On Mon, 23 Jan 2006, Mikkel L. Ellertson wrote: > >>Now, you can not directly tell the system to ignore the file systems > >>on the new drive without disabling the drive all together. > > > > Disagree. I made no new entries in /etc/fstab to mount the second, > > newly-installed drive's partitions. Therefore they should have been > > ignored. So it seems that there are only two possibilites: (1) The > > correct drive was booted (channel A/ide0 master) but the wrong > > partitions were found at mount time; or (2) The incorrect drive was > > booted (channel B/ide1 master) and, again, the wrong partitions were > > selected for mounting. > > > Look at /etc/fstab - are the partitions specified as > /dev/<something>, or are you using partition labels? If you use > partition labels, and you have more then one partition with the > same label, you get unpredictable results. Actually you get very predictable results judging from the FC4 source: You'll end-up mounting the last entry in /proc/partitions whose label matches the label in /etc/fstab. Since /proc/partitions is (apparently) ordered alphanumerically by device filename, I'll always get the ide1-attached drive which has that same label. All else remaining equal, I should also see log messages about the duplicate labels, but the old filesystems labelled /usr and /var were incomplete (recall FC2 and FC3 install problems for certain IDE CD-ROM drives; this older disk drive was last used during a failed install), so I couldn't search the log files. Of course I'm running FC3 not FC4, so the details may differ a little. But all's well that ends well. I just wanted to mount some old Prolog and Python source.... > Mikkel > * Nick Geovanis | IT Computing Svcs | Northwestern Univ | n-geovanis@ | northwestern.edu +------------------->
