should not try to fsck ntfs volumes on boot

Status changed to 'Confirmed' because the bug affects multiple users.

Is this doable for oneiric?

Anything is happening with this bug or not?
Error is still there.

> So, ntfs-3g should symlink fsck.ntfs to ntfsfix or ntfsck,
> though the latter fails to check my partitions,

It fails because the development of ntfsck has been interrupted years ago before anything useful was developed.

> so maybe ntfsfix would be better.

ntfsfix does not check the volume, it fixes common errors preventing from mounting, and triggers a check by Windows. If you can mount, you do not need ntfsfix.

If you want to call fsck.ntfs, do so with an option avoiding the checks, until some kind soul strives to develop them.

Yes, that's fine. But what should I do when I start my Oneiric and it bothers me with fsck.ntfs for my two windows partitions? It's not a solution to say
'If you want to call fsck.ntfs, do so with an option avoiding the checks, until some kind soul strives to develop them.'
Haha.

With Natty I never had such trouble before. When it is not possible to check a partition, Ubuntu shouldn't come with such stupid error messages. It should ignore such things it can't handle. I don't want that the boot process gets interrupted and wants me to press S to skip these check. I have to do this 2 times on every boot up. This can't be true. I never had it before and I don't want it for the future. I need a really solution. Such things shouldn't happen. What about newbies which a trying Ubuntu the first time?

> But what should I do when I start my Oneiric and it bothers me with
> fsck.ntfs for my two windows partitions?

As there is no possible check, you should not request a check. Checks at boot time are determined by the last field in /etc/fstab. <quote from man fstab> : If the sixth field is not present or zero, a value of zero is returned and fsck will assume that the filesystem does not need to be checked </quote>

> I need a really solution.

There are two possibilities : either you avoid the check, or you find someone volunteering to develop a checker for free.

1) This problem will affect all users that have ntfs checks requested in their fstab after upgrading to oneiric. These are supposedly *very many*.

2) The real problem is not that the check fails, but that oneiric's startup screen FALSELY reports that there were "serious problems" mounting the drives. The screen should report that the drives could not be checked since no program to do this was available. The message as it is will badly scare users, since the must assume that their data is damaged, which, most likely, is not the case.

3) Not requesting the check will of course avoid this bug. But since users could safely request checking in fstab up to natty without receiving any errors, we must assume that checks *are* requested for ntfs drives.

4) This is a show stopper for oneiric.

Users should not have to intervene on every boot. Please, at least fix that much.

 I agree that this is a serious problem that might affect a lot of users.
And needs to be fixed soon.

This bug is serious. Why does nothing happens to fix it?

You may like to support bug 920545 "Upgrade NTFS-3G to new version [precise]

One of a few interesting bugfixes is:

 * ntfsfix: new option -d to clear the dirty flag if fix is successful

    And some more with regards to ntfsfix:

    * ntfsfix: fixed volume dirty flag test
    * ntfsfix: new option to clear the list of bad sectors
    * ntfsfix: fixed compilation on Sparc
    * ntfsfix: fixed a bug causing a segmentation violation
    * ntfsfix: repair self-located MFT data bug

pls see also comment #3 in duplicate bug https://bugs.launchpad.net/ubuntu/+source/ntfs-3g/+bug/875237/comments/3

/etc/fstab has to be set up correctly too, if you prefer automounting

The right answer here, since ntfsck doesn't do anything useful, is to ensure we're not trying to check ntfs volumes at boot.

Could someone who's seeing this bug attach their /etc/fstab?

IMO the proper attitude would be to create a dummy fsck.ntfs which reports that "no checks has been done" (where?) and returns OK

There are a few ways in which we might choose to fix that; that wouldn't
be my preferred one, but this is a colour-of-the-bikeshed argument.
Rather than getting into solutions, can you answer Steve's question?

sorry, i slipped. here's mine producing error. normally i just keep it at 0

# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc nodev,noexec,nosuid 0 0
# / was on /dev/sda7 during installation
UUID=3e46fa3b-b898-4f19-b1ec-873c3ef66d53 / ext4 errors=remount-ro 0 1
# swap was on /dev/sda6 during installation
UUID=610dc6e3-7481-4879-a2f8-bb5ab265551e none swap sw 0 0

/dev/sda5 /media/commune ntfs-3g errors=remount-ro 0 2

Thanks. I guess that was added manually, then; I'll work around this in mountall.

This bug was fixed in the package mountall - 2.36

---------------
mountall (2.36) precise; urgency=low

  * If no specific fsck.* checker is found for a given filesystem type
    (except for a short list where we know that checkers really should be
    present), skip checking that filesystem rather than emitting scary error
    messages (LP: #838091). This duplicates some logic from 'fsck -A', but
    that seems reasonable given mountall's purpose.
 -- Colin Watson <email address hidden> Thu, 12 Apr 2012 18:58:16 +0100

not going to happen for oneiric, closing