1 |
On Wed, 9 Jan 2013 14:48:33 +0000 (UTC) |
2 |
Grant Edwards <grant.b.edwards@×××××.com> wrote: |
3 |
|
4 |
> On 2013-01-09, Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
5 |
> > On Wed, 9 Jan 2013 02:47:07 +0000 (UTC) |
6 |
> |
7 |
> > The data on a medium can corrupt, and it can corrupt silently for a |
8 |
> > long time. |
9 |
> |
10 |
> And I'm saying I've never seen that happen. |
11 |
> |
12 |
> So you're saying that the data on a medium can corrupt without being |
13 |
> detected by the block encodings and CRCs used by the disk controller? |
14 |
|
15 |
No, I'm not saying that *at*all* |
16 |
|
17 |
I've been saying all along that data which you never go near can |
18 |
corrupt silently while you are not using it. When you do eventually get |
19 |
around to reading it, the electronics will do what they are designed to |
20 |
do and properly detect a problem that already happened. |
21 |
|
22 |
> |
23 |
> > At some point it may deteriorate to where it passes a cusp |
24 |
> > and then you will get your first visible sign |
25 |
> |
26 |
> No, the first visible sign in the scenario you're describing would be |
27 |
> a read returning erroneous data. |
28 |
|
29 |
That's what I said. The first VISIBLE sign is an error. You want to |
30 |
catch it before then. |
31 |
|
32 |
Analogy time: A murderer plans to do Grant. By observing Grant and only |
33 |
observing Grant, the first visible sign of an issue is the death of |
34 |
Grant. Obviously this is sub-optimum and we should be looking at a few |
35 |
more things than just Grant in order to preserve Grant. |
36 |
|
37 |
> |
38 |
> > - read failure. You did not see anything that happened prior as it |
39 |
> > was silent. |
40 |
> |
41 |
> If a read successfully returns correct data, how is it "silent"? |
42 |
|
43 |
I never used those words and never said "successfully returns correct |
44 |
data". At best I said something equivalent to "If a read returns". |
45 |
|
46 |
The point I'm trying hard to make is that all our fancy hardware merely |
47 |
gives an *apparency* of reliable results that are totally right or |
48 |
totally wrong. It looks that way because the IT industry spent much |
49 |
time creating wrappers and APIs to give that effect. Under the covers |
50 |
where the actual storage happens it is not like that, and errors can |
51 |
happen. They are rare. |
52 |
|
53 |
Lucky for us, these days we have precision machinery and clever |
54 |
mathematics that reduce the problem vastly. I know in my own case the |
55 |
electronics offer a reliability that far exceeds what I need so I can |
56 |
afford to ignore rare problems. Other people have different needs. |
57 |
|
58 |
|
59 |
-- |
60 |
Alan McKinnon |
61 |
alan.mckinnon@×××××.com |