xfsrestore(8) System Manager's Manual xfsrestore(8)
xfsrestore - XFS filesystem incremental restore utility
xfsrestore [ options ] -f source [ -f source ... ] dest
xfsrestore [ options ] - dest
xfsrestore -I [ subopt=value ... ]
xfsrestore restores filesystems from dumps produced by xfsdump(8). Two modes of operation are available: simple and cumulative.
The default is simple mode. xfsrestore populates the specified destination directory, dest, with the files contained in the dump media.
The -r option specifies the cumulative mode. Successive invocations of xfsrestore are used to apply a chronologically ordered sequence of
delta dumps to a base (level 0) dump. The contents of the filesystem at the time each dump was produced is reproduced. This can involve
adding, deleting, renaming, linking, and unlinking files and directories.
A delta dump is defined as either an incremental dump (xfsdump -l option with level > 0) or a resumed dump (xfsdump -R option). The deltas
must be applied in the order they were produced. Each delta applied must have been produced with the previously applied delta as its base.
xfsrestore keeps state information in the xfsrestorehousekeepingdir, to inform subsequent invocations when used in cumulative mode, or in
the event a restore is interrupted. To ensure that the state information can be processed, a compatible version of xfsrestore must be used
for each subsequent invocation. Additionally, each invocation must run on a system of the same endianness and page size.
The options to xfsrestore are:
Each invocation of xfsrestore creates a directory called xfsrestorehousekeepingdir. This directory is normally created directly under
the dest directory. The -a option allows the operator to specify an alternate directory, housekeeping, in which xfsrestore creates
the xfsrestorehousekeepingdir directory. When performing a cumulative (-r option) restore or resuming (-R option) a restore, each
successive invocation must specify the same alternate directory.
Specifies the blocksize, in bytes, to be used for the restore. For other drives such as DAT or 8 mm , the same blocksize used for the
xfsdump operation must be specified to restore the tape. The default block size is 1Mb.
Use the specified program to alert the operator when a media change is required. The alert program is typically a script to send a
mail or flash a window to draw the operator's attention.
-e Prevents xfsrestore from overwriting existing files in the dest directory.
-f source [ -f source ... ]
Specifies a source of the dump to be restored. This can be the pathname of a device (such as a tape drive), a regular file or a
remote tape drive (see rmt(8)). This option must be omitted if the standard input option (a lone - preceding the dest specification)
-i Selects interactive operation. Once the on-media directory hierarchy has been read, an interactive dialogue is begun. The operator
uses a small set of commands to peruse the directory hierarchy, selecting files and subtrees for extraction. The available commands
are given below. Initially nothing is selected, except for those subtrees specified with -s command line options.
ls [arg] List the entries in the current directory or the specified directory, or the specified non-directory file entry. Both
the entry's original inode number and name are displayed. Entries that are directories are appended with a `/'.
Entries that have been selected for extraction are prepended with a `*'.
cd [arg] Change the current working directory to the specified argument, or to the filesystem root directory if no argument is
pwd Print the pathname of the current directory, relative to the filesystem root.
add [arg] The current directory or specified file or directory within the current directory is selected for extraction. If a
directory is specified, then it and all its descendents are selected. Entries that are selected for extraction are
prepended with a `*' when they are listed by ls.
delete [arg] The current directory or specified file or directory within the current directory is deselected for extraction. If a
directory is specified, then it and all its descendents are deselected. The most expedient way to extract most of the
files from a directory is to select the directory and then deselect those files that are not needed.
extract Ends the interactive dialogue, and causes all selected subtrees to be restored.
quit xfsrestore ends the interactive dialogue and immediately exits, even if there are files or subtrees selected for
help List a summary of the available commands.
-m Use the minimal tape protocol. This option cannot be used without specifying a blocksize to be used (see -b option above).
Allows xfsrestore to restore only files newer than file. The modification time of file (i.e., as displayed with the ls -l command) is
compared to the inode modification time of each file on the source media (i.e., as displayed with the ls -lc command). A file is
restored from media only if its inode modification time is greater than or equal to the modification time of file.
-o Restore file and directory owner/group even if not root. When run with an effective user id of root, xfsrestore restores owner and
group of each file and directory. When run with any other effective user id it does not, unless this option is specified.
Causes progress reports to be printed at intervals of interval seconds. The interval value is approximate, xfsrestore will delay
progress reports to avoid undue processing overhead.
-q Source tape drive is a QIC tape. QIC tapes only use a 512 byte blocksize, for which xfsrestore must make special allowances.
-r Selects the cumulative mode of operation. The -a and destination options must be the same for each invocation.
Specifies a subtree to restore. Any number of -s options are allowed. The restore is constrained to the union of all subtrees speci-
fied. Each subtree is specified as a pathname relative to the restore dest. If a directory is specified, the directory and all files
beneath that directory are restored.
-t Displays the contents of the dump, but does not create or modify any files or directories. It may be desirable to set the verbosity
level to silent when using this option.
Specifies the level of detail used for messages displayed during the course of the restore. The verbosity argument can be passed as
either a string or an integer. If passed as a string the following values may be used: silent, verbose, trace, debug, or nitty. If
passed as an integer, values from 0-5 may be used. The values 0-4 correspond to the strings already listed. The value 5 can be used to
produce even more verbose debug output.
The first form of this option activates message logging across all restore subsystems. The second form allows the message logging
level to be controlled on a per-subsystem basis. The two forms can be combined (see the example below). The argument subsys can take
one of the following values: general, proc, drive, media, inventory, and tree.
For example, to restore the root filesystem with tracing activated for all subsystems:
# xfsrestore -v trace -f /dev/tape /
To enable debug-level tracing for drive and media operations:
# xfsrestore -v drive=debug,media=debug -f /dev/tape /
To enable tracing for all subsystems, and debug level tracing for drive operations only:
# xfsrestore -v trace,drive=debug -f /dev/tape /
-A Do not restore extended file attributes. When restoring a filesystem managed within a DMF environment this option should not be used.
DMF stores file migration status within extended attributes associated with each file. If these attributes are not preserved when the
filesystem is restored, files that had been in migrated state will not be recallable by DMF. Note that dumping of extended file
attributes is also optional.
-B Change the ownership and permissions of the destination directory to match those of the root directory of the dump.
-D Restore DMAPI (Data Management Application Programming Interface) event settings. If the restored filesystem will be managed within
the same DMF environment as the original dump it is essential that the -D option be used. Otherwise it is not usually desirable to
restore these settings.
-E Prevents xfsrestore from overwriting newer versions of files. The inode modification time of the on-media file is compared to the
inode modification time of corresponding file in the dest directory. The file is restored only if the on-media version is newer than
the version in the dest directory. The inode modification time of a file can be displayed with the ls -lc command.
-F Inhibit interactive operator prompts. This option inhibits xfsrestore from prompting the operator for verification of the selected
dump as the restore target and from prompting for any media change.
-I Causes the xfsdump inventory to be displayed (no restore is performed). Each time xfsdump is used, an online inventory in
/var/lib/xfsdump/inventory is updated. This is used to determine the base for incremental dumps. It is also useful for manually
identifying a dump session to be restored (see the -L and -S options). Suboptions to filter the inventory display are described
-J Inhibits inventory update when on-media session inventory encountered during restore. xfsrestore opportunistically updates the online
inventory when it encounters an on-media session inventory, but only if run with an effective user id of root and only if this option
is not given.
Specifies the label of the dump session to be restored. The source media is searched for this label. It is any arbitrary string up
to 255 characters long. The label of the desired dump session can be copied from the inventory display produced by the -I option.
Insert the options contained in options_file into the beginning of the command line. The options are specified just as they would
appear if typed into the command line. In addition, newline characters (
) can be used as whitespace. The options are placed before
all options actually given on the command line, just after the command name. Only one -O option can be used. Recursive use is
ignored. The destination directory cannot be specified in options_file.
-Q Force completion of an interrupted restore session. This option is required to work around one specific pathological scenario. When
restoring a dump session which was interrupted due to an EOM condition and no online session inventory is available, xfsrestore cannot
know when the restore of that dump session is complete. The operator is forced to interrupt the restore session. In that case, if
the operator tries to subsequently apply a resumed dump (using the -r option), xfsrestore refuses to do so. The operator must tell
xfsrestore to consider the base restore complete by using this option when applying the resumed dump.
-R Resume a previously interrupted restore. xfsrestore can be interrupted at any time by pressing the terminal interrupt character (see
stty(1)). Use this option to resume the restore. The -a and destination options must be the same.
Specifies the session UUID of the dump session to be restored. The source media is searched for this UUID. The UUID of the desired
dump session can be copied from the inventory display produced by the -I option.
-T Inhibits interactive dialogue timeouts. xfsrestore prompts the operator for media changes. This dialogue normally times out if no
response is supplied. This option prevents the timeout.
Specifies a subtree to exclude. This is the converse of the -s option. Any number of -X options are allowed. Each subtree is speci-
fied as a pathname relative to the restore dest. If a directory is specified, the directory and all files beneath that directory are
Specify I/O buffer ring length. xfsrestore uses a ring of input buffers to achieve maximum throughput when restoring from tape
drives. The default ring length is 3. However, this is only supported when running multi-threaded which has not been done for Linux
yet - making this option benign.
- A lone - causes the standard input to be read as the source of the dump to be restored. Standard input can be a pipe from another
utility (such as xfsdump(8)) or a redirected file. This option cannot be used with the -f option. The - must follow all other
options, and precede the dest specification.
The dumped filesystem is restored into the dest directory. There is no default; the dest must be specified.
A base (level 0) dump and an ordered set of delta dumps can be sequentially restored, each on top of the previous, to reproduce the con-
tents of the original filesystem at the time the last delta was produced. The operator invokes xfsrestore once for each dump. The -r
option must be specified. The dest directory must be the same for all invocations. Each invocation leaves a directory named xfsrestore-
housekeeping in the dest directory (however, see the -a option above). This directory contains the state information that must be communi-
cated between invocations. The operator must remove this directory after the last delta has been applied.
xfsrestore also generates a directory named orphanage in the dest directory. xfsrestore removes this directory after completing a simple
restore. However, if orphanage is not empty, it is not removed. This can happen if files present on the dump media are not referenced by
any of the restored directories. The orphanage has an entry for each such file. The entry name is the file's original inode number, a
".", and the file's generation count modulo 4096 (only the lower 12 bits of the generation count are used).
xfsrestore does not remove the orphanage after cumulative restores. Like the xfsrestorehousekeeping directory, the operator must remove it
after applying all delta dumps.
A dump consists of one or more media files contained on one or more media objects. A media file contains all or a portion of the filesys-
tem dump. Large filesystems are broken up into multiple media files to minimize the impact of media dropouts, and to accommodate media
object boundaries (end-of-media).
A media object is any storage medium: a tape cartridge, a remote tape device (see rmt(8)), a regular file, or the standard input (currently
other removable media drives are not supported). Tape cartridges can contain multiple media files, which are typically separated by (in
tape parlance) file marks. If a dump spans multiple media objects, the restore must begin with the media object containing the first media
file dumped. The operator is prompted when the next media object is needed.
Media objects can contain more than one dump. The operator can select the desired dump by specifying the dump label (-L option), or by
specifying the dump UUID (-S option). If neither is specified, xfsrestore scans the entire media object, prompting the operator as each
dump session is encountered.
The inventory display (-I option) is useful for identifying the media objects required. It is also useful for identifying a dump session.
The session UUID can be copied from the inventory display to the -S option argument to unambiguously identify a dump session to be
Dumps placed in regular files or the standard output do not span multiple media objects, nor do they contain multiple dumps.
Each dump session updates an inventory database in /var/lib/xfsdump/inventory. This database can be displayed by invoking xfsrestore with
the -I option. The display uses tabbed indentation to present the inventory hierarchically. The first level is filesystem. The second
level is session. The third level is media stream (currently only one stream is supported). The fourth level lists the media files
sequentially composing the stream.
The following suboptions are available to filter the display.
(where n is 1, 2, or 3) limits the hierarchical depth of the display. When n is 1, only the filesystem information from the inventory
is displayed. When n is 2, only filesystem and session information are displayed. When n is 3, only filesystem, session and stream
information are displayed.
(where n is the dump level) limits the display to dumps of that particular dump level.
The display may be restricted to media files contained in a specific media object.
(where value is a media ID) specifies the media object by its media ID.
(where value is a media label) specifies the media object by its media label.
Similarly, the display can be restricted to a specific filesystem.
(that is, [hostname:]pathname), identifies the filesystem by mountpoint. Specifying the hostname is optional, but may be useful in a
clustered environment where more than one host can be responsible for dumping a filesystem.
identifies the filesystem by filesystem ID.
(that is, [hostname:]device_pathname) identifies the filesystem by device. As with the mnt filter, specifying the hostname is
More than one of these suboptions, separated by commas, may be specified at the same time to limit the display of the inventory to those
dumps of interest. However, at most four suboptions can be specified at once: one to constrain the display hierarchy depth, one to con-
strain the dump level, one to constrain the media object, and one to constrain the filesystem.
For example, -I depth=1,mobjlabel="tape 1",mnt=host1:/test_mnt would display only the filesystem information (depth=1) for those filesys-
tems that were mounted on host1:/test_mnt at the time of the dump, and only those filesystems dumped to the media object labeled "tape 1".
Dump records may be removed (pruned) from the inventory using the xfsinvutil program.
An additional media file is placed at the end of each dump stream. This media file contains the inventory information for the current dump
session. If the online inventory files in /var/lib/xfsdump/inventory are missing information for the current dump session, then the inven-
tory information in the media file is automatically added to the files in /var/lib/xfsdump/inventory. If you wish to incorporate the
inventory information from the media file without restoring any data, you may do so using the -t option:
# xfsrestore -t -f /dev/tape
This is useful to rebuild the inventory database if it is ever lost or corrupted. The only caveat is that xfsrestore needs to read through
the entire dump in order to reach the inventory media file. This could become time consuming for dump sessions with large media files.
xfsdump is tolerant of media errors, but cannot do error correction. If a media error occurs in the body of a media file, the filesystem
file represented at that point is lost. The bad portion of the media is skipped, and the restoration resumes at the next filesystem file
after the bad portion of the media.
If a media error occurs in the beginning of the media file, the entire media file is lost. For this reason, large dumps are broken into a
number of reasonably sized media files. The restore resumes with the next media file.
When xfsdump dumps a filesystem with user quotas, it creates a file in the root of the dump called xfsdump_quotas. xfsrestore can restore
this file like any other file included in the dump. This file can be processed by the restore command of xfs_quota(8) to reactivate the
quotas. However, the xfsdump_quotas file contains information which may first require modification; specifically the filesystem name and
the user ids. If you are restoring the quotas for the same users on the same filesystem from which the dump was taken, then no modifica-
tion will be necessary. However, if you are restoring the dump to a different filesystem, you will need to:
- ensure the new filesystem is mounted with the quota option
- modify the xfsdump_quotas file to contain the new filesystem name
- ensure the uids in the xfsdump_quotas file are correct
Once the quota information has been verified, the restore command of xfs_quota (8) can be used to apply the quota limits to the filesystem.
Group and project quotas are handled in a similar fashion and will be restored in files called xfsdump_quotas_group and xfsdump_quo-
To restore the root filesystem from a locally mounted tape:
# xfsrestore -f /dev/tape /
To restore from a remote tape, specifying the dump session id:
# xfsrestore -L session_1 -f otherhost:/dev/tape /new
To restore the contents a of a dump to another subdirectory:
# xfsrestore -f /dev/tape /newdir
To copy the contents of a filesystem to another directory (see xfsdump(8)):
# xfsdump -J - / | xfsrestore -J - /new
dump inventory database
rmt(8), xfsdump(8), xfsinvutil(8), xfs_quota(8), attr_set(2).
The exit code is 0 on normal completion, and non-zero if an error occurred or the restore was terminated by the operator.
For all verbosity levels greater than 0 (silent) the final line of the output shows the exit status of the restore. It is of the form:
xfsdump: Restore Status: code
Where code takes one of the following values: SUCCESS (normal completion), INTERRUPT (interrupted), QUIT (media no longer usable), INCOM-
PLETE (restore incomplete), FAULT (software error), and ERROR (resource error). Every attempt will be made to keep both the syntax and the
semantics of this log message unchanged in future versions of xfsrestore. However, it may be necessary to refine or expand the set of exit
codes, or their interpretation at some point in the future.
Pathnames of restored non-directory files (relative to the dest directory) must be 1023 characters (MAXPATHLEN) or less. Longer pathnames
are discarded and a warning message displayed.
There is no verify option to xfsrestore. This would allow the operator to compare a filesystem dump to an existing filesystem, without
actually doing a restore.
The interactive commands (-i option) do not understand regular expressions.
When the minimal rmt option is specified, xfsrestore applies it to all remote tape sources. The same blocksize (specified by the -b option)
is used for all these remote drives.
xfsrestore uses the alert program only when a media change is required.
Cumulative mode (-r option) requires that the operator invoke xfsrestore for the base and for each delta to be applied in sequence to the
base. It would be better to allow the operator to identify the last delta in the sequence of interest, and let xfsrestore work backwards
from that delta to identify and apply the preceding deltas and base dump, all in one invocation.