destination server.
Reboot the destination server to mount the devices.
If a pool on the devices you moved is not automatically activated, activate the pool. At a server console, enter
nsscon
nss /poolactivate=pool
Create the eDirectory Storage objects for the NSS pool and each of its volumes.
In iManager, click
.
Click .
Browse to select the destination server.
Select the pool, then click .
In the lower right, select to view all volumes on the selected pool.
iManager opens to the Volumes page with the server and pool preselected.
For each volume in the selected pool, select the volume, then click .
Repeat Step 4.d through Step 4.f for each NSS pool and its volumes.
Allow the eDirectory tree to stabilize.
This can take several minutes.
Run the vldb repair command.
On the primary VLDB server, at the command prompt, enter
vldb repair
The VLDB Repair adds a GUID entry for each of the recommissioned volumes to the VLDB. Having the GUIDs in the VLDB allows you to create DFS junctions on NetWare servers that point to the volumes on the Linux machine. You cannot create a DFS junction on an NSS volume on Linux. If the device you moved contains an NSS volume with a DFS junction in it, you cannot follow the existing junction while the volume is on a Linux server.
Make sure the devices are up and working as expected on the Linux System.
For each device, you can optionally upgrade the format for its partitions (or segments, as specified in EVMSGUI) on the device, or leave the partitions in the old format.
Upgrading the partitions on the device to the NetWare 6.5 and later partition format is strongly recommended, but it is not necessary unless you want to mirror a partition on the device. Before you can create the mirror, you must upgrade the partition format of the partitions on the device you moved.
At a server console, open the EVMSGUI by entering
evmsgui
Right-click the segment, then select .
Repeat Step 8.b for every partition on the device that was moved to the OES Linux server.
Click to save the changes.
If you want to mirror the upgraded partition from Step 8, see Section 9.16, Mirroring an Existing Pool with NSSMU.