Snapshot (information technology)

from Wikipedia, the free encyclopedia

Under snapshot or checkpoint ( English snapshot ) is understood in the information technology is a snapshot of a system or object.

Distributed Systems

In the area of distributed systems , the term is used for a recording of the global status at a specific point in time. This can be achieved, for example, with a snapshot algorithm .


Any content on a computer screen is copied to the buffer using the "Print" button, also known as "PrtSc" (Print Screen). This is a snapshot (actually a screenshot ) of the screen content. After calling up a word processing or image editing program , the screen content can be inserted there from the buffer.

Snapshot of a website

The snapshot of a website in a browser is a special form of screenshot or screenshot. The screen content that is not directly visible or that can only be accessed by scrolling can also be included in the snapshot.

Version management

In computer programs, a snapshot is a program version that is released directly from the version management database and not as an official version ( "release" ). Usually these versions are generated from the current state of the version management of the source code or from the working copy of a developer's project.

Mass storage

A distinction is made between the redirect-on-write and copy-on-write processes .

With redirect-on-write , a snapshot is a special storage area that stores older or more recent versions of changed data. It does not contain a complete copy of the data, but is filled step-by-step with every change by redirecting all changes to the snapshot . If the snapshot is to be read now , it is first checked whether the part to be read is there. If this is the case, this part is used, otherwise the original file is read.

With copy-on-write file systems, only the top-level metastructure is copied when a snapshot is created. Since copy-on-write file systems never overwrite active data blocks, but create a new block with the changed data and delete the old one as soon as it is no longer referenced, the data and metadata blocks referenced in the snapshot branch are retained if they are saved in the Main branch, as they are still referenced by the snapshot. In this way, the snapshot is created automatically during normal operation.

In file systems

Snapshots can be implemented in the file system, where e.g. B. read access to older versions of the directory tree in a special directory.

Examples are the WAFL file system, UFS in FreeBSD 5, the Btrfs , ZFS and NTFS with the Volume Shadow Copy Service on Windows .

At the block level

Some disk arrays and logical volume managers can keep snapshots of data blocks on their virtual hard disks or logical volumes .


Data backups can be made from snapshots without having to rely on the consistency of the current database. Some applications and file systems also offer the option of putting the data in use in a defined state and delaying new write accesses until the underlying system has created a snapshot.

So z. For example, a database can be stopped during operation and a snapshot can be created. The database can then be started again immediately and the snapshot z. B. be secured elsewhere. Modern database systems, however, enable consistent data backups without stopping them and creating snapshots.

If a snapshot is accessible in the file system, users can e.g. B. restore it yourself after an accidental overwriting.

It is important that no write operation is active at the time the snapshot is created in order to avoid inconsistent data in the snapshot. Snapshots can only be used if the consistency of the current database is guaranteed.

See also