Re: The Tragedy Of systemd

On 09/10/2024 11:36, Richard Kettlewell wrote:

Marc Haber <mh+usenetspam1118@zugschl.us> writes:

Lars Poulsen <lars@cleo.beagle-ears.com> wrote:

But redhat has funding from and obligations to groups that run large
datacenters that need to be managed from control desks based on
parameterized templates, and to support these needs, they built
something that works for that crowd.

>
And still those server people complain that systemd feels more like
its geared for desktop machines. Noone cares how long it takes to boot
a server,

 ??? the end users of whatever service the server is providing are likely
to care. Certainly if one of our dev servers (e.g. git, wiki, bug
tracker) has to be rebooted in the middle of the the day[1] then work
will gradually grind to a halt until it’s back.
 

Conversely with SSDS and stripped down to daemons only, a server doesn't have that much TO bring up.

   [1] avoided if possible of course, but sometimes shit happens.
 

and many server jockeys would love the possibility to turn
off the parallelism of systemd when booting (for reproducibility,
sacrificing speed).

 

What I dislike mainly is that its code for an ego trip. Log files that need a program to list them? Why?
New code that breaks existing startup scripts in new and exciting ways and a fairly low level of community knowledge makes for a bad user experience.
--
If I had all the money I've spent on drink...
..I'd spend it on drink.
Sir Henry (at Rawlinson's End)