1 |
Am Sun, 26 Mar 2017 14:16:56 +0200 |
2 |
schrieb Alan McKinnon <alan.mckinnon@×××××.com>: |
3 |
|
4 |
> On 26/03/2017 14:14, tuxic@××××××.de wrote: |
5 |
> > Hi, |
6 |
> > |
7 |
> > ok, seems that at least the data of one of the three partitions |
8 |
> > of my sdcard is toasted... |
9 |
> > |
10 |
> > But it would be interesting to check, what the initial (?) reason |
11 |
> > for that failure is: Hardware or logic -- sdcard or filesystem. |
12 |
> |
13 |
> that failure very seldom happens with spinning disks |
14 |
> it very often happens with SDcards |
15 |
> |
16 |
> chances are *huge* that it's the card itself |
17 |
> |
18 |
> > |
19 |
> > Is there any flash-memory-friendly and -aware checker out there, |
20 |
> > which detects sectors which cannot be recovered back to |
21 |
> > functioning? |
22 |
> |
23 |
> Not really, to my knowledge no-one has figured out a metric that |
24 |
> shows increased odds of pending failure for SD cards. |
25 |
> |
26 |
> They just stop working one day. A lot like fuses actually - there's |
27 |
> no way to examine a fuse and predict when it's likely to blow. |
28 |
> |
29 |
> Lesson to be learned: SD cards should only contain stuff you are |
30 |
> happy to lose, or of which you have several backups |
31 |
|
32 |
Many SD cards are optimized for FAT usage, thus they do wear leveling |
33 |
only where FAT systems do most writes and rewrites: In the file |
34 |
allocation table. Almost anything else works different. |
35 |
|
36 |
Maybe better try f2fs as it more likely spreads writes evenly across |
37 |
the SD card. |
38 |
|
39 |
-- |
40 |
Regards, |
41 |
Kai |
42 |
|
43 |
Replies to list-only preferred. |