These discussions usually end up with the same list of pointless arguments.
systemd has this one annoying bug that really sucks
Nobody says systemd is perfect, but all things considered, many people agree that it's better than the alternatives. I've found systemd modules to be easier to maintain than bash scripts, anyway. Just as a reminder, on my gentoo system, the average init script is 86 lines long, the longest 789. systemd modules are usually around 10. Sure, both OpenRC and upstart have a few nice features that systemd lacks, but overall, I consider systemd to be superior to either.
Why would I want systemd? sysvinit works fine for me.
That's great. If you had your way, we'd still be using linux 2.4 and motif. The features that systemd brings are important and waaaay overdue. Sure, you could implement cgroups, container spawning and read-only roots in your sysvinit bash scripts, but eventually your bash scripts will be 5k LoC each, making them brittle and unmaintainable. Which is why nobody does it, and which is why we need a proper init system that turns common requirements into a fool-proof one-liner.
But it's about choice
I can probably name a dozen packages that everyone on every distribution uses, for which there are no alternatives. Yet, there are no debates about any of them. I really don't see why it's bad that we converge on a new and modern init system, but not a problem at all that we're stuck with the same kernel, the same ssh implementation, the same util-linux, the same zlib, and for most of us the same shell implementation and the same bootloader.
The whole debate is not technical, but emotional.
But all those things aside, isn't devuan a good thing?
We'll see. It takes more than a proof-of-concept and an announcement to make a distribution I can trust. Their announcement raises a few red flags, so does their aggressive donation drive. But the most important test is their handling of security issues. OpenSSL will have a critical update in two days. If I was considering switching to devuan, I'd keep a close eye on the availability of updates and their communications around them.
For comparison: Gentoo Linux Security Advisories, Debian Security Advisories, Devuan Security Advisories. Notice anything odd on the third link?
systemd has this one annoying bug that really sucks
Nobody says systemd is perfect, but all things considered, many people agree that it's better than the alternatives. I've found systemd modules to be easier to maintain than bash scripts, anyway. Just as a reminder, on my gentoo system, the average init script is 86 lines long, the longest 789. systemd modules are usually around 10. Sure, both OpenRC and upstart have a few nice features that systemd lacks, but overall, I consider systemd to be superior to either.
Why would I want systemd? sysvinit works fine for me.
That's great. If you had your way, we'd still be using linux 2.4 and motif. The features that systemd brings are important and waaaay overdue. Sure, you could implement cgroups, container spawning and read-only roots in your sysvinit bash scripts, but eventually your bash scripts will be 5k LoC each, making them brittle and unmaintainable. Which is why nobody does it, and which is why we need a proper init system that turns common requirements into a fool-proof one-liner.
But it's about choice
I can probably name a dozen packages that everyone on every distribution uses, for which there are no alternatives. Yet, there are no debates about any of them. I really don't see why it's bad that we converge on a new and modern init system, but not a problem at all that we're stuck with the same kernel, the same ssh implementation, the same util-linux, the same zlib, and for most of us the same shell implementation and the same bootloader.
The whole debate is not technical, but emotional.
But all those things aside, isn't devuan a good thing?
We'll see. It takes more than a proof-of-concept and an announcement to make a distribution I can trust. Their announcement raises a few red flags, so does their aggressive donation drive. But the most important test is their handling of security issues. OpenSSL will have a critical update in two days. If I was considering switching to devuan, I'd keep a close eye on the availability of updates and their communications around them.
For comparison: Gentoo Linux Security Advisories, Debian Security Advisories, Devuan Security Advisories. Notice anything odd on the third link?
Comment