[gentoo-dev] Re: rfc: improve file system mounting and unmounting in OpenRC

[Duncan]

Ian Stakenvicius posted on Tue, 04 Aug 2015 17:17:51 -0400 as excerpted:

> So what you are suggesting here now is that you want to (A) potentially
> break mounting with the need to externally manage mounts via services in
> openrc instead of just using /etc/fstab, and (B) also break services if
> something doesn't start, which is one of the reasons why you wanted to
> go through with this per-mount service in the first place.  My point is
> that no, we should keep localmount as succeeding even if one of the
> dependent services fails to mount, *just like it does right now*, *for
> the same reasons* as it succeeds on failure right now.

+1

IMO, localmount must continue to succeed /by/ /default/, even if some 
mounts fail, because it's basically legacy, and must maintain legacy 
behavior.  Turning it into a wrapper "internally" is fine, but the 
overall localmount must still succeed, as too much depends on that 
behavior as it is.

That also solves the problem (at least as long as the localmount legacy 
now-wrapper continues to exist) of people with all sorts of strange 
layouts having to adjust their dependencies manually when they upgrade, 
as if they don't do anything, the localmount wrapper will still continue 
to function as it always has.

Meanwhile, once the individual mount services are available, people who 
want a more precise boot can setup individual dependencies as they wish, 
and then flip a localmount option if they like that lets it fail, but it 
will _only_ fail if they flip that option.  Critically, when they do so, 
they can also configure exactly which mounts localmount can fail on, with 
(implementor's choice) either (1) other mounts continuing to always 
succeed, or (2) other mounts being setup as entirely separate services, 
that localmount doesn't depend on or wrap any longer.

The general idea is that users can setup the separate mount deps for 
services and mounts they really care about.  But localmount will still 
let them glob those they don't care about together to handle as one, and 
will let them decide whether that glob can as a whole fail, or not, 
depending on what way their "don't care" goes.

Again, if users do nothing, localmount as a wrapper continues to glob all 
(not noauto) localmounts, just as it does now, and continues to succeed, 
just as it does now.  At that point, however, it'll be possible to break 
individual mounts out of the glob if desired, but critically, it'll be 
user's responsibility to configure deps as appropriate, if they do.

And any current localmount-depending services will by default continue to 
depend on local-mount, so it'll continue to work just as it does now, 
unless users actually care enough to break out the individual mounts and 
setup deps for them accordingly.

However, as an aid to doing so, "ported" services, instead of killing 
their localmount dep, will keep it, but will have a nice comment listing 
exactly what directories they use, so users who want to setup individual 
mount deps will know exactly what mount deps they need, depending on 
their system-unique mountpoint setup.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman