Zaloha2.sh

Zaloha2 is a small and simple directory synchronizer:

• Zaloha2 is a BASH script that uses only FIND, SORT and AWK.
• All you need is the Zaloha2.sh file (contains ~118 kB of program code and ~100 kB of documentation).
• No compilation, installation and configuration is required.
• Cyber-secure: No new binary code, no new open ports, no talking to any "uncles" on the Internet, easily reviewable.
• Three operation modes are available: the Local Mode, the Remote Source Mode and the Remote Backup Mode.
• Local Mode: Both directories are available locally (local HDD/SSD, flash drive, mounted Samba or NFS volume).
• Remote Source Mode: the source directory is on a remote source host that can be reached via SSH/SCP and the backup directory is available locally.
• Remote Backup Mode: the source directory is available locally and the backup directory is on a remote backup host that can be reached via SSH/SCP.
• Zaloha2 does not lock files while copying them. No writing on either directory may occur while Zaloha2 runs.
• Zaloha2 always copies whole files via the operating system's CP command or the SCP command (= no delta-transfer like in RSYNC).
• Zaloha2 is not limited by memory (metadata is processed as CSV files, no limits for huge directory trees).
• Zaloha2 has optional reverse-synchronization features.
• Zaloha2 can optionally compare the contents of files.
• Zaloha2 prepares scripts for case of eventual restore (can be optionally switched off to shorten the analysis phase).

Full documentation is available both online as well as inside of Zaloha2.sh.

On Linux/Unics, Zaloha2 runs natively. On Windows, Cygwin is needed.

What are synchronization programs good for ?

Read article here.

How does Zaloha2.sh work ?

Explained in five sentences (the Local Mode)

• FIND is executed to scan both directories to obtain CSV metadata about files and subdirectories.
• The CSV metadata is compared by a sequence of sorts and AWK processing steps.
• The results (= prepared synchronization actions) are presented to the user for confirmation.
• If the user confirms, the synchronization actions are executed.
• A non-interactive regime is available as well.

Explained by an Interactive Flowchart

Open Interactive JavaScript Flowchart here.

An insight into the CSV metadata is given here

Open CSV data model here.

Explained in full detail

Read the relevant sections in the Documentation.

Obtain Zaloha2.sh

The simplest way: Under the green button "Code" above, choose "Download ZIP". From the downloaded ZIP archive, extract Zaloha2.sh and make it executable (chmod u+x Zaloha2.sh).

For running the Simple Demo, extract also the scripts Simple_Demo_step1/2/3/4/5/6/7.sh and make them executable.

The integrity of Zaloha2.sh can be verified by its SHA-256 hash. The hash of current version is:

sha256sum Zaloha2.sh
099ebd75cf6c181217a74789ff1e531ef9c2975e3bcfc75275add4d6259225b8  Zaloha2.sh

Usage Examples

Basic Usage: Synchronize test_source to test_backup:

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup"

Basic usage with Colors (on terminals with ANSI escape codes):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --color

Besides the standard synchronization, if there are files in test_backup that are newer (younger) than the corresponding files in test_source, copy them back to test_source (Reverse-Update):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --revUp

Besides the standard synchronization, if there are files in test_backup that do not exist in test_source, and those files are newer (younger) than the last run of Zaloha2, copy them back to test_source (Reverse-New):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --revNew

Do Not Remove objects from test_backup (unless the removals are unavoidable), just add new files and subdirectories to test_backup (and/or update existing files in there):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --noRemove

Exclude subdirectories .git from the synchronization. If such subdirectories already exist in test_backup, they are considered obsolete (= removals are prepared):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" \
           --findSourceOps='( -type d -a -name .git ) -prune -o'

Exclude files with ending .NFO from the synchronization. If such files already exist in test_backup, they are considered obsolete (= removals are prepared):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" \
           --findSourceOps='( -type f -a -name *.NFO ) -o'

Generally Exclude files with ending .NFO from the synchronization (never do anything with them). Note that the option --findGeneralOps is used instead of --findSourceOps. The + sign tells Zaloha2 to use the passed FIND expression in addition to the internally defined defaults:

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" \
           --findGeneralOps='+( -type f -a -name *.NFO ) -o'

Disable the internally defined defaults for --findGeneralOps that exclude the well-known Linux and Windows "trash" and "lost+found" subdirectories: Use the option --findGeneralOps without the leading + sign to replace the internally defined defaults (here by an empty string):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" \
           --findGeneralOps=

A more complex example with several FIND expressions: Do not backup any subdirectories .git, .svn and .cache in test_source. If they already exist in test_backup, prepare their removals. These are also clear cases for --findSourceOps. Further, completely ignore (in both directories) any coredump files and the subdirectories mount and database located directly under test_source and test_backup. The reasons are that arbitrary external media can be mounted under mount and the backup of database is implemented by dedicated tools provided by the database vendor (also not on file level). These are again clear cases for --findGeneralOps. Additionally, preserve the internally defined defaults for --findGeneralOps (this is achieved by the leading + sign). Finally, for a good visual appearance, use the expression by expression method:

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup"      \
           --findSourceOps='( -type d -a -name .git ) -prune -o'    \
           --findSourceOps='( -type d -a -name .svn ) -prune -o'    \
           --findSourceOps='( -type d -a -name .cache ) -prune -o'  \
           --findGeneralOps='+'                                     \
           --findGeneralOps='( -type f -a -name core ) -o'          \
           --findGeneralOps='-path ///d/mount -prune -o'            \
           --findGeneralOps='-path ///d/database -prune -o'

Compare files Byte-By-Byte instead of by just their sizes and modification times (warning: this might take much time):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --byteByByte

Do not prepare scripts for the case of restore (No Restore, saves processing time and disk space):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --noRestore

Instead of GNU AWK, use MAWK, the very fast AWK implementation based on a bytecode interpreter:

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --mawk

Produce less screen output (No Progress Messages from the analysis phase):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --noProgress

Do Not Execute the actions (just prepare the scripts, see the Documentation for more details):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" --noExec

Put Zaloha2 Metadata (working files of Zaloha2.sh) to a separate storage, e.g. for speed reasons (please note that if Zaloha2.sh is used for multiple synchronizations, then each such use (instance) of Zaloha2.sh requires a separate Metadata directory):

Zaloha2.sh --sourceDir="test_source" --backupDir="test_backup" \
           --metaDir="/fast_storage/Zaloha_metadata/data_of_instance_0001"

Show the Documentation (quit viewing by typing q):

Zaloha2.sh --help

So much typing?

Some command lines may get lengthy. But: Usually you do not create new setups each day, nor you type the commands always from scratch. After having fine-tuned a given synchronization/backup process, you are advised to materialize the command line of Zaloha2 into a wrapper script. The wrapper script can then be attached to a desktop icon, for instance, so that the whole process can be started just by double-clicking on that icon. The advantage of wrapper scripts is that you can do more in them than just invoke Zaloha2: e.g. mount the backup media before, unmount it after, specially handle sensitive data and so on.

Usage Screenshot (of the Local Mode)

Performance tuning and Performance data (in the Local Mode)

• The option --noRestore switches off creation of the restore scripts. If you do not need the restore scripts, use this option to shorten the analysis phase.
• The option --mawk instructs Zaloha2 to use MAWK instead of (usually) GNU AWK. MAWK is an AWK implementation based on a bytecode interpreter and is significantly faster. To utilize MAWK, it must be installed on the local system.

Performance was measured on following system and data:

• Standard commodity PC with Intel i3, 3.3GHz, 4GB RAM
• Source directory: local Seagate HDD 500GB, 7200rpm, SATA III
• Backup directory: a second local Seagate HDD 500GB, 7200rpm, SATA III
• Filesystem: ext4 on both HDDs
• Data synchronized: 110 GB, 88.000 files
• Operating system: Linux (Fedora 30)
• Binary utilities used by Zaloha2: GNU find, GNU sort, mawk
• Zaloha2 options: --noRestore YES, --mawk YES, --detectHLinksS NO, --byteByByte NO

Measured performance of the analysis phase:

• first run: 25 seconds (filesystem data not cached in the OS: the FINDs 2 x 12 secs, the sorts and AWKs 1 sec)
• second run: 5 seconds (filesystem data cached in the OS: the FINDs 2 x 2 secs, the sorts and AWKs 1 sec)

Performance of the execution phase:

• depends on how many files need synchronization: zero to several minutes

The Remote Source Mode

In the Remote Source Mode, the source directory is on a remote source host that can be reached via SSH/SCP and the backup directory is available locally. This mode is activated by the option --sourceUserHost.

The Remote Source Mode explained in eight sentences

• The script for FIND scan of the source directory (file 210) is prepared locally.
• The script is then copied to the remote source host via SCP and executed via SSH.
• The obtained CSV metadata (file 310) is copied back from the remote source host via SCP.
• The FIND scan of the backup directory occurs locally.
• The CSV metadata is compared by a sequence of sorts and AWK processing steps (occurs locally).
• The results (= prepared synchronization actions) are presented to the user for confirmation.
• If the user confirms, the synchronization actions are executed (via SSH and SCP).
• A non-interactive regime is available as well.

Usage Example of the Remote Source Mode

# Establish the SSH master connection
ssh -nNf -o ControlMaster=yes -o ControlPath='~/.ssh/cm-%r@%h:%p' 'user@sourcehost'

# Run Zaloha2.sh
Zaloha2.sh --sourceDir="test_source_remote"    \
           --backupDir="test_backup_local"     \
           --sourceUserHost='user@sourcehost'  \
           --sshOptions='-o ControlMaster=no -o ControlPath=~/.ssh/cm-%r@%h:%p'     \
           --scpOptions='-o ControlMaster=no -o ControlPath=~/.ssh/cm-%r@%h:%p -T'

# Terminate the SSH master connection
ssh -O exit -o ControlPath='~/.ssh/cm-%r@%h:%p' 'user@sourcehost'

The Remote Backup Mode

In the Remote Backup Mode, the source directory is available locally and the backup directory is on a remote backup host that can be reached via SSH/SCP. This mode is activated by the option --backupUserHost.

The Remote Backup Mode explained in eight sentences

• The FIND scan of the source directory occurs locally.
• The script for FIND scan of the backup directory (file 220) is prepared locally.
• The script is then copied to the remote backup host via SCP and executed via SSH.
• The obtained CSV metadata (file 320) is copied back from the remote backup host via SCP.
• The CSV metadata is compared by a sequence of sorts and AWK processing steps (occurs locally).
• The results (= prepared synchronization actions) are presented to the user for confirmation.
• If the user confirms, the synchronization actions are executed (via SSH and SCP).
• A non-interactive regime is available as well.

Usage Example of the Remote Backup Mode

# Establish the SSH master connection
ssh -nNf -o ControlMaster=yes -o ControlPath='~/.ssh/cm-%r@%h:%p' 'user@backuphost'

# Run Zaloha2.sh
Zaloha2.sh --sourceDir="test_source_local"     \
           --backupDir="test_backup_remote"    \
           --backupUserHost='user@backuphost'  \
           --sshOptions='-o ControlMaster=no -o ControlPath=~/.ssh/cm-%r@%h:%p'     \
           --scpOptions='-o ControlMaster=no -o ControlPath=~/.ssh/cm-%r@%h:%p -T'

# Terminate the SSH master connection
ssh -O exit -o ControlPath='~/.ssh/cm-%r@%h:%p' 'user@backuphost'

Comparing contents of files via SHA-256 hashes

The option --byteByByte cannot be used in the Remote Modes, because the CMP command needs local access to both compared files.

For the Remote Modes, a new option --sha256 has been introduced. This option causes both FIND scans to additionally invoke SHA256SUM on each encountered file. Note that the SHA-256 hashes of the files are calculated on the hosts where the files are located. Also no contents of files are transferred over the network, just the SHA-256 hashes. The SHA-256 hashes are then compared to detect files that appear identical but their contents differ.

Other uses of the SHA-256 hashes

The option --sha256 is not limited to the Remote Modes – it can be used in the Local Mode too. CSV data about files that contain their SHA-256 hashes can be used for other purposes as well:

Assume you have files with identical (duplicate) contents scattered across your filesystem, and you want to keep only one copy per unique file. In other words: de-duplicate files by content. The finding of files with duplicate contents can be achieved by sorting the CSV file 330 by the SHA-256 hashes (= by column 13) and evaluating the result, e.g. by a simple AWK program that prints out the files where the SHA-256 hash equals to the SHA-256 hash in the immediately preceding record in the sorted file.

Performance tuning in the Remote Source and Remote Backup Modes

• The options --noRestore and --mawk (see above) are relevant for the Remote Modes as well.
• The option --findParallel instructs Zaloha2 to run the FIND scans of the source and backup directories in parallel. As these scans run on different hosts, this will further save time.
• Last, SCP can be tuned for higher speed by choosing suitable ciphers and compression levels. See SCP documentation for details. Use the option --scpOptions to pass additional command-line options to the SCP commands invoked by Zaloha2. Faster SCP will shorten both the analysis phase (faster copying of CSV metadata over the network) as well as the actual execution phase.
• Eventually use the High performance SSH/SCP – HPN-SSH that removes the performance limitation due to statically defined internal flow control buffers.

Development status

• From the perspective of my own requirements and ideas, Zaloha2 is a completed work.
• Zaloha2 successfully passes all my test cases.
• Eventual (conclusive) Problem Reports are welcome (via Issues).
• Enhancement Requests so far they won't contradict the "small and simple synchronizer" idea (= no program code inflation, no new processing steps, no increase in runtime etc).

Add-on script Zaloha2_Snapshot.sh

An add-on script to create hardlink-based snapshots of the backup directory exists: Zaloha2_Snapshot. This allows to create Time Machine-like backup solutions.

Full list of changes (Zaloha2.sh compared to Zaloha.sh)

Some design changes break backward compatibility with Zaloha.sh (the predecessor of Zaloha2.sh). For this reason, this repository is a new repository created specially for Zaloha2.sh.

License

MIT License