[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: systemd: some more questions



Tom H writes:
 > On Thu, Feb 13, 2014 at 3:12 AM, Gian Uberto Lauri <saint@eng.it> wrote:
 > >
 > > let's assume that I have to execute a script after S10checkfs/
 > > S11mountall and before S13networking.
 > >
 > > Do I find a clear, reliable set of instructions to achieve this goal?
 > >
 > > At this time my fear is that systemd turns out as a very nice tool for
 > > personal pc but a PITA for servers/vm (as happened for upstep).
 > 
 > You've chosen an interesting case! :)
 > 
 > My first reaction was to write: "If I were you I'd create a sysvinit
 > script that'll be slotted in between mountall and networking by
 > insserv rather than try to figure out how to order it in between S10
 > and S13 by systemd directly."
 > 
 > But I thought that given that "/etc/init.d/mountall.sh" is an early
 > boot service, I checked and found that it's overridden in systemd by
 > an empty "/lib/systemd/system/checkfs.service".
 > 
 > So I'd check what systemd unit (units) does (do) the job of
 > mountall.sh and create a systemd service that Requires=/Wants= and is
 > After= that unit (those units) as well as WantedBy=/RequiredBy= and
 > Before= "network.target".

The only thing that is "wrong" in your plan is "So I'd check what
systemd unit (units) does (do) the job of mountall.sh".

The "right" thing would have been "So I'd check the documentation to
see what unit...": one place with things explainev vs. what looks like
reverse engineering.

That's not for me, some more tinkering or some less tinkering does not
change the things this much. 

But I fear a situation like this:

Random J. Hacker: "We could use a Debian stable server for this, but I
	   	  		   have to go and see how to port our boot script ..."
Project M. Anager: "Go and see?"
Random J. Hacker: "Yes, they changed the boot process and the documentation
	   	  		   does not cover all"
Project M. Anager: "No, we will stick to $whatsoever, we have no budget for
		   		    with old sysV init, we have non budget for..."

[Please note: mrs. Project M. Anager at least knows what she is
 talking about]

On an issue like this documentation is critical. Let's hope that RH
interest in systemd will give us this kind of documentation...

-- 
 /\           ___                                    Ubuntu: ancient
/___/\_|_|\_|__|___Gian Uberto Lauri_____               African word
  //--\| | \|  |   Integralista GNUslamico            meaning "I can
\/                 coltivatore diretto di software       not install
     già sistemista a tempo (altrui) perso...                Debian"

Warning: gnome-config-daemon considered more dangerous than GOTO


Reply to: