On 2010-09-29 03:36, tsu2 wrote:
>
> Carlos E. R.;2229976 Wrote:
> OK, I’ll bite.
> Is this something based on personal observation only, or do you know
> that this causes a setting (eg some kind of “dirty disconnect”) that is
> preserved and even more surprisingly somehow communicated to a different
> machine if the USB storage device is re-attached to another machine?
Documented. In NTFS partitions or sticks (which is the OP case), there is a “journal”, same as do
reiserfs, xfs, ext3/4… When the filesystem is properly closed, the journal is applied and reset.
When you plug that stick into linux, it detects that the log is not empty, and not knowing how to
apply the log, it refuses to mount it. This is explained in the ntfs-3g documentation.
And the problem of the OP was precisely that his external NTFS disk was not mounted when plugged in,
it did not appear automatically - which matches precisely with the above “feature”.
Notice:
The post I replied to was one from BlueAce3:
Re: Mounting a NTFS USB Drive w/ family photos
group: opensuse.org.help.hardware.64-bit
to which you have replied in “opensuse.org.help.hardware” instead. It is true that BlueAce3 also
posted here, but I did not see it, nor replied to.
We are having a strange crosspost.
And the post from “cameron” you mention below I have no idea which one it is. :-?
> Again, the only time I’ve seen consequences is if a file operation is
> in progress, which shouldn’t be a surprise.
>
> In cameron’s situation, I suspect a file operation in progress (ie.
> maybe indexing?). Once file indexing has been completed I wouldn’t be
> surprised if he could then remove/re-attach the drive normally.
I can not comment on “cameron” problem as I don’t have that post.
In general, FAT disks are quite resilient to the “removal” treatment. After all, it was designed for
floppies, which could be easily removed after waiting for the light to go off. But… modern disks
are much bigger, and the operating systems can cache and delay write operations for undefined
periods of time. There is a variable risk with sudden removal of a disk, depending a lot on the
applications running. The problems might not manifest themselves right after the removal, they can
be delayed days or months.
It is true, man people think that nothing bad happens… till one day, it happens. Yes, I have seen
many people with destroyed data on sticks.
And in the case of journaled filesystems, automounting usually fails on the very next plug-in. As
the OP saw.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)