Convert RZ to TAR.BZ2 Online (Swapping One Bzip2-Based Format for Another)

Two Formats That Both End With Bzip2, Starting From Different Places

Rzip, written by Andrew Tridgell (also the creator of Samba and rsync) with its last stable release, version 2.1, dating to February 2006, compresses in two stages: a first pass matching repeated data across a window up to 900 MB, then a second pass running whatever remains through bzip2 itself. TAR.BZ2 is simpler by comparison — a plain tar archive with bzip2's own compression applied directly on top, no preceding long-range matching stage at all. Since rzip typically compresses an existing tar archive in the first place, converting RZ to TAR.BZ2 usually means decompressing the rzip layer to recover that tar archive, then recompressing it with bzip2 alone.

The practical question this conversion answers is whether rzip's extra matching stage was actually contributing anything for the specific data involved. If it was, dropping it in favor of plain bzip2 produces a larger TAR.BZ2 than the original TAR.RZ; if the source data didn't have much long-range repetition to begin with, the two can end up close in size, since bzip2's own block-sorting still handles ordinary, shorter-range redundancy on its own.

This conversion tends to come up when the .tar.bz2 extension is specifically what a downstream tool, build script, or package repository expects, rather than because rzip's own compression did anything wrong — plenty of Linux packaging systems and automated build pipelines check for a .tar.bz2 suffix directly, and a .rz or .tar.rz file simply won't match that check no matter how well it compressed the same underlying data.

What Rzip's First Stage Actually Adds on Top of Plain Bzip2

Bzip2 alone works in fixed blocks of roughly 900 KB, applying the Burrows-Wheeler transform and Huffman coding within each block independently — it has no mechanism for recognizing that a chunk of data near the start of a file matches another chunk 500 MB later, since each block is processed without reference to any other block. Rzip's first stage exists specifically to catch that exact case: its matching pass scans across the full window before bzip2 ever runs, replacing long-distance repeats with short references so bzip2 only has to compress what's left.

On the project's own benchmark using the Samba source tree, this combination reached a 9.55x compression ratio versus bzip2 alone's 4.78x on the same data — a difference the project attributes directly to that first-stage long-range matching, since the second stage in both cases is the same bzip2 algorithm working on different quantities of pre-processed data.

Because rzip's matching stage already removes the long-range duplication before bzip2 sees the data, rzip in practice can finish faster than running bzip2 alone on the original file, even though rzip is nominally doing more total work across its two stages — bzip2 simply processes a smaller, pre-shrunk stream by the time its turn comes.

What Converting RZ to TAR.BZ2 Gains and What It Costs

• Gain — near-universal tool support: bzip2 ships by default on virtually every Linux distribution and macOS, and Windows tools like 7-Zip and PeaZip open .tar.bz2 without extra configuration, unlike .rz which needs rzip specifically.
• Gain — a widely recognized file extension: .tar.bz2 is a common, documented convention across open-source software distribution; .rz remains an obscure extension outside Linux/Unix circles.
• Lose — rzip's long-range matching advantage: on files with repetition spread across hundreds of megabytes, the documented 9.55x vs. 4.78x ratio gap on the project's own benchmark shows real compression left on the table by skipping that stage.
• Lose — the original compression work already done: going from RZ to TAR.BZ2 means fully decompressing and then recompressing, discarding rzip's matching pass entirely rather than building on it.
• Keep — the same final entropy coder: both formats end with bzip2, so the underlying compression technique in the last stage is identical either way.
• Gain — a genuinely portable output for build pipelines: package repositories and automated scripts frequently look for a .tar.bz2 extension specifically, something the .rz or .tar.rz naming won't satisfy regardless of how the data compressed.

Which Tools Handle Each Side of This Conversion

Decompressing the source requires the rzip binary, GPL-licensed and packaged in Debian and Ubuntu's repositories, run as rzip -d filename.tar.rz to recover the underlying tar archive. From there, GNU tar's own -j flag (tar -cjf output.tar.bz2 recovered-directory) or the standalone bzip2 command handles the compression step to produce the final .tar.bz2 file — both are standard, preinstalled tools on virtually any Linux or macOS system.

On Windows, 7-Zip and PeaZip both create .tar.bz2 archives directly through their graphical interfaces, while PeaZip additionally lists native support for reading the original .rz format, making it a practical single-application route for handling both halves of this specific conversion without switching tools.

There's no dedicated Windows binary maintained by the rzip project itself, so decompressing the source .rz file on Windows depends specifically on PeaZip rather than the original author's own software, unlike bzip2, which has long-standing, widely available Windows ports from multiple sources.

On the rzip side, both the original rzip and PeaZip run entirely offline once installed, so decompressing an .rz file doesn't require any network access — a relevant point for anyone converting sensitive backup or log data that shouldn't leave a local machine during the process.

Real Problems Reported Converting Rzip Archives to TAR.BZ2

A recurring complaint involves someone converting a large backup file from RZ to TAR.BZ2 expecting a similar file size, then finding the result noticeably larger — the consistent explanation is that the original data had significant long-range repetition, the exact pattern rzip's 900 MB matching window was built to catch and plain bzip2's much smaller block size cannot see at all.

A second documented issue involves someone with an .rz file and no rzip installation available, discovering that few mainstream tools besides PeaZip can open it — a consequence of rzip staying a niche Linux/Unix utility after its last release in 2006, never achieving the broad archiver support that bzip2 itself gained over the same period.

A third reported pattern involves confusing rzip with its incompatible successor lrzip: a file compressed with one cannot be decompressed with the other despite the similar name, and attempting it produces an error rather than a usable tar archive to recompress.

A fourth pattern shows up in automated build scripts written around a fixed .tar.bz2 extension check, which fail silently or error out when pointed at a .rz or .tar.rz file instead — a maintenance issue tied entirely to the script's own narrow assumption about file naming, not to any real incompatibility between what the two formats actually contain once decompressed.

RZ and TAR.BZ2 Compared on the Details That Matter

Questions About Converting RZ Files to TAR.BZ2

Will the TAR.BZ2 file always be bigger than the original RZ file?
Often, but not always. It depends on how much long-range repetition the original data had — rzip's matching window catches redundancy spread across hundreds of megabytes that plain bzip2's smaller blocks simply cannot see.

Why do both formats use bzip2 if rzip is supposed to be better?
Rzip uses bzip2 as its second stage, after a first-stage matching pass has already removed long-range duplication. TAR.BZ2 uses bzip2 alone, with no such preprocessing step, which is exactly why the two formats can differ significantly in size on the same source data.

Is there a one-step tool that converts RZ directly to TAR.BZ2?
Not a dedicated one. The standard route decompresses with rzip first, then recompresses the recovered tar archive with bzip2 or GNU tar's own -j flag as a separate step.

Can I open an RZ file on Windows before converting it?
PeaZip, a free archiver for Windows, lists RZ among its supported formats. Mainstream tools like 7-Zip and WinRAR have no documented native support for it.

Does converting to TAR.BZ2 change any of the original file permissions?
No. Both rzip and bzip2 compress raw byte streams without touching tar's own header fields, so Unix permissions, ownership, and modification times recorded in the tar layer survive the conversion unchanged.

Does the bzip2 compression level matter when rebuilding the TAR.BZ2?
Yes. Bzip2 supports levels 1 through 9, controlling its block size up to the format's 900 KB maximum, with 9 as its own default — a higher level generally compresses better at the cost of more memory and processing time during the rebuild.