Cross volume links in Data Migrator
Data Migrator allows you to move data from primary to secondary storage based on policies that you define. When a file is migrated, a cross volume link is left, indicating the new location of the file's data. A cross volume link is a special file on the local file system that "points" to the file on a remote file system. The cross volume link stores the migrated file's metadata and is used to construct a handle for the file on the remote file system.
When a read request for a migrated file is received, the storage server reads the cross volume link, constructs the file handle, retrieves the file from the secondary storage, and sends the file to satisfy the read request. In this way, the read request is serviced and the client need not be aware that the file is not actually stored on the local file system.
Enhanced cross volume links (CVL-2) are supported on all NAS Platforms and offer the following benefits:
-
Remote file systems might be on a storage device attached to a remote server (not necessarily another NAS Platform) accessible through the NFSv3 protocol. This capability, called remote migration, allows the storage server to migrate files to a separate storage device, such as content archival or compliance products.
Remote migration to a Hitachi Content Platform (HCP) system, is supported through the HTTP or HTTPS (cloud only) protocol for new remote migration paths, but pre-existing remote migration paths that were created in releases that supported the NFSv3 protocol are still supported. You cannot, however, create paths to an HCP system using the NFSv3 protocol.
Access to files located on the external storage device is a licensed feature, requiring an external volume link (XVL) license and a Data Migrator license. See the Server and Cluster Administration Guide for information on adding a license key.
-
Local access to migrated file attributes increases performance for getattr/lookup/readdir+ requests.
For enhanced cross volume links (CVL-2), all file attributes for a migrated file are stored on the local file system. When an operation requires the attributes of a migrated file, the storage server responds with locally stored attribute information, which provides better performance.
In addition, local read caching can be used to improve performance when accessing migrated files. See the Server and Cluster Administration Guide for information on local read caching.
-
Quota management and CVL-2 links.
File-size-based quotas consider the size of a file to be its logical length (the length shown by 'ls -l') and as such ignore the fact that files may resize on an external location.
Usage-based quotas compute the size of a file from the number of local blocks that it currently uses. A file which has been externally migrated uses zero local (non-metadata) blocks, so will appear to a usage-based quota to be of size zero.