[sf-lug] systemd criticism

[jim]

On my laptop running Ubuntu 12.04.X

$ man systemd
No manual entry for systemd

$ ps aux | grep system
102        987  0.0  0.0  25652  2576 ?        Ss   Aug01   5:02 
dbus-daemon --system --fork --activation=upstart
root      2402  0.0  0.1  95548 12348 ?        S    Aug01   0:00 
/usr/bin/python /usr/lib/system-service/system-service-d
jim      12614  0.0  0.0  13592   932 pts/4    S+   11:58   0:00 grep 
--color=auto system

~$ ps aux | grep init
root         1  0.0  0.0  24568  2524 ?        Ss   Aug01   0:02 /sbin/init


 From Wikipedia
*systemd* is a system management daemon 
<http://en.wikipedia.org/wiki/Daemon_%28computing%29> designed for Linux 
<http://en.wikipedia.org/wiki/Linux> and programmed exclusively for the 
Linux API 
<http://en.wikipedia.org/wiki/Linux_kernel_interfaces#Kernel.E2.80.93user_space_API>. 
For systems using it, it is the first process 
<http://en.wikipedia.org/wiki/Process_%28computing%29> which is executed 
<http://en.wikipedia.org/wiki/Execution_%28computing%29> in user space 
<http://en.wikipedia.org/wiki/User_space> during the Linux startup 
process <http://en.wikipedia.org/wiki/Linux_startup_process>. Therefore, 
systemd serves as the root of the user space's process tree 
<http://en.wikipedia.org/wiki/Process_tree>.


JS: I think of the kernel as process 0 and init as process 1.
I read the Wikipedia excerpt as systemd replacing init, i.e.
systemd has process ID 1 and init need not apply.

JS: I find the LKML link uninformative for my experience
(i.e. I feel the need for answers to questions about the
underlying scheme of things--kernel kicks off init and
things go from there: how does systemd fit in, as a
replacement for init or as PID 2 or what, and then what's
the role of init?)





On 08/26/2014 02:35 PM, Jeff Bragg wrote:
> Mostly more discussion:
>
> http://lwn.net/Articles/602579/
>
>
> On Tue, Aug 26, 2014 at 2:33 PM, Jeff Bragg <jackofnotrades at gmail.com 
> <mailto:jackofnotrades at gmail.com>> wrote:
>
>     Here are a couple of links from my history that explain some of
>     the relevant issues.  The first link is an article, the second
>     discussion about it, as far as I can tell.
>
>     http://ewontfix.com/14/
>     https://news.ycombinator.com/item?id=7210064
>
>
>     On Tue, Aug 26, 2014 at 2:25 PM, Jeff Bragg
>     <jackofnotrades at gmail.com <mailto:jackofnotrades at gmail.com>> wrote:
>
>         Brian:  Yes, it does.  Sorry, I probably should have provided
>         some warning.
>
>         Akkana:  It is light on concrete criticisms. I've come across
>         discussions with more detail, I just didn't have the links to
>         hand when I sent this.  I'll try to track them back down and
>         add them here.
>
>
>         On Tue, Aug 26, 2014 at 2:03 PM, Akkana Peck
>         <akkana at shallowsky.com <mailto:akkana at shallowsky.com>> wrote:
>
>             Jeff Bragg:
>             >> I'm sympathetic to the points made here.
>             >>
>             >> https://lkml.org/lkml/2014/8/12/459
>
>             I think I'm sympathetic, but ... what are the actual points?
>             I see a lot of flameage, and a lot of namecalling, and no
>             concrete
>             reasons he hates systemd so much, no details on how it
>             messed up
>             his system.
>
>             It would be a lot more helpful to write about regressions
>             caused by
>             systemd. Like:
>
>             Brian Wood writes:
>             > I accidentally got a lot of junk in my log file... the
>             same error
>             > over and over.  i tried to find a way to clean out the
>             junk, but
>
>             A couple of lines like that contain more detail than
>             Christopher
>             Barry's whole rant. Or:
>
>             > wasn't able to in part because the log was in binary.
>
>             Binary logs? Really? Now that's a good concrete argument
>             against
>             systemd.
>
>             I haven't delved into systemd much, and I'm curious to
>             hear about
>             the good and bad about it. I did fight with upstart quite
>             a bit, and
>             found it a big step backward from SysV init files --
>             mostly because
>             everything was done with undocumented compiled binaries
>             rather than
>             self-documenting shell scripts, so if you wanted to change
>             anything
>             about the boot process, you had to download (sometimes
>             quite a lot
>             of) C source, read it, and maybe rebuild and install it. I
>             gather
>             that's true of systemd too, and that it affects a lot more
>             of the
>             system than upstart (more than just boot).
>
>             But I haven't hit any actual problems with systemd yet on
>             Jessie
>             or Sid.  So I'm curious to hear what the problems are, and
>             their
>             workarounds, since I'm sure I'll hit them eventually. 
>             "systemd is a
>             trojan. systemd is a medusa ... groups with agendas ...
>             just don't
>             believe in freedom" just isn't very enlightening.
>
>                     ...Akkana
>
>             _______________________________________________
>             sf-lug mailing list
>             sf-lug at linuxmafia.com <mailto:sf-lug at linuxmafia.com>
>             http://linuxmafia.com/mailman/listinfo/sf-lug
>             Information about SF-LUG is at http://www.sf-lug.org/
>
>
>
>
>
>
> _______________________________________________
> sf-lug mailing list
> sf-lug at linuxmafia.com
> http://linuxmafia.com/mailman/listinfo/sf-lug
> Information about SF-LUG is at http://www.sf-lug.org/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://linuxmafia.com/pipermail/sf-lug/attachments/20140827/ff0c5093/attachment.html>