1 |
On 26/03/17 17:24, tuxic@××××××.de wrote: |
2 |
> On 03/26 04:50, Bill Kenworthy wrote: |
3 |
>> On 26/03/17 15:26, tuxic@××××××.de wrote: |
4 |
>>> On 03/26 03:04, Bill Kenworthy wrote: |
5 |
>>>> On 26/03/17 14:25, tuxic@××××××.de wrote: |
6 |
>>>>> On 03/26 05:50, tuxic@××××××.de wrote: |
7 |
>>>>>> On 03/26 11:21, Adam Carter wrote: |
8 |
>>>>>>> Step 1: dd the contents into an image |
9 |
>>>>>>> |
10 |
>>>>>>> ddrescue is probably a better option than plain dd. |
11 |
>>>>>>> |
12 |
>>>>>>> step 2: put the sdcard to one side. |
13 |
>>>>>>>> step 3: loopback mount a copy of the image (not the original) |
14 |
>>>>>>>> step 4: try recovering the filesystem on the loopback, if it fails ... try |
15 |
>>>>>>>> something else on another image copy |
16 |
>>>>>>> |
17 |
>>>>>>> |
18 |
>>>>>>> Yep, once you've got the image mounted loopback, you can run |
19 |
>>>>>>> testdisk/photorec depending on how bad it is. |
20 |
>>>>>> |
21 |
>>>>>> Hi all, |
22 |
>>>>>> |
23 |
>>>>>> thanks a lot for all help! :) |
24 |
>>>>>> |
25 |
>>>>>> Currently I am ddresucueing the flashcard to the harddisc. |
26 |
>>>>>> Next I will try to mount the sdcard. |
27 |
>>>>>> |
28 |
>>>>>> What reliable sdcard-reader can one recommend ? |
29 |
>>>>>> (...sorry if this sentence sounds harsh...I it by no means meant |
30 |
>>>>>> that way...I am no native speaker... :) |
31 |
>>>>>> |
32 |
>>>>>> Cheers |
33 |
>>>>>> Meino |
34 |
>>>>>> |
35 |
>>>>>> |
36 |
>>>>> |
37 |
>>>>> Hi, |
38 |
>>>>> |
39 |
>>>>> Is the assumption correct, that -- if ddrescue could read each |
40 |
>>>>> partitions of the sdcard without stuttering, retries and errors -- |
41 |
>>>>> the sdcard itsself is ok and "only" the logical structure |
42 |
>>>>> (fs, superblock etc) got damaged? |
43 |
>>>>> Or do I overlook something? |
44 |
>>>>> |
45 |
>>>>> (Background: I dont want to put a sdcard into the bin, if |
46 |
>>>>> fdisking & reformatting that beast would gives me back an ok |
47 |
>>>>> media...) |
48 |
>>>>> |
49 |
>>>>> Cheers |
50 |
>>>>> Meino |
51 |
>>>>> |
52 |
>>>>> |
53 |
>>>>> |
54 |
>>>>> |
55 |
>>>> |
56 |
>>>> The dd gets you the best chance to work on the data before it completely |
57 |
>>>> fails. In my experience the sdcard will only get worse ending with total |
58 |
>>>> failure - if it hasn't already. |
59 |
>>>> |
60 |
>>>> If the dd dump comes up rubbish and cant be recovered, the actual sdcard |
61 |
>>>> will be worse. You can run "strings" against the image to see if there is |
62 |
>>>> any text in there (or even cat the /dev/sdcard node through strings) to see |
63 |
>>>> if the bits are still there. |
64 |
>>>> |
65 |
>>>> I dont know of a cdparanoia type recovery utility for sdcards but I suspect |
66 |
>>>> sdcard design means that approach wont work. |
67 |
>>>> |
68 |
>>>> BillK |
69 |
>>>> |
70 |
>>>> |
71 |
>>>> |
72 |
>>> |
73 |
>>> Hi Bill, |
74 |
>>> |
75 |
>>> I got mixed results: There are three partitions on the sdcard from |
76 |
>>> which I could fully recover (even mount it directly via loop device) |
77 |
>>> the first and the third one. |
78 |
>>> |
79 |
>>> The second one is screwed up. |
80 |
>>> |
81 |
>>> Running fsch.ext4 against the image it starts with "bad superblock" |
82 |
>>> and suggests two alternatives. |
83 |
>>> |
84 |
>>> I started fsch.ext4 again while using -b to define the alternate |
85 |
>>> superblock and it starts to ask me *zillions of question, which |
86 |
>>> I all answered with 'yes' in a first attempt (I have a backup of the |
87 |
>>> image...). |
88 |
>>> The result was an image, which I could mount again. |
89 |
>>> But beside 'lost+found' with some small rests of something which |
90 |
>>> may be files in a previous life nothing was there.... |
91 |
>>> |
92 |
>>> Currently it looks to me, that something has totally messed up the fs |
93 |
>>> there. |
94 |
>>> |
95 |
>>> What do you think? |
96 |
>>> |
97 |
>>> Cheers |
98 |
>>> Meino |
99 |
>>> |
100 |
>>> |
101 |
>>> |
102 |
>>> |
103 |
>>> |
104 |
>> Sounds like its toast :( |
105 |
>> |
106 |
>> I have never had a lot of luck with any of the ext file systems - you have |
107 |
>> to baby them and they corrupt very easily compared to others. I try and |
108 |
>> avoid them ... |
109 |
>> |
110 |
>> BillK |
111 |
>> |
112 |
>> |
113 |
>> |
114 |
> |
115 |
> Hi Bill, |
116 |
> |
117 |
> ...the SDcard is for my Android tablet, which runs kernel 3.6.x.something (Lollipop) |
118 |
> if I remember correctly. With the App Linux Deploy I installed Linux |
119 |
> on another partition of the sdcard and used to chroot into it. |
120 |
> |
121 |
> With what filesystem did you made good experiences of, Bill? My GENTOO PC |
122 |
> (with which I am currently writing this email) also uses ext4...and |
123 |
> your last mail makes me nervous...very nervous... |
124 |
> |
125 |
> For my tablet I have to use an filesystem, which is supported by and |
126 |
> compiled into the kernel. Unfortunately, there is no alternative |
127 |
> Android build for this tablet... |
128 |
> In this case it is ext4 and vfat...and of that both I think ext4 is |
129 |
> better... |
130 |
> |
131 |
> Cheers |
132 |
> Meino |
133 |
> |
134 |
> |
135 |
|
136 |
I used to prefer reiserfs ... more robust in some ways, less in others |
137 |
... BUT I almost always was able to rescue files or whole filesystems |
138 |
with reiserfs, not so with ext*. These days I prefer btrfs ... again |
139 |
not perfect but with OS disks are single and the rest are bcache/ssd |
140 |
fronted btrfs raid10's my main issue is running out of space. |
141 |
|
142 |
Problems with the ext series were inability to deal with dirvish backup |
143 |
(corrupted), running out of inodes, terminal corruption when running out |
144 |
of space, silent corruption with hibernation, data loss/corruption on |
145 |
abrupt power loss, and it goes on ... |
146 |
|
147 |
|
148 |
Current sdcards are either vfat for win compatibility (no choice) or |
149 |
btrfs (raspberry pi). Just turned off my older rpi model 1B which is |
150 |
ext 4 earlier this morning - been corrupted and reimaged a few times! |
151 |
|
152 |
File systems seem to be very much YMMV - each use, load and environment |
153 |
present a different set of requirements and problems. |
154 |
|
155 |
BillK |