1 |
Albert Hopkins <marduk@×××××××××××.org> writes: |
2 |
|
3 |
> On Friday, July 8 at 11:55 (-0500), Harry Putnam said: |
4 |
> |
5 |
> [..] |
6 |
>> Somehow I managed to really hurt the installation ... here is what I |
7 |
>> remember having done: |
8 |
>> |
9 |
>> Some how I got mixed up when running as root, and attempted to mount a |
10 |
>> users encfs directory. (Its a single user machine so it my users |
11 |
>> directory) |
12 |
>> |
13 |
>> That should just fail with some kind of permission error since no one, |
14 |
>> even root, can mess with someone elses' encfs directory. |
15 |
>> |
16 |
> This is not entirely the case. No one can enter an encfs mount |
17 |
> (destination) but the person that mounted it (by default), but anyone |
18 |
> can *mount* the encfs "source" (and thus become the owner of the mount). |
19 |
> All other Unix permissions are retained. |
20 |
> |
21 |
>> But once I'd done that I could no longer even `ls' the subject |
22 |
>> directory. Not as user, not as root. A simple `ls' would totally |
23 |
>> hang the terminal. |
24 |
>> |
25 |
>> Of course I tried to umount but really it never actually mounted. |
26 |
>> |
27 |
> It probably *did* mount, but... |
28 |
> |
29 |
>> I started getting this error: `Transport endpoint is not connected' |
30 |
|
31 |
> Usually that means the background process that actually handles the |
32 |
> enc/dec has died or is otherwise not responding. |
33 |
|
34 |
>> I could see roots attempt to mount the darn thing in ps wwaux output |
35 |
>> and killed that pid. |
36 |
>> |
37 |
> That's probably why you got the above error. But technically if you |
38 |
> just kill the process, the kernel still thinks it's mounted. |
39 |
|
40 |
>> Eventually (after posting several days ago on encfs list. I resorted |
41 |
>> to umounting /home (after full backup of course) and reformatted it. |
42 |
>> |
43 |
> That's seems a bit extreme... |
44 |
> |
45 |
>> I was then able to deleted encfs_raw and encfs_mnt. |
46 |
>> |
47 |
> Did you try simply rebooting or manually unmounting? That's probably |
48 |
> all that was needed. |
49 |
|
50 |
Yes, both. First the umounting, then reboot. So no I don't think it |
51 |
was all that was needed. |
52 |
|
53 |
>> But here is the real kicker. Even after all that, and in fact another |
54 |
>> full round of mostly the same stuff, including another reformat. So |
55 |
>> two reformats and two reboots. Even with that, I still cannot create a |
56 |
>> new enc_raw and enc_mount of the same name as the old one. |
57 |
|
58 |
> What do you mean by "cannot"? Do you get an error? Does dmesg tell you |
59 |
> anything? |
60 |
|
61 |
I meant the earlier behavior would recur, without the `Transport |
62 |
endpoint is not connected' part. That is, mkdir ~/.encjunk ~/.junk |
63 |
|
64 |
encfs ~/.encjunk ~/.junk Would result in completely hung |
65 |
terminal when I answered the passwd prompt. |
66 |
|
67 |
[...] |
68 |
|
69 |
> If you totally remove the source and target directories, there is no |
70 |
> other information stored, which allows you to (e.g) encfs directory on a |
71 |
> vfat-formated USB stick and move it mount it on a different machine. |
72 |
> All that encfs knows about an encrypted directory is in *one* file on |
73 |
> the source directory (.encfs6.xml). Once that file is gone, there is no |
74 |
> such thing as an encfs. |
75 |
|
76 |
There must have been plenty of pilot error in the original goings on. |
77 |
I tried to report it as it actually happened but I must have had some |
78 |
of it wrong or whatever... because just now, prompted by your input. |
79 |
(and there have been at least two reboots (as mentioned earlier)) |
80 |
|
81 |
I was able to generate the desired mount name with no problems. |
82 |
|
83 |
So don't have to rework any scripts or continue puzzling over what was |
84 |
going on. In other words... as has happened a few other times, |
85 |
writing about the problem has seemed to cure it.. |
86 |
|
87 |
I'm good to go now ... thanks. |