From Bugzilla Helper:
User-Agent: Mozilla/5.0 Galeon/1.2.6 (X11; Linux i686; U;) Gecko/20020830

Description of problem:
(First of all, see Bug 79558 for the background on this one.)

The following uses two DiskOnKeys, but I think that it will happen with other
devices (say, USB floppies).

Step 1) I boot the machine with a DiskOnKey (which I'll call here DiskOnKey 1)
inserted in a USB port. It gets assigned to sda, and pdfstab adds the following
line to /etc/fstab:

/dev/sda                /mnt/diskonkey          auto    noauto,owner,kudzu 0 0

Step 2) I remove DiskOnKey 1.

Step 3) I insert a second DiskOnKey (which I'll call here DiskOnKey 2). It gets
assigned to sdb, and at this point /etc/fstab becomes:

/dev/sdb                /mnt/diskonkey          auto    noauto,owner,kudzu 0 0

Step 4) I remove DiskOnKey 2. No changes in /dev/fstab.

Step 5) I re-insert DiskOnKey 1. Now /dev/fstab has the following two "kudzu"
llines:

/dev/sdb                /mnt/diskonkey          auto    noauto,owner,kudzu 0 0
/dev/sda                /mnt/diskonkey1         auto    noauto,owner,kudzu 0 0

This behavior seems inconsistent, since:
- the mount point for the devices was exchanged (confuses the user).
- Inserting a different device erased the previous entry in step 3 but not in
step 5.

By the way, this behavior is not always reproducible exactly like that, but
always results in a different error. For instance, if in Step 1 I plug the
DiskOnKey *after* I boot:

After Step 3)
/dev/sda                /mnt/diskonkey          auto    noauto,owner,kudzu 0 0
/dev/sdb                /mnt/diskonkey1         auto    noauto,owner,kudzu 0 0
After Step 4)
/dev/sdb                /mnt/diskonkey1         auto    noauto,owner,kudzu 0 0
After Step 5)
/dev/sdb                /mnt/diskonkey1         auto    noauto,owner,kudzu 0 0

So I ended with only /dev/sda plugged and only the /dev/sdb line in /etc/fstab.

This might have something to do with Bug 77382, or not (or it is an easier test
to find that bug).

Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
See the steps in the description.
    

Actual Results:  Kudzu exchanges the devices and the mountpoints, or ends up
with only a mountpoint for the wrong device.

Expected Results:  The assignment of the mountpoints and devices should remain
constant (or at least correct :-)

Additional info: