Buchan Milne <[EMAIL PROTECTED]> writes:
> Guillaume Cottenceau wrote:
> > Following bug #3967, I'd like to add a
> > /usr/share/doc/postfix-2.0.12/MANDRAKE.SPECIFIC.CHROOT_README
> > file. I'd write something like the following. Please share with
> > me your wise comments on it.
>
> Instead of providing documentation on the long way to fix it, why not
> provide an easier way to fix it?
I have some concerns about that, please read my comment on said
bug.
> > -----8<----------8<----------8<----------8<----------8<-----
> >
> > For example, if your /etc/resolv.conf is updated when you launch
> > the network on your laptop, you will want to create a file named
> > /etc/sysconfig/network-scripts/ifup.d/postfix containing:
> >
> > -=-=---=-=---=-=---=-=---=-=---=-=--
> > #!/bin/sh
> > # update resolv.conf in postfix chroot environment
> > cp -f /etc/resolv.conf /var/spool/postfix/etc/resolv.conf > /dev/null
> > -=-=---=-=---=-=---=-=---=-=---=-=--
> >
> > You may want to update also other files depending on your
> > configuration.
>
> Why not have the postfix init script check an option in a /etc/sysconfig
> file for a descriptive variable (POSTFIX_ALWAYS_UPDATE_CHROOT), and if
> it is true, check every file under /var/spool/postfix/{var,usr,lib} to
> see if a newer one is outside the chroot, if so, update the one in the
> chroot.
When to do so? On boot? On ifup? On postfix start? On postfix
reload? My main concern about that is that I'm not sure there is
a so good default for that decision.
Another problem is that I'm not sure it's always good to maintain
a full correspondances between the files in chroot and outside.
And finally, I'm not sure it's good to have a default behaviour
that is too far away from default postfix one (so that our users
can read "default" postfix documentation and not be too lost),
that's why a quick doc would be a good solution IMHO.
> IMHO, postfix should be restarted when an interface comes up
> (if it was running).
Or reloaded? Or not (as Luca said)?
I think taking decisions for that default behaviour is somewhat
dangerous.
--
Guillaume Cottenceau - http://people.mandrakesoft.com/~gc/
