Description of problem: I have mounted two NTFS windows partitions with /etc/fstab. When I save data on one of my partitions quite often (for example when I'm programming and save my code every time before I execute it) it happens that my system doesn't shut down completely, it hangs by the latest halt message "Halting System...". Above the "Halting System..." message, an error of ntfs-3g occurs: ntfs-3g: Unknown option '-n'. ... I have noticed this error the second time within one week. (I hope that this belongs to the component kernel) Version-Release number of selected component (if applicable): kernel 2.6.23.1-49.fc8 How reproducible: I'm sorry but I'm not sure why this error occurs. I think it's because the NTFS partitions couldn't be clearly unmounted. The error that is displayed can be created with "ntfs-3g -n" in a terminal Actual results: System doesn't turn off, hangs by "Halting System..." Have to hard-reset my computer Expected results: Shouldn't hang ^^ Additional info: Because I haven't found any log entry: I have got a photo of my screen that shows this error. This was taken by a mobile phone camera, but it's readable xD.
Created attachment 271601 [details] "Screen Shot" of the error
Created attachment 271611 [details] /etc/fstab
same problem here
Very strange. It's as if fstab-decode is generating -n as the option it needs for the halt script to perform the ntfs umount. Bill, does this make any sense to you?
What do you have in /proc/mounts normally?
I suppose the simplest way to debug this is to slap a set -x in __umount_loop in /etc/init.d/functions and see what the output is.
I notice that you're running ntfs-3g-1.1030, which is a bit out of date. I was unable to reproduce this with the latest package for f8, which is ntfs-3g-1.1120-1. To be fair, I downgraded to ntfs-3g-1.1030, and I can't reproduce it with that package either.
I'm not an expert at shell-scripting at all, but this error could be generated by the halt script: line 193-197 # Remount read only anything that's left mounted. # echo $"Remounting remaining filesystems readonly" mount | awk '{ print $3 }' | while read line; do fstab-decode mount -n -o ro,remount $line done So it's possible be that the ntfs partition couldn't be unmounted, so it will be remounted read-only, but that fails. But that's just a wild guess! (please ignore if that's not possible) xD
If ntfs mount doesn't support -n, it really needs to...
Szaka, ntfs-3g doesn't support the -n option to mount. Would you be opposed to adding it? (The current code is doing some odd "ignore mtab" options that take effect at compile time, not as a runtime option)
No. But the fundamental problem here is that remount is not supported and doing so can cause data loss the same way as if the partition weren't unmounted at all. Please just use unmount at the moment.
We cannot hardcode different options for different file systems, especially since we're just calling 'normal' mount/unmount - if it can't support it, it should just ignore it.
Created attachment 296100 [details] patch for ntfs-3g to ignore -n mount opt
This is resolved in F-9 and rawhide.