Open a Unix dump backup on Windows and browse it as a folder tree. Load the whole set and the level 0 and its increments are layered into one view, each file marked with the level it came from. The restore command plays no part in the process.
Version 26.7 · Windows 11, Windows 10, Windows 8.1, Windows 7 and Windows Server
A dump set is a level 0 base plus its increments, so the whole sequence has to be read together before any single file resolves.
Inside the set
What the DUMP file extractor resolves from a level 0 and its increments, before any file leaves the set.
A level 0 base and the incrementals stacked on it are merged into a single view. The most recent version of every path is resolved across the sequence, with the source level shown beside each file.
The dump header names the source filesystem and the byte order of the machine that wrote it. ext2, ext3 and UFS are decoded on that basis, so a big endian set from old hardware opens the same as a little endian one.
Each directory entry in a dump carries its inode metadata. The owner, group, UID, GID and mode are surfaced against the file so the original ownership is visible before extraction, not reconstructed afterwards.
A dump records the inode type of every entry. Symlinks, hard links, block and character devices and named pipes are labelled by type in the listing instead of being flattened into blank placeholder files.
The directory map is written ahead of the file data in a dump. A set whose tail was lost still yields a full listing, and each entry is flagged by whether its data block survived.
No write ever touches the dump files. They are memory mapped read only for the duration and released untouched, which keeps a set usable as evidence or as a second attempt later.
The level 0 and its increments are indexed together, so the current version of any file resolves in a single pass.
Run the Free DemoGetting from a folder of level numbers to the file somebody asked you for.
Point the DUMP file extractor at the folder holding the archives. A level 0 is a snapshot of one moment. The file you want may have changed after it, so the increments beside it matter as much as the full.
Archives with no extension load the same way. The name was never the identifier.
Each archive opens with a header recording the filesystem it came from, the level it ran at, the date and the byte order of the machine. The restore command depends on these values. Here they are read directly off the file.
Old and new filesystem layouts, either byte order, UFS alongside ext.
Levels are sorted and layered so the newest copy of every path wins, which is what a real restore does across a chain. The result is one tree to look at rather than a folder to work through in order.
Works without /etc/dumpdates, since the levels and dates come out of the headers.
A filesystem dump runs to a great many inodes and you will not scroll it. Type part of a path and the list narrows across the whole set, showing which level each match came from.
The level column is usually the quickest way to separate two copies of the same path.
Only the selected entries are written, into an ordinary Windows folder. No volume is formatted, nothing is mounted and nothing on the disk is overwritten.
The archives are opened for reading. They are the same files afterwards.
If a level between your full and the date you want was lost, anything changed in that window is not in the set at all. The listing shows which levels are present so you can see the hole before you go looking for a file that was never there.
Sets found on tape are often incomplete. Better to know at the start.
Not a list of files. A filesystem taken apart at the inode level and written down in passes.
dump maps the filesystem before writing anything, then lays down directories and file contents in separate passes. Symbolic links and hard links get their own inode records along the way. The structure is reconstructed from those records rather than read off a table.
Every inode record carries its permission bits, numeric UID and GID and its timestamps. They are shown against each entry, since nothing on a Windows machine corresponds to a numeric owner.
dump only ever handled ext2, ext3 and the BSD UFS filesystems. It also refuses a subdirectory above level 0. What you hold is a partition rather than a folder somebody chose. Everything that was mounted there is in it.
Yes. The usual answer is to go and find a Unix box, which assumes you have one and that it matches.
DUMP was written to be read back by exactly one program. Every guide says the same thing, install restore and run it. Each of them quietly assumes a stack of things about your situation.
What restore expects
What you have
The gap is the whole problem. Nothing is wrong with the archives. The environment that made them readable was decommissioned years ago.
The first three are real and cost nothing. Each of them needs a Unix host you may not have.
| Route | No Unix needed |
Reads a whole set |
One file at a time |
Nothing to format |
|---|---|---|---|---|
| restore in list mode Free See what is inside, restore -t -f archive |
||||
| restore in interactive mode Free Its own shell with ls, cd, pwd, add, extract and help, restore -i -f archive |
||||
| restore by name or rebuild Free restore -x for one path, restore -r onto a volume freshly made with mke2fs |
||||
| DUMP file extractor Demo Loads the whole set, then writes out your selection |
Interactive mode is the answer when a Unix host is already available. It costs nothing. Two things catch people there. Paths have to be given with a leading dot and slash. Whatever you extract lands in the directory you started restore from. Its own manual warns that rebuild mode can be detrimental to your health, which is fair. It wants an empty volume, layers the levels on in order and leaves a restoresymtable behind for the next one. The DUMP extractor tool is for the case where there is no host to run any of it on.
Rarely from anything current. These arrive from the era when a filesystem and a tape drive were the whole backup strategy.
The BSDs shipped dump and restore as standard. Solaris had ufsdump and ufsrestore, close relatives under another name. HP-UX carried its own flavour again. A great many Linux servers were backed up with the ext version long after better options existed. It was the default answer for a decade because it understood the filesystem directly, needed no scripting and wrote straight to tape.
What arrives now is the sediment. A tape library read out before disposal, a drawer of cartridges from a decommissioned server, an archive folder inherited with a business. The machines are gone, the operator has retired and somebody needs one document out of it.
A dump set needs nothing above the entry tier.
Standard, $49
Two machines. DUMP at any level is covered, together with the entry tier archive formats.
Professional, $99
For Veeam, Acronis, ShadowProtect, tape and virtual disks. A dump set needs none of it.
Try the DUMP extractor software on the folder you were given. It sorts the levels, builds one tree and shows which archive each file came from, before anything is bought.
Buy Standard Compare all editions
Reflects Backup Extractor 26.7. BitRecover has been building data recovery and format conversion tools since 2011.
A dump set is not one backup. It is a full plus a run of increments. Your file sits in exactly one of them.
Level 0 captures the whole filesystem. Every level above it holds only what changed since the last dump at a lower number, so the scope shrinks as the number climbs. That makes the set compact and it makes finding one file awkward.
Which level ran when is recorded in /etc/dumpdates, back on whichever machine produced the set. That file lived on the server rather than with the backups, so it is almost never in the folder you were handed. The good news is that the header above carries the same facts, which is where they get read from instead.
This is why a single archive so often disappoints. People open the level 0, find a version of the file from weeks earlier and conclude the backup is useless. The newer copy was in an increment nobody thought to open.
What comes up once the folder turns out to hold several archives instead of one.
Nothing that ships with it. The format was only ever meant to be read back by restore, which expects a Unix host and a filesystem it recognises.
Levels run 0 through 9. Level 0 is the whole filesystem and every level above it carries only the changes made since a lower one ran, so a single file may sit in any of them.
Whichever level last touched it, which the DUMP extractor tool shows against every entry. Load them together in the DUMP file extractor and search across the set rather than opening each one hoping.
It is the log kept by the server doing the dumping, recording which level ran when. It explains a set of archives. It is almost never kept with them. Not having it is normal.
No. Search the listing, tick the entry and export it to a folder. Nothing has to be formatted and no filesystem gets rebuilt.
It matters to restore, which is fussy about filesystem type and byte order. The DUMP file extractor reads the header and works it out instead of asking you.
Each inode record carries them and the listing shows them beside the entry. Windows has nowhere to put a numeric owner, so they stay as readable detail.
Only in how it reaches you. Once the tape image is a file on disk it loads like any other, including a multi volume set written across several cartridges.
Usually. Directory information sits near the front, so a listing often survives a corrupt or partial tail even when the data behind it does not. You see what the listing contains before deciding anything.
Standard at $49, the same tier as the other Unix archive formats. Professional adds the commercial imaging formats you do not need for this.
Load the folder, let the levels sort themselves out and take the file you came for.
Version 26.7 · Windows 11, Windows 10, Windows 8.1, Windows 7 and Windows Server · 30 day money back