Administrator Guide

Variable Name Description Default
During backup, if this variable is set, an additional directory listing is added to the
backup data stream. Because of the additional process required, this addition
could aect the backup data stream size and performance.
During recovery, if this variable is set and if the backup data stream was
generated with this variable turned on, the NDMP server handles deleting les
and directories that are deleted between incremental backups.
Setting this variable requires additional processing time and enlarges the backup
data stream size (how much it changes depends on the number of elements in
the backup data set). If this feature is not important to the end user, it should not
be set.
BASE_DATE Used by TSM for token-based backup, as an alternative to using the LEVEL
environment variable.
When BASE_DATE is set to 00, a full backup is performed.
After a full backup completes, a token can be retrieved by retrieving the
DUMP_DATE environment variable. This token can then be passed in later
backups as the value of BASE_DATE. The backup performed in this case is an
incremental backup relative to the time when the token was generated.
When BASE_DATE is set to -1, token-based backup is disabled.
-1
DEREF_HARD_LINK Controls whether hard link les data content are backed up for all instances of
the same le. Valid values are Y and N.
N
Incremental Backups
Each time a backup is performed, the NDMP server stores the timestamp for the backup. When the NDMP server performs an
incremental backup, it uses the timestamp stored for the previous full or incremental backup to determine if a directory or le needs to be
included.
Both supported backup types (dump and tar) support incremental backup. The algorithm for traversing the backup target directory is the
same. However, because inode-based le history generation has dierent requirements to support DAR, the backup data stream generated
is dierent:
• dump: Each directory visited will be backed up and a le history entry will be generated. It does not matter whether the directory has
changed.
• tar: Backs up and generates a le history entry only for the directories that have changed.
Therefore, the amount of data backed up using a tar backup will be less than that of a dump backup. The size dierence depends on the
number of directories in the backup data set.
Handling Hard Links
NDMP backup handles hard link les in the most ecient way by default. That is, the hard link les’ data content will be backed up only
once. After the backup operation encounters the rst hard link le and backs up its content, the backup process remembers the inode
number of that le. Subsequently, when the backup operation encounters les with the same inode number, only the header is backed up.
When this backup data stream is restored, the hard link les will be recovered as hard link les.
This mode of backup could create a problem in the case of a selective restore when the selected les or directories to be restored contain
hard link les that are not the rst instance encountered during backup. In this case, the restore fails and an NDMP message is sent to the
DMA server indicating the rst instance of the le that should also be included in the selective restore.
452
FluidFS Administration