Slackware: Is systemd inevitable?
|
To paraphrase Patrick in the interview: While nobody objects to a faster boot process, increased complexity is rarely a good thing and may not represent a good trade-off in this particular case. And since no software actually requires systemd, there's no pressing need to replace SysV init in Slackware or anywhere else. (Well, Mr. Poettering did try to press the issue by having systemd absorb udev, but that just resulted in a udev fork.)
Personally, I'm not convinced the reduced boot time offered by systemd is such a great thing. After all, laptop users avoid reboots altogether with suspend and hibernation, while server operators reboot once in a blue moon and typically have to wait longer for the BIOS POST than the OS boot process anyway. |
I think this is a well-reasoned write up regarding systemd:
http://monolight.cc/2011/05/the-systemd-fallacy/ Yeah, I'm not buying it. Complexity, non-modularity, etc. I hardly ever reboot, so touting faster boot times is NOT a selling point, aside from the other strikes against it. |
Oh boy, here we go again!
|
See Previous Thread
Please refer to this LQ thread on this subject.
http://www.linuxquestions.org/questi...ystemd-885228/ If you would like to comment on an aspect of systemd with regards to Slackware that was not already included or referenced in that 513 post thread then please speak up. Please read all posts in that thread (the systemd related ones) and read all referenced links before posting. |
Everyone can say what they want, but I think that at the end, the situation can be summarized with this:
Quote:
I think a lot of people here should read it if it's not already done. It would help to stop spreading misinformation like I can read all day about systemd on this forum |
Quote:
|
Quote:
I didn't have an opinion on this before, but certainly I do now. That page has really put me off systemd. I don't care how cool or efficient it may or may not be. I'm not even going to look at it after reading that. |
Quote:
|
Let it mature and stabilize, and then take a look at it. After all, it doesn't look like an alien idea: it is basically an inetd for system boot. If it works and its issues are resolved, then why shouldn't we use it?
But traditional linux boot is very very simple to understand conceptually, even if the particular scripts that contain it are not. I am not sure systemd is equally simple to fix once things get dirty. |
I remember when we were having the same discussion about *udev* as a replacement for hotplug.
|
For the average desktop user it probably wouldn't make any difference to use systemd. I used it on Arch for a while and it seemed OK, though the quick boot made no real difference to me. Equally, I'm not sure systemd does anything that can't already be done. To my mind it comes down to managing change and looking at the way in which change is effected and what it represents. For me systemd represents a kind of centralisation of power and influence to RHEL/Poettering and perhaps an end to a certain kind of diversity within the Linux ecosystem. For that reason, and that reason alone, I hope that Pat never puts it in Slackware. The complexity of systemd is very far from the simplicity that Slackware provides.
|
As always I trust Pat to make sane decisions with Slackware; I trust his judgement.
|
Quote:
This is why Slackware is the only Linux distribution I can use without losing my sanity. The 'keep it simple' approach is very much in line with my own personal philosophy. The Slackware team do an excellent job with their cautious approach. If they're eventually railroaded into using systemd, I trust them to implement it in a logical and sane manner. FWIW: I think that the present init system in Slackware is superb. It is easily understood by novices like me. It is also extensible and flexible. There is no real need for anything else, and this, I think, is where a lot of the angst toward systemd stems from. |
I wish DevFS, Hotplug, and HAL were still being developed. BSD systems still use them without so much as a hiccup compared to the issues with udev and systemd powered Linux systems. Udev is still broken last I checked with many of it's rules and even I still have issues with it creating redundant rules that break stuff. Pain in the arse...
I respect whatever solution Patrick goes with, and you can see where other distributions including Linux from Scratch have worked tirelessly to avoid systemd at all costs, but I sincerely hope Patrick does what he can to keep systemd out as long as possible. Bad software does not a stable system make. |
All times are GMT -5. The time now is 09:07 PM. |