supervision - discussion about system services, daemon supervision, init, runlevel management, and tools such as s6 and runit
 help / color / mirror / Atom feed
From: Gerrit Pape <pape@smarden.org>
To: supervision@list.skarnet.org
Subject: Re: aborting at stage 1?
Date: Tue, 19 Jun 2007 18:32:02 +0000	[thread overview]
Message-ID: <20070619183202.24886.qmail@f2f18f9094771c.315fe32.mid.smarden.org> (raw)
In-Reply-To: <Pine.LNX.4.64.0704282013160.28059@jmaa.math.ist.utl.pt>

On Sat, Apr 28, 2007 at 08:34:23PM +0100, Jorge Almeida wrote:
> On Sat, 28 Apr 2007, Vincent Danen wrote:
> >* Jorge Almeida <jalmeida@math.ist.utl.pt> [2007-04-28 13:20:21 +0100]:
> >>What happens if some vital init task fails? Suppose that checking the
> >>root filesystem fails. The system should halt, but how to do it? If the
> >>script ends with "init 0" in case the previous steps fail, then some
> >>changes must be made in /etc/runit, according to the man page of
> >>runit-init. But if the root file system is mounted read-only...
> >
> >This has pretty much nothing to do with runit.  For instance, my
> >/etc/runit/1 is:
> >
> Thank you for your reply, but I'm kind of lost here. The system I'm
> trying to setup is LFS. I don't want to use general boot scripts, I'm
> trying to keep the boot scripts (i.e., stages 1 and 3 scripts) simple
> and customized to my system (sacrifying generality). I noticed that the
> rc script in the link you provided uses /sbin/halt. In my Gentoo system,
> /sbin/halt belongs to the sysvinit package, so I suppose it will not be
> available in a sysvinit-free, pure runit, system. Is this correct? If
> so, the point is how to halt the system during stage 1 when something
> goes wrong and the root filesystem is not writable. From what I
> understood of the runit man pages, the only way to halt the computer
> when something goes wrong during stage 1 is to have the script
> /etc/runit/1 to issue "init 0". This will jump over stage 2 but will go
> to stage 3, which will need the root filesystem to be writable.

Hi,

if something goes wrong in stage 1, the /etc/runit/1 script should exit
100.  runit will then skip stage 2, and enter stage 3, see runit(8).
Possibly some things in stage 3 won't work if the root fs is mounted
read-only, but finally /etc/runit/3 will exit, and runit will either
halt or reboot after looking at /etc/runit/reboot.  For read-only
filesystems, runit supports /etc/runit/reboot being a symlink, even a
dangling one, which can point into a ramdisk mount point.  This applies
to all files/directories the runit programs need write access to BTW, if
not, I'd consider it as a bug.

HTH, Gerrit.


  parent reply	other threads:[~2007-06-19 18:32 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-04-28 12:20 Jorge Almeida
2007-04-28 16:57 ` Vincent Danen
2007-04-28 19:34   ` Jorge Almeida
2007-04-29  0:22     ` Mike Buland
2007-04-29 23:12       ` Vincent Danen
2007-04-30 10:06         ` Jorge Almeida
2007-06-19 18:32     ` Gerrit Pape [this message]
2007-06-19 19:02       ` Jorge Almeida

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20070619183202.24886.qmail@f2f18f9094771c.315fe32.mid.smarden.org \
    --to=pape@smarden.org \
    --cc=supervision@list.skarnet.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).