[Mageia-dev] GRUB can't see HD from chroot unless parent /dev is bind-mounted in chroot ?

[Colin Guthrie]

'Twas brillig, and Bruno Cornec at 17/01/13 10:35 did gyre and gimble:
> Colin Guthrie said on Thu, Jan 17, 2013 at 09:45:27AM +0000:
> 
>> If you would like to detail the problems with squid and point at a
>> detailed bug report I can take a look or at least advise on the best
>> steps to take to get a really solid system.
> 
> I've detailed my feedback here:
> http://brunocornec.wordpress.com/2013/01/07/a-mageia-based-firewall-with-auto_inst-and-lots-of-other-stuff-like-chrooted-squid/

I've not read fully, but you don't even appear to have written a systemd
unit to tell systemd about your chrooted environment or customised the
init script for squid to run inside a chroot.

Systemd has one of the best set of documentation available on any
project I've worked on. There are so many man pages, wiki posts and
blogs documenting the various features.

I appreciate you may not like it but it really simplifies things
generally. Instead of two screens of bash you need to read you've got
half a dozen attributes to read.

As detailed in the link I already sent it should be quite trivial to
configure things in a chroot and systemd is expecting to find a pid file
from somewhere then that will be listed in the initscript or systemd
unit both of which can be easily customised to look for that file
directly inside the chroot.

In the case of squid, it's the "pidfile" line of the initscript:

http://svnweb.mageia.org/packages/cauldron/squid/current/SOURCES/squid.init?view=markup

See line 12.

Just editing that and including the correct path would have "made
systemd happy".


Really this is just a (perfectly understandable) lack of understanding.
I appreciate with any change it takes an amount of effort to learn new
things. But really sysvinit was not "simple". There are lots of gotchas
to watch out for (like a user doing a restart and having the daemon
inherit the users execution environment which can have many weird
effects and fallouts and which will cause things to behave differently
on next boot; and like not knowing if some process is started by an
initscript or via inittab and why you should use a script vs. inittab
entry (the latter having some kind of trivial restart logic built in
being the usual reasoning)). Lots of things come out of the woodwork
when you really dig in to it. Systemd's not perfect and I accept that
but it is getting there and the breadth and depth of the documentation
is still one of the major benefits IMO. sysadmins of tomorrow should
have a much, much clearer understanding of what's going on on their
systems and how they work.

YMMV of course :)

Col

-- 

Colin Guthrie
colin(at)mageia.org
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited http://www.tribalogic.net/
Open Source:
  Mageia Contributor http://www.mageia.org/
  PulseAudio Hacker http://www.pulseaudio.org/
  Trac Hacker http://trac.edgewall.org/