Hi,

On Mon, Jan 16, 2006 at 06:28:07PM +0100, Mike Dornberger wrote:
> did you "reformat" the new partition(s) hdb5 (and hdb6) with something like
> (as root):
> 
> # mkfs -t ext3 /dev/hdb5
> (or # mke2fs -j /dev/hdb5)

Yes, I did. I think that the problem was that I didn't
reboot the system after creating the partitions. I did:

  # mkfs.ext3 -cv /dev/hdb5

immediately after creating the partitions. Only after
writing the bug report I found in the manual that it's
recommended to reboot the system.

[...]

> You have probably lost all your data at your former hdb5. (The fs might
> "see" the filenames but the content is probably erroneous and/or
> inaccessible.)

[...]

Thanks for the explanation! I think that what happened was
what you described: there were only file names, not files at
hdb5. But no problem: I had a backup in another partition
(hdb7). So, I didn't lost any data.

I used parted to delete the two new partitions and create
them again. Then I restored the data from the other
partition. This time I rebooted the system about two or
three times between operations. The 32 bit programs that are
in this partitions are running normally and the result of
"df" command is meaningful:

  # df
  Filesystem  1K-blocks      Used Available Use% Mounted on
  /dev/hdb1     6728280   3403348   2983152  54% /
  /dev/hdb3     5763648   3972324   1498540  73% /home
  /dev/hdb8     3842348   1072988   2574176  30% /mnt/ia32
  /home         5763648   3972324   1498540  73% /mnt/ia32/home
  /tmp          6728280   3403348   2983152  54% /mnt/ia32/tmp
  tmpfs          511380         0    511380   0% /dev/shm

However, the partition got a different number (it's now
hdb8) and cfdisk reports a problem:

  # cfdisk /dev/hdb

  FATAL ERROR: Bad logical partition 8: enlarged logical
  partitions overlap
            Press any key to exit cfdisk

There is no error, according to parted:

root@:~# parted /dev/hdb print
Disk geometry for /dev/hdb: 0kB - 80GB
Disk label type: msdos
Number  Start   End     Size    Type      File system  Flags
1       32kB    7000MB  7000MB  primary   ext3         boot
2       7000MB  8003MB  1003MB  primary   linux-swap   
3       8003MB  14GB    5996MB  primary   ext3         
4       14GB    80GB    66GB    extended               
8       14GB    18GB    3997MB  logical   ext3         
9       18GB    62GB    44GB    logical   ext3         
5       62GB    68GB    5996MB  logical   ext3         
6       68GB    74GB    5996MB  logical   ext3         
7       74GB    80GB    6029MB  logical   ext3         
Information: Don't forget to update /etc/fstab, if necessary.

On the one hand, probably the problem was caused by a wrong
usage of cfdisk.  Then, this bug could be changed to
wishlist: "More clear warning that the system must be
rebooted". Something like a blank screen with only the
message: "ERROR: The kernel failed to read the new partition
table.  You must reboot your system!". This might not be
necessary for experienced Unix users, but Linux is receiving
many people migrating from other OS.

On the other hand, it's strange that parted and cfdisk
disagree about how healthy are the partitions.

Thank you for your reply,

Best regards,

Jakson

Attachment: signature.asc
Description: Digital signature

Reply via email to