1 |
On Mon, Oct 14, 2013 at 07:38:19AM +0800, Patrick Lauer wrote: |
2 |
> On 10/14/2013 07:29 AM, Mike Gilbert wrote: |
3 |
> > On Sun, Oct 13, 2013 at 7:21 PM, Patrick Lauer <patrick@g.o> wrote: |
4 |
> >> On 10/14/2013 03:32 AM, William Hubbs wrote: |
5 |
> >>> All, |
6 |
> >>> |
7 |
> >>> from what I'm seeing, we should look into converting /etc/mtab to a |
8 |
> >>> symlink to /proc/self/mounts. |
9 |
> >>> |
10 |
> >>> Are there any remaining concerns about doing this? |
11 |
> >> |
12 |
> >> Apart from breaking umount -a and some other things? |
13 |
|
14 |
What "other things"? AFAICT from the debian bug report[1] problems are far |
15 |
much more likely when it's not a symlink, unless you're on kernel <2.6.26 |
16 |
(which is admittedly more likely for a Gentoo user, but it would not be the |
17 |
mainstream Gentoo user, by any stretch of the imagination.) |
18 |
|
19 |
> >> (The breakage is visible e.g. with umount -a tmpfs, which used to be |
20 |
> >> quite useful if you had a few chroots with /var/tmp/portage as tmpfs and |
21 |
> >> wanted to reset them. Now it'll also punt random things like /run if |
22 |
> >> you're lucky - and in the past it knocked out the OpenRC state directory |
23 |
> >> reliably) |
24 |
> >> |
25 |
> > |
26 |
> > I don't follow this: it seems like umount -a is supposed to unmount |
27 |
> > all filesystems. umount -a -t tmpfs would unmount all tmpfs |
28 |
> > filesystems. /run should be included in that set, even if mtab is a |
29 |
> > regular file. |
30 |
> > |
31 |
> |
32 |
> And the magic trick is to keep "system mounts" like /run out of |
33 |
> /etc/mtab (willful desynchronization) so that umount -a doesn't nuke |
34 |
> them by accident. |
35 |
> |
36 |
> ... why else would you keep such data in two non-synchronized locations?! :D |
37 |
> |
38 |
|
39 |
You wouldn't: the only reason I've clung to the idea of a file, are the |
40 |
transitional problems mentioned in the debian bug, ie information not available |
41 |
in /proc/mounts that is available in /etc/mtab. Clearly that was fixed 2 years |
42 |
ago, including for NFS and smb. |
43 |
|
44 |
Given the CLONE_NEWNS issue (which one might use in the situation you describe): |
45 |
"At this point, /etc/mtab *must* be a symlink to avoid breakage. Note that |
46 |
/proc/mounts is now a symlink to /proc/self/mounts for this reason: each process |
47 |
has potentially different mounts"; going forward it really does not seem like a |
48 |
good idea not to default to the symlink. |
49 |
|
50 |
I agree it should not be forced on existing installs, nor on an admin who wants |
51 |
to do something funky (or is using an ancient kernel for some reason.) However |
52 |
I would definitely support a news item about the default change, along with |
53 |
a recommendation for existing installs also to change, unless the admin has |
54 |
a reason not to. |
55 |
|
56 |
Just as a heads-up, that the ground has changed, and the vast majority really |
57 |
do not want to be running without that symlink. |
58 |
|
59 |
And yeah, script that properly (or a new namespace if it works) you lazy git ;p |
60 |
|
61 |
Regards, |
62 |
steveL. |
63 |
|
64 |
[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=494001 |
65 |
|
66 |
-- |
67 |
#friendly-coders -- We're friendly, but we're not /that/ friendly ;-) |