fragview

Created: 2011-07-18 21:39
Updated: 2018-10-31 13:34
License: other
c++

README.md

About

Files stored on disk usually divided into small chunks called blocks, which can reside anywhere on disk partition. When file blocks are not adjacent, one can say it's fragmented. There is nothing bad in fragmentation, it helps filesystem to use disk space more efficiently instead. But if file divided in too many fragments its read speed can decrease dramatically, especially if fragments located far one from another.

Fragview's main purpose is to visualize disk content by displaying its map. It displays bunch of small colored squares (called clusters), each representing a number of blocks. Square's color is white for unoccupied blocks and colored for occupied ones. If cluster contains part of fragmented file its color will be red, and for non-fragmented -- blue.

Fragview uses its own metric to determine fragmented file or not. It is based on estimation of worst read speed. Files whose read speed more than two times lower than read speed of continuous ones will be considered fragmented.

Compiling and installing

$ sudo apt-get install libboost-dev libgtkmm-3.0-dev libglibmm-2.4-dev libsqlite3-dev
$ mkdir build && cd build
$ cmake -DCMAKE_BUILD_TYPE=Release ..
$ make
$ sudo make install

Binaries

There are three binaries you can run:

  • fragview
  • fileseverity
  • fragdb

fragview is a GTK+ GUI application which enables you to browse by clicking and scrolling. fragdb is a command-line utility whose purpose is to collect number of fragments of files in specified directory recursively while storing its result in sqlite3 database. Then it is possible to make some simple reports like displaying ten most fragmented files or files having more than 100 fragments. fileseverity is similar to filefrag utility from e2fsprogs, but displays severity metric rather than fragment count. A few words about this metric can be found in draftpad.md file.

License

fragview is distributed under the terms of the MIT License. See LICENSE.MIT for full text.

Cookies help us deliver our services. By using our services, you agree to our use of cookies Learn more