Sync Job Manager

Description

This script automates the task of syncing folders and projects (databases) on client computers with a central server. It builds on other software for data synchronization between distributed systems. At the moment support for Git & Unison is implemented, but other protocols like rsync are planned for the future. In case of Git all branches in for those projects, which are stated in the *.conf files, will be synced with the server.

Usage

for pulling changes from the server

./server-sync.sh pull $conf-file

for pushing to the server

./server-sync.sh push $conf-file

Options

cmd	descr
--force, -f	syncing will be overwrite destination (in case of push) or overwrite local files (in case of pull)
-c <client>	select client, at the moment only git and unison are available; if none is given, all syncing clients are used
--env <name>	select syncing environment, e.g. 'work', the *.conf files can be configured to only apply to a certain environment; all other environments are ignored

for setting the config (name + email address only possible for git repos)

./server-sync.sh set-conf

Configuration

• Git repositories have to be set up and wired by yourself for each client and the server
• *.conf-files: the path to the local Git project and the remote url of the server repo have to be provided in conf-files
• *.conf files should be in the home directory under ~/.sync-conf or can be provided as an argument to the call
• best would be to set up an easy authentication method with the server, like a key-based method, in order to avoid prompts when running the script
• for setting the git config for a repo, you need to provide the name + email address next to the repository path

Requirements

Can only be run from a Bash, e.g. MacOs or Linux distribution with git and unison installed (and a transfer protocol as ssh on client and server)

Currently planned

• lock files
• specify if only push, only pull or both modes are applied to each repo
• gui for convenient setup of conf files
• error report in case of conflicts or failed checkouts At the moment the script is simple and not laid out to sync in collaborative projects. E.g. in case of git, when pushing the server branches are expected to be fast-fowarded. The same applies to pulling. Merges cannot be handled. In the future this should be possible.