[CentOS] What are the differences between systemd and non-systemd Linux distros?

Fri Oct 19 11:07:46 UTC 2018
Simon Matter <simon.matter at invoca.ch>

> Valeri Galtsev wrote:
>> On 10/17/18 7:55 PM, Warren Young wrote:
> <snip>
>>> Benno Rice is right: Lennart Poettering gets stuff done.
> Because he's funded. And I strongly suspect that a lot of that funding
> comes from M$'s interest in Upstream.
> <snip>
>> With all due respect, many people just stopped offering any argument
>> about systemd, and simply fled elsewhere which in _their_ opinion
>> (and I am one of them) lies better in what they with their education
>> and life experience is more reasonably resembling system suitable
>> for servers.
>> Servers are key word for me. You can see me using macintosh laptop in
>> variety of places, that doens't mean MacOS will be my choice for server,
>> so
>> don't count laptopls into any statistics. The same is true about a bunch
>> of other sysadmins I know, who mostly use Apple laptops, whereas run
>> Linux, or UNIX-like, or [truly] UNIX servers.
> Actually, I've got CentOS on my 9 yr old Netbook, that I use while
> traveling. Otherwise, my home workstation is CentOS 6, and I am NOT
> looking forward to EOL.
> But Valeri's correct: people are tired of screaming and yelling about
> systemd, because we've had years now of the response being "tough, it's
> the Wave of the Future", and Poettering is like upper management: they
> know, I mean, Everything, so why should they need to talk to end users (or
> working sysadmins)?
> Lack of screaming and yelling filling this venue is more because "what's
> the point?", and we have to get work done.


A lot was already said but let me underline a few things from my personal
point of view:

- the upgrade path from EL6 to EL7 is completely broken.

That's certainly a good thing for upstream, because they can sell even
more support and training. I don't blame them for trying to make money, I
just say from the technical point of view it's not the best solution.

For home users it doesn't hurt too much but for the enterprise market it's
bad. Migrating complex systems is a huge amount of work and takes a lot of
time and manpower. In the end it means higher costs.

- the migration to systemd is not really finished carefully in EL7.

Just look into upstream's Bugzilla and see how many issues still exist and
will probably not be fixed.

I show you a simple example: we happen not mount some NFS filesystems on
servers like this in /etc/fstab:

ftp:/var/ftp/pub     /mnt/nfs          nfs   bg,soft   0 0

Now, with every Linux since the last millennium one could simply bring
down the system into maintenance mode with 'telinit 1', and all worked
Now try the same with EL7, do a 'systemctl rescue' or 'systemctl
emergency' and see what happens. With lightning speed it does the wrong
thing, brings down networked services, brings down the network, and
doesn't unmount the NFS filesystems. Then try a 'df' or 'lsof' in rescue
mode, it all hangs.

Of course I found a solution, mount it with the option
'x-systemd.requires=network-online.target' and it behaves correctly. But
really, it's broken, because it's always clear that NFS mounts always only
work WITH network!

That's just a single small example how things don't work as expected.

- migrating from EL6 --> FreeBSD seems easier than migrating from EL6 -->

That's really an important point, because those who started using Linux
with Linux/systemd will be bound to Linux/systemd with their knowledge,
switching to a *BSD or other Unix will be difficult. For me, I don't like
to be limited in such ways.

In other words, systemd is a new operating system which still lacks a
kernel :-)

One thing I know for sure: if the *BSD folks were ever going to invent
something like systemd, they will do it in a way which hurts less.