Hard links to files (NTFS volumes or UNIX)

A hard link is a directory entry for a file. Every file can be considered to have at least one hard link. On NTFS volumes or on UNIX systems, each file can have multiple hard links. Therefore, a single file can appear in many directories (or even in the same directory with different names). A Volume serial number (VSN) and a File Index indicates the actual file, unique on the volume. Collectively, the VSN and File Index are referred to as the file ID.

During a backup, if the backup selection list includes hard-linked files, the data is backed up only once. NetBackup uses the first file name reference that is found in the directory structure. If a subsequent file name reference is found, the reference is backed up as a link to the name of the first file. To back up subsequent references means that only one backup copy of the data is created, regardless of the number of multiple hard links.

If all hard-link references are restored, the hard-linked files continue to point to the same ID as the other files to which they are linked. However, if all the hard links are not restored, you can encounter anomalies as shown in the following examples.

Example 1

Assume that three hard links point to the same data. During a backup of Link2 and Link3, Link2 is encountered first and backed up. Then Link3 is backed up as a link to Link2. The three files are all hard-linked to the same data.

Figure: Example of hard links to the same data

Example of hard links to the same data

The original copies of Link2 and Link3 are backed up to tape, then deleted. Only Link1 is left on the disk.

Figure: Example of hard links backed up to tape and disk

Example of hard links backed up to tape and disk

During a subsequent restore, Link2 and Link3 are restored. The restored files, however, do not point to the same file ID as Link1. Instead, they are assigned a new file ID or inode number and the data is written to a new place on the disk. The data in the new location is an exact copy of what is in Link1. The duplication occurs because the backup does not associate Link2 and L3 with Link1.

Figure: Example of restored hard links

Example of restored hard links

Example 2

Assume that you attempt to restore only Link3. Here, NetBackup cannot link Link3 to Link2 because Link2 does not exist. The restore can complete only if it can link to Link2. A secondary restore request to the NetBackup server automatically restores Link2, which contains the data. Link2 can now be successfully restored.

More Information

About the Backup Selections tab

Changing backup selections for standard policies

Changing backup selections for database policies

Changing backup selections for Oracle or DB2 policies

Reducing backup time

Verifying the backup selections list

Path rules for Microsoft Windows file backups

Path rules for Windows disk image (Raw) backups

Path rules for Windows registry backup

Pathname rules for UNIX clients

About the path rules for NetWare NonTarget clients

Path rules for NetWare Target clients

Path rules for clients that run extension products

Backup selections list directives

Backup selections list directives for multiple data streams

Files that are excluded from backups by default