disk full, bug identified
Posted: 23 Oct 2020, 15:10
Hello all,
I hope I didn't "discover" something you guys have been discussing here earlier. Anyways it happened to me just like to many others in threads on this forum, I was recovering a 2TB disk and got my 4TB disk filled but not all data was recovered.
The explanations I found here were about testdisk not finding EOF in corrupted files and behaving grabby by creating large files until the next file header... BUT
The real error was another in my case ... while some files were indeed a bit (not by too much) bigger than others what was really happening was some kind of looping, or better explained as wrong pointing to the next read sector.
i.e. in one of my photo folders one of the .JPG images was interpreted as a folder, more specifically as a hardlink to root / folder of the disk so every here and then testdisk would misinterpret a file as a link and in start a recovery of the whole disk all over into that folder, interestingly on the next iteration it would ignore the bad file and finish there where it started ... something like:
recover: file.jpg
file.jpg header ----> /the/whole/disk/up/until/ ---> file.jpg eof
So far i had to delete this looped trees 6 times to free up space and restart a recovery from the next file ... but every here and there this happens and fills the disk
p.s. you can move all the files from the nested folder to the root folder because there are a lot of files recovered in the nested one that are not in the original
Anyone has an idea about this ?
Thx
I hope I didn't "discover" something you guys have been discussing here earlier. Anyways it happened to me just like to many others in threads on this forum, I was recovering a 2TB disk and got my 4TB disk filled but not all data was recovered.
The explanations I found here were about testdisk not finding EOF in corrupted files and behaving grabby by creating large files until the next file header... BUT
The real error was another in my case ... while some files were indeed a bit (not by too much) bigger than others what was really happening was some kind of looping, or better explained as wrong pointing to the next read sector.
i.e. in one of my photo folders one of the .JPG images was interpreted as a folder, more specifically as a hardlink to root / folder of the disk so every here and then testdisk would misinterpret a file as a link and in start a recovery of the whole disk all over into that folder, interestingly on the next iteration it would ignore the bad file and finish there where it started ... something like:
recover: file.jpg
file.jpg header ----> /the/whole/disk/up/until/ ---> file.jpg eof
So far i had to delete this looped trees 6 times to free up space and restart a recovery from the next file ... but every here and there this happens and fills the disk
p.s. you can move all the files from the nested folder to the root folder because there are a lot of files recovered in the nested one that are not in the original
Anyone has an idea about this ?
Thx