Tag Archives: 10.8

Counterpart – safely live cloning OS X Server


Pushed yesterday to GitHub is version 1.2 of Counterpart. It adds a helpful feature that makes it incredibly simple to safely live clone OS X server; optional server & database backup.

During a live disk clone, the Open Directory and PostgreSQL databases utilised by OS X server may not be in a consistent state on the disk. Because of this, the cloned database data may be corrupted, and important Server and services data may not be backed up correctly. Counterpart solves this by providing an option to archive Open Directory & dump PostgreSQL to disk before attempting a clone. Whilst this was possible before with the help of a pre-clone script, adding these features directly in Counterpart has made it easier than ever to implement a cloned backup solution on OS X Server.

More information is available in the Counterpart README, but backing up OS X server data is as simple as throwing the new -b flag as follows:

counterpart -s / -d /Volumes/BootClone -b password

A lot more information on exactly how Counterpart achieves its backups and cloning process is available in the README on GitHub.

Counterpart on GitHub


Introducing Counterpart: a live bootable clone tool for OS X

These days, backing up Mac OS X Server is easier than ever, with plenty of options available for taking incremental backups of system and service data, and great support from Time Machine’s ServerBackup process in making sure that databases and Open Directory are included in hourly backups.

All of these great options have one thing in common though: restoration. In a production environment, as lovely as backups are; admins are often hounded by management and users when hardware fails, and time elapses whilst operating systems are re-installed and service data rolled back from backup drives. To combat this downtime, I like to keep a fully bootable backup connected to each production server that is as up to date as possible. In many situations, this allows you to get a slightly data-delayed version of your environment up and running until you can schedule proper downtime to restore backed up data. For years, I utilised Mike Bombich‘s fantastic tool Carbon Copy Cloner, along with it’s excellent Scheduled Tasks functionality for this exact purpose. With it’s recent move to shareware, I decided that there really should be a free, open source script for cloning Mac OS X that can be scheduled with launchd, and provides proper logging and statistical data. It is with this that I announce the release of Counterpart; a wrapper script for rsync on OS X that is capable of producing bootable clones of live Mac systems. I have been using this for a little over 6 months to backup hundreds of OS X systems, and it has made my life so much easier a couple of times during hardware failures.

Counterpart utilises rsync, the fantastic data and synchronisation utility, and wraps it in a script that provides the correct settings and filesystem exceptions to create a bootable clone of a live Mac system, whilst error checking and providing comprehensive logging, statistics and monitoring data. It is bundled with detailed documentation and instructions on scheduling clones. Coming soon is a companion Nagios plugin script to monitor clones and provide performance data, meaning you can be sure clones are completing successfully, and get insights into your backup data like this:

Counterpart Clone Statistics

It should also be said that whilst Counterpart was first envisaged to backup Mac OS X Server instances, there is nothing stopping it being used to back up standard OS X clients, and it would be a great, fully scriptable, free way to create a bootable clone of your OS before an upgrade or significant modification to your system.

I have uploaded Counterpart as a new project on GitHub, and welcome any feedback that you may have on using it. I am also happy to answer any questions on getting it set up in your environment, and I am best contacted using this form.

Counterpart on GitHub


Monitoring the new OS X Server Caching Service with Nagios

In followup to last week’s post on Monitoring Mac OS X Server Software Update Server, here is my new script to monitor the Caching Service on OS X Mountain Lion Server (Server.app 2.2+).

The Caching Service is a deceptively magical new service which automatically caches Apple update and Mac App Store content with no need for client configuration. A very good writeup of what the service actually does can be found at Noel Alonso’s blog.

In classic Apple style, the new Caching Service looks pretty bare on the surface, with a simple toggle and a slider to change the size of your cache. Whilst advanced configuration of the service is definitely possible; from a cursory glance, none of the options or statuses give you a really good idea of what is happening in the background.

This is where the following script works nicely. Checking that the service is active, registered, and accessible to clients provides a good way to monitor that caching of content should occur, but the performance data that the script returns gives you some excellent ways to monitor and analyse cache size and efficiency. As you can see in the graph above, I am able to get a good visual indication over time of how much content my server is caching, as well as how much cached content it is providing to clients versus downloading directly from Apple. In the next few weeks, I will write some articles on the RRDtool commands I am using to produce these graphs.

It is really satisfying to be able to see nice chunks of bandwidth that no longer have to come in via the internet. Enjoy!

check_osx_caching.sh on GitHub