Bug 112807

Summary: Saving to FAT32: Corruption
Product: [Applications] kget Reporter: Xavion <Xavion.0>
Component: generalAssignee: KGet bugs <kget-bugs-null>
Status: RESOLVED NOT A BUG    
Severity: normal CC: nicolasg
Priority: NOR    
Version First Reported In: unspecified   
Target Milestone: ---   
Platform: unspecified   
OS: Linux   
Latest Commit: Version Fixed/Implemented In:
Sentry Crash Report:

Description Xavion 2005-09-18 03:32:16 UTC 
Version:           v0.8.4 (using KDE 3.4.2-1.2.fc4.kde, Fedora)
Compiler:          Target: i386-redhat-linux
OS:                Linux (i686) release 2.6.12-1.1447_FC4

I don't have any way to prove this, but I think KGet might be the cause.

I recently decided to put my "Downloads" directory onto my FAT32 partition and instructed KGet to filter all files to that dir by default.

Within a day or so (i.e. after a few downloads had been completed), I started getting errors on that drive and had to fix the errors using Windows.  This happened three times (after three downloading sessions).

I didn't have any such problems beforehand and the corruptions stopped occurring after I changed KGet's filters back to my Ext3 partition.

Is this possibly a KGet bug?
Comment 1 Nicolas Goutte 2005-09-18 03:39:10 UTC 
What kinds of errors? File name errors? (If yes, you should check how your partition is declared to Linux, especially the code page used on the file system and the "io charset" used by Linux.)
Comment 2 Xavion 2005-09-19 09:10:05 UTC 
Thanks for replying.

The files became corrupted in that directory (e.g. very strange filenames and massive filesizes) and one of the daemons was constantly using CPU time to try and read it (even with no apps running).  While this was happening, the rest of the partition automatically became read-only.  I had to reboot the machine into Windows (each time) and fix the errors from there.

Here's the relevant line from my /etc/fstab file:
  /dev/hda3 /mnt/Common vfat
    auto,users,rw,umask=0007,suid,gid=501,shortname=mixed

Where do I find out the other info (i.e. "code page used" and "io charset")?

Reg
Chris
Comment 3 Thiago Macieira 2005-09-19 12:38:27 UTC 
I don't think KGet can cause what you're describing even if it wanted to. This has to be a kernel bug or a disk hardware failure.

You'll have to talk to your distribution developers so that they can take it up to the kernel developers.