What separates the Men from the Boys. Fedora vs. Antergos
This morning began with a nice cup of coffee and I commenced going on line. I’ve been energized of late with the discovery of Antergos.
But this am I got a punch in the stomach. Antergos saw an update needed to be applied. I let is do its update noting it included an update to Arch kernel 3.19.2-1, so when the update finished, I rebooted into the new kernel.
I was taken directly into the Desktop with no greeter login/password intervention. That seemed odd. Then I noticed I had no network connection despite nm-applet running. There wasn’t an option to edit my connections either – it was blank.
I’ve run into this with rolling releases before, and particularly with kernel updates.
I plugged a cat5 cable in and got a connection, then search on Antergos forums to see if there was any help. None. It was early and probably I was among the few ‘lucky souls’ who tried to update.
Joe Average is simply ‘dead in the water’ when this happens. They don’t have the requisite knowledge to recover. My Netbook became a paper weight. I was miffed.
So, I got to thinking about the real virtue of ‘rolling releases’. My conclusion is, when it works it is fine. When things go a kilter your system can become borked, unusable.
Looking back to how Microsoft does it, they have restore points.
You have to tip your hat to them for having years ago seen the importance of such a feature.
With Antergos you get a sausage. Okay, I see a post:
with suggestions for a ‘work-around’ to roll-back the kernel.
This is not elegant to say the least. And who knows if it would even work depending on the ‘state’ of each machine (no two being the same).
So, I am reassessing sipping my coffee and decide, I am going back to Fedora. This would and cannot happen on that platform.
Better yet, I went ahead and configured the partitions with Btrfs. And, installed yum-plugin-fs-snapshot which automatically snapshots your filesystems during updates.
The benefit of this plugin comes exactly in this situation – a rollback to the last ‘snapshot’ or restore point can be made in 5 seconds.
Disaster avoided. Imagine this happening in an Enterprise environment.
You’d be put up against the nearest wall and shot. Enterprise cannot have downtime period the frickin’ Amen. No excuse is acceptable.
This is what separates the Men from the Boys.
I do love what the developers of Antergos have done thus far, but is wouldn’t survive in the real business world. It’s okay for hobbyists.
So with yum, you get package management with history roll-back capability and all applications are precompiled by the vendor, not you.
With Arch and yaourt, the work of doing source code compiling and build are deferred until you choose to install and then depending on the application you wait minutes to as much as hours for the build and then final compress into an Arch package is complete ‘before’ you can install. It’s a most novel approach, but it is a trade-off.
If I need an app that isn’t rpm packaged, I can rpmbuild it myself or just build the tar.gz. That’s not an option for ‘Joe Six-Pack’ and they ‘should’ have a ‘recovery’ option in Antergos. They don’t. And that’s a major factor to consider when you pick which Distro to go with.
Seriously, I think Antergos is filling the biggest unmet need with Arch – making ‘Arch made Easy’ for Everyone. I think the installer is slick, seamless and professional. I hope the developers understand and are reading my comments. You are doing great work. Keep working and make Antergos the best Distro. I think you can do it.
Good Luck. – Dietrich