1 |
On 2013-01-08, Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
2 |
|
3 |
>> When a hard drive starts to fail, you don't unknowingly get back |
4 |
>> "rotten" data with some bits flipped. You get either a "seek error" |
5 |
>> or "read error", and no data at all. IIRC, the same is true for |
6 |
>> attempts to read a failing CD. |
7 |
> |
8 |
> I see what Florian is getting at here, and he's perfectly correct. |
9 |
> |
10 |
> We techie types often like to think our storage is purely binary, the |
11 |
> cells are either on or off and they never change unless we |
12 |
> deliberately make them change. We think this way because we wrap our |
13 |
> storage in layers to make it look that way, in the style of an API. |
14 |
> |
15 |
> The truth is that our storage is subject to decay. Harddrives are |
16 |
> magnetic at heart, and atoms have to align and stay aligned for the |
17 |
> drive to work. Floppies are infinitely worse at this, but drives are |
18 |
> not immune. Writeable CDs do not have physical pits and lands like |
19 |
> factory original discs have, they use chemicals to make reflective and |
20 |
> non-reflective spots. The list of points of corruption is long and |
21 |
> they all happen after the data has been committed to physical storage. |
22 |
|
23 |
True. But, in my experience, the chances of any of those failures |
24 |
resulting in a successful read of incorrect data is vanishly small. |
25 |
|
26 |
> Worse, you only know about the corruption by reading it, there is no |
27 |
> other way to discover if the medium and the data are still OK. He |
28 |
> wants to read the medium occasionally |
29 |
|
30 |
That may be a good idea, and will detect media failures. |
31 |
|
32 |
> and verify it |
33 |
|
34 |
That's the part I think is pointless in practice (if you're trying to |
35 |
detect failing media). |
36 |
|
37 |
> while the backups are still usable, and not wait for the point of no |
38 |
> return - the "read error" from a medium that long since failed. |
39 |
|
40 |
My point is that _comparing_data_to_a_backup_ just isn't a useful, |
41 |
practical way to detect failing hard drives, optical drives, or CDs. |
42 |
I've seen a lot of hard drives, optical drives, floppy drives, |
43 |
flopies, and CDs fail. The failure mode in every case has been a "seek |
44 |
error" or "read error" resulting in _no_data_ rather than a read |
45 |
returning erroneous data. |
46 |
|
47 |
It seems that in laboratory conditions, people have managed to see |
48 |
erroneous data, but I'm not convinced worrying about it is worthwhile. |
49 |
|
50 |
IMO, having backup data _is_ very valuable, but regularly reading |
51 |
files and comparing them to backup copies isn't a useful way to detect |
52 |
failing media. |
53 |
|
54 |
You're much more likely to detect failing RAM (which is useful, but |
55 |
there are better ways to do it). |
56 |
|
57 |
-- |
58 |
Grant Edwards grant.b.edwards Yow! I think I am an |
59 |
at overnight sensation right |
60 |
gmail.com now!! |