Re: repeated failure to properly shutdown

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Sat, 23 Jul 2016 01:06:14 +0200
    From:        Rhialto <rhialto%falu.nl@localhost>
    Message-ID:  <20160722230614.GB4434%falu.nl@localhost>

  | Similarly, I noticed that if /var is a tmpfs (or any initially empty
  | directory really), then /etc/rc.d/mountcritlocal fails because it wants
  | to cd to /var/run and that has not been created (if that ever happens).

I'm not sure I'd call that one a bug - an empty /var is missing all
kinds of stuff that is likely to be needed - I think if you want to
run that as a tmpfs, you'd also need to supply a mtree file to
populate it correctly, and run that as part of the setup - though
I'll admit, expecting that to happen before mountcritlocal runs might
be asking too much.

Does mountcritlocal really fail though?   All it appears to be doing is
removing the contents of /var/run - in a sub-shell.   An error message
when the cd fails I'd expect, and the sub-shell will exit because of that,
but the rest of muntcritlocal should continue I'd expect - not that there is
any more of it that will do anything useful, clearing /var/run happens after
the critical filesystems are mounted - which it must, as /var might be one
of them.

About the only problem I see is that /var/run/lvm will fail to be created
if /var/run does not exist.   Perhaps mountcritlocal should create /var/run ?
(would just mean s/mkdir/mkdir -p/) - and perhaps either run the subshell
that clears /var/run with stderr redirected to /dev/null, or only run it
if /var/run exists ?

kre