We are in the process of migrating this forum. A new space will be available soon. We are sorry for the inconvenience.

File corruption: HEAD request is ok but GET fails


alex89k
10-17-2016, 04:25 PM
Hi,
I have a 10TB account and at least one file that I put in it is no longer downloadable.

The file is correctly listed in the parent directory with correct metadata (size, name, md5).
HEAD requests receive a correct response with right metadata (size, name, md5).
GET requests receive a 404 Not Found error.
The file is not segmented.
I see this behaviour from any client I use: Hubic web client, Swift Explorer, manual HTTP requests according to Swift specifications, my Java software that I'm developing.

The account is alessandro dot brencich at gmail dot you-know-what.
The file path is "default/Light/00356.dat".

The (obvious) bad thing is that the user discovers that the file is not recoverable only when he tries to download it, probably when he needs it because he has no other copy of it. (Ok, he should always have another copy, I know...)

I want to clarify that I'm not "shocked" for the incident. Hubic offers 10 TB for 50 euros while other services like GD or DB would want over 1.000 euros (over 20 times the Hubic fee!) for the same space. A little corruption can happen (I saw it on 2 big files, while over 20.000 other files are ok). The only things I would like to know are:

(1) Is the file really corrupted forever or this is a temporary condition and the file will be recovered?

(2) In the first case (the file is definitively corrupted) is there a way to know it? I'm developing a general Hubic backup software and wouldn't be a great effort to implement a re-upload of one file if it gets corrupted... but the software should have a way to know it (possibly different from downloading the whole file every week only to know if it is still there).

Thank you.