Also see: The 9 Best Linux Distros
Over the past year, I've spent more time than ever using rolling release Linux distributions. My experiences have been positive and negative, depending on the distribution and system updates applied.
Having tried a number of different rolling release distros, I'll be speaking frankly in this article about a solid case against rolling release distributions. But before you jump to any conclusions, it's worth reading the entire piece to better understand where I'm going with this.
Newer isn't always better
I know for a fact that newer feels better, especially with regard to software. Updates usually come with new features, bug fixes, and even performance improvements! There are some reasonable benefits to updating one's software on a regular basis, and lucky for us, Linux packaging repositories make this practically automatic. But what happens when the updates come in too quickly, and perhaps are untested?
For example, while most distribution specific repos offer well tested software updates, sometimes the user repos can end up pushing software that might need a little more time in the debugging department. Whether it's a rolling release or a fixed release, the software offered by your distribution is generally considered to be safe and stable.
Unfortunately there are occasions with desktop environments or kernel related updates where bugs do crop up. On two of my rolling release PCs, a kernel update completely froze out my network cards – wired networking even, nothing uncommon at all. This network card issue affected two completely separate systems because of a kernel bug.
Clearly, I was free to roll back to a working kernel, which was easy enough for me as an experienced user. For a newer user, well, it would have left many feeling pretty disenfranchised.
With a fixed release distribution, this isn't even an issue. You could burn a CD of Ubuntu's latest 14.10 release running kernel 3.16, while still using Ubuntu 14.04 with kernel 3.13 and try out the newest release. When you discover that on the live image, networking is a bust – it's no big deal. Thankfully, you were merely testing things out on a CD and can simply remove it, file a bug report and go on with your day.
Do you see the difference? For advanced users, side-stepping a kernel bug is simple enough, however for newbies, not so much. Someone coming from a fixed release kernel would never consider that a new kernel is likely to be the issue. Fixed releases are definitely winners in this space if you'd rather not trouble shoot stuff with each update bug that crops up.
Fresh installation vs old installation
I'll be the first to admit I tend to be a big fan of the dedicated home directory. By the same token, there is something to be said for a clean installation of an operating system vs upgrading an existing one. All the cruft and other unused data can be easily removed by treating yourself to a clean installation of your Linux installation.
With fixed release distributions, this is easily done. But if you're relying on a rolling release, you will never get this benefit unless you decided to wipe the drive and re-install yourself. Some might argue that there is no single advantage to one over the other. I disagree as a fixed release schedule entices the user to entertain the idea of a clean install vs using the upgrade tool.
Considering the upgrade tool for your fixed release distro is bound to fail eventually, you're eventually going to end up doing a clean installation. Rolling distributions will have to wait until things become so messed up that rolling back packages or config file editing isn't going to fix the problem at hand. Point being, nothing runs better than a fresh Linux installation – not even a 5 year old rolling release install.
Fixed releases are more stable
I accept that I will end up with grief for making this statement, but without question, if you want stability, you want a fixed release distribution of Linux. Rolling releases can be more convenient, but they also open up doors to challenges that might not otherwise be introduced. Taking the time to allow others to test out updates made to desktop environments and kernel updates are something I believe many Linux users overlook.
Granted, distributions such as Arch do a very good job at making sure any "surprises" are mentioned ahead of time in their awesome mailing list. But let's be honest, unless you're a super-geek, you're not going to read a mailing list very frequently. Unless you're working in IT or you have the discipline to think about it, odds are you're just as forgetful as I am when it comes to this sort of thing. Stability through self-research before updating is a skill in itself and one that most of us lack.
With a fixed release of Linux, on the other hand, you're free to be pretty forgetful. Forgot to update for two weeks? No big deal, just install the updates and get on with your day. Even though there are occasions where a bug "could" be introduced, it's rarely going to break something critical at the desktop or kernel level.
To be clear: exceptions with both fixed and rolling releases do exist, but for the most part I've found that a fixed release distribution is more stable simply because it limits the opportunity for the user to break it.
Supporting fixed release distributions
I believe strongly that, for a business environment, a rolling release is just begging for extra work. Fixed release distributions in these environments allow IT staff to button things down and manage any needed updates in easy to digest segments. As a bonus, the IT team isn't left troubleshooting kernel bugs until they've set aside the time to drill deep into the upcoming release of their preferred fixed release distro.
Remember from above, less moving parts means less potential for breakage. If you control the update cycle with regard to kernel and desktop environment upgrades, you've won half the support calls right off the bat when dealing with common workstation errors.
At this point, one might surmise that I really dislike using rolling release distributions. Or worse, I have little to no experience with them. The answer to this, of course, was addressed at the beginning of this article. I've been using rolling release distros for roughly a year now. I can see the appeal of using them, but I also feel that there needs to be a clearer set of guidelines for when they make sense as an option.
For desktop users, I believe newbies should avoid using rolling release distros until they're more comfortable rolling back packages and/or overcoming sudden breakage. Whether it's something as silly as a broken GNOME extension that worked yesterday, or your Ethernet card suddenly stopped working, if you can't overcome these issues – avoid rolling releases.