Planning for vFiler unit creation
Prerequisites for vFiler unit creation
The following prerequisites must be met before you can create a vFiler unit:
- You must create at least one unit of storage (qtrees or volumes--traditional or flexible) before creating the vFiler unit.
- The storage unit for vFiler unit configuration information must be writable. It must not be a read-only file system, such as the destination volume or qtree in a SnapMirror relationship.
- The IP address used by the vFiler unit must not be configured when you create the vFiler unit.
Storage guidelines
When deciding which storage units to assign to vFiler units, remember the following guidelines:
- The first storage unit assigned to the vFiler unit contains the vFiler unit configuration information. It is called the primary storage unit. Although you can remove storage units from a vFiler unit at any time after the vFiler unit is created, the primary storage unit must remain for as long as the vFiler unit exists.
- The primary storage unit has the same security characteristics it had before it was transferred to the vFiler unit. When you create a new vFiler unit, C$ share-level permissions are restricted to administrators only, but file-level security is not modified. The vFiler unit administrator can set more restrictive file-level permissions.
- If the qtree or volume to be used as the primary storage unit contains an /etc directory, the data in the directory is lost after you add the qtree or volume to a vFiler unit. Data in qtrees and volumes used as non-primary storage units is preserved.
- A volume assigned to a vFiler unit must not be the storage system's root volume. You can, however, assign qtrees in the root volume to a vFiler unit.
- A volume assigned to a vFiler unit can be a traditional volume or a FlexVolTM volume. You cannot assign aggregates to a vFiler unit.
For information about traditional volumes, FlexVol volumes, and aggregates, see the Data ONTAP Storage Management Guide.
- FlexCache volumes can be created on the default vFiler unit, vfiler0, but cannot be assigned or moved to any other vFiler unit.
- If the vFiler unit administrator needs to create qtrees on the vFiler unit, assign volumes instead of qtrees to the vFiler unit when creating the vFiler unit. This is because qtrees can be created only at the root of a volume.
- If you anticipate that you might have to move the disks that are used for the vFiler unit's storage from one storage system to another, assign volumes, not qtrees, to the vFiler unit.
Note
You can move traditional volumes and FlexVol volumes between vFiler units; however, you cannot move FlexCache volumes between vFiler units.
- When managing NFS exports, CIFS shares, quotas, and options, vFiler unit administrators need to enter the complete path names of the storage resources used by the vFiler units in commands and configuration files. Therefore, choose volume and qtree names appropriately, so that you can share with the vFiler unit administrators the complete path names beginning with filer_name:/vol/vol_name.
Naming guidelines
When deciding the vFiler unit name, remember the following guidelines:
- The name can contain up to 31 alphanumeric ASCII characters.
- The name can contain the dash (-) and the underscore (_) characters, but the dash character must not be the first character.
- The name is case-insensitive.
- The name is unique.
You might also want to incorporate the name of the physical storage system in the vFiler unit name to enable you to easily determine the physical storage system with which a vFiler unit is associated; for example, mycompanyss1_vfiler1.
- The name vFiler0 is reserved.
Language guideline
vFiler unit administrators need to edit the /etc/quotas and /etc/usermap.cfg files for their vFiler units. These files support Unicode and root volume UNIX encoding. To ensure that vFiler unit administrators who do not use Unicode-capable editors can edit the files, create vFiler units on a storage system whose root volume language can be used for editing.
Quotas guideline
Creating a vFiler unit involves changing the ownership of a volume or qtree from the hosting storage system to a specified vFiler unit. This change requires that quotas be turned off for the affected volume. You can turn quotas back on for the volume after the vFiler unit is created.
Clustering considerations
Remember the following considerations for establishing vFiler units on a cluster:
- You can create up to 32 vFiler units on each member of a cluster, depending on the memory capacity of the hosting storage systems. See Adjusting the vFiler unit limit for details.
- Each member of a cluster must have a MultiStore license to take over its partner with a MultiStore license.
- The vFiler units hosted by the storage systems of the cluster are created and configured independently. That is, each storage system can host a different number of vFiler units, and the vFiler unit configurations on the storage systems can be different from each other.
- In takeover mode, the functioning storage system takes over all vFiler units created on the failed storage system. These vFiler units include the vFiler units you create and the unit called vFiler0. Therefore, for vFiler units on the failed storage system to work correctly after the takeover, each network interface used by a vFiler unit in a cluster must have a partner interface.
SAN considerations
When you create vFiler units on a storage area network (SAN), keep the following LUN and iSCSI considerations in mind:
Copyright © 1994-2005, Network Appliance, Inc. All rights
reserved.
Part No. 210-01078-B0
File: basicad4.htm
Last updated: 08-Dec-2005 13:31:46