| 1 |
On Mon, Aug 03, 2015 at 04:38:59PM -0700, Daniel Campbell (zlg) wrote: |
| 2 |
> -----BEGIN PGP SIGNED MESSAGE----- |
| 3 |
> Hash: SHA256 |
| 4 |
> |
| 5 |
> On 08/03/2015 12:47 AM, Brian Dolbec wrote: |
| 6 |
> > On Mon, 3 Aug 2015 00:22:42 -0700 "Daniel Campbell (zlg)" |
| 7 |
> > <zlg@g.o> wrote: |
| 8 |
> > |
| 9 |
> > |
| 10 |
> >> I'm having a hard time understanding why we need daemons to |
| 11 |
> >> handle our filesystems. Can you give me a use case that |
| 12 |
> >> /etc/fstab is insufficient for solving? |
| 13 |
> > |
| 14 |
> >> - -- Daniel Campbell - Gentoo Developer |
| 15 |
> > |
| 16 |
> > |
| 17 |
> > It is about defining proper dependencies and not blindly returning |
| 18 |
> > a success result when there were actual failures to start some |
| 19 |
> > files systems. So in some ways it is a bugfix. But it is actually |
| 20 |
> > a re-design which will overcome shortcomings/limitations in the |
| 21 |
> > fstab, netmount, localmount designs. |
| 22 |
> > |
| 23 |
> > Net result should be better configurability, proper error |
| 24 |
> > reporting, proper service order startup,... |
| 25 |
> > |
| 26 |
> > Downside, it will likely mean a little migration/transistion. |
| 27 |
> > |
| 28 |
> > I'm in favour of the change. Good work William. |
| 29 |
> > |
| 30 |
> > |
| 31 |
> I'm okay with a change as long as it's relatively manageable and |
| 32 |
> offers some real benefits. If I understand correctly, this new |
| 33 |
> mounting will allow us to declare mounting dependencies the same way |
| 34 |
> we declare service/daemon dependencies, correct? |
| 35 |
> |
| 36 |
> So say I want to have an ownCloud instance that provides a single /usr |
| 37 |
> or /etc for any Gentoo system that wants it on my local network. Is |
| 38 |
> that a use case that would benefit from this new mounting? I'm just |
| 39 |
> trying to understand which use cases benefit and why, and what it is |
| 40 |
> that fstab isn't good enough for right now. As a developer, I want to |
| 41 |
> be able to support users on this if/when it hits mainline OpenRC. |
| 42 |
|
| 43 |
fstab is *not* going anywhere. |
| 44 |
|
| 45 |
The difference right now is that you just have two services that control |
| 46 |
all file system mounts and imo do a bad job of it. ;-) |
| 47 |
|
| 48 |
netmount and localmount always succeed, regardless of whether anything |
| 49 |
they mount fails. |
| 50 |
|
| 51 |
Under the new system, you will have services like mount.home, mount.usr, |
| 52 |
mount.var etc, which will actually be able to report failure if they do |
| 53 |
not mount their file systems. |
| 54 |
|
| 55 |
localmount and netmount will be kept, for now, but you will have to |
| 56 |
configure them to have dependencies like |
| 57 |
|
| 58 |
rc_need="mount.foo mount.bar mount.bas" etc, depending on which file |
| 59 |
systems are local or network. |
| 60 |
|
| 61 |
These versions of localmount and netmount will also change behaviours, |
| 62 |
because they will be able to fail if a filesystem they need fails to |
| 63 |
mount. |
| 64 |
|
| 65 |
Does that make sense? |
| 66 |
|
| 67 |
William |
Archives