DF is not very reliable in my opinion.
What is the result with 'DU'
Hi,
I have got a strange symptom at one of my friend's WL-HDD with an internal 40 GB harddisc.
When I use fdisk, everything seems to be fine:
Disk /dev/discs/disc0/disc: 40.0 GB, 40007761920 bytes
255 heads, 63 sectors/track, 4864 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/discs/disc0/part1 1 63 506016 82 Linux swap
/dev/discs/disc0/part2 64 4864 38564032+ 83 Linux
The disc is detected as a 40 GB harddisc and the second partition is allocated almost 100% of disc size.
However, a df -h gives the following output:
[root@WL-HDD root]$ df -h
Filesystem Size Used Avail Use% Mounted on
rootfs 3.0M 3.0M 0 100% /
/dev/root 3.0M 3.0M 0 100% /
/dev/discs/disc0/part2 966M 917M 0 100% /opt
Total disc space is now down to less than 1 GB! And this is also affecting any new files that are being created, so basically, the disc is full with just one out of 40 GB being used.
Can anybody help me explaining why this is the case?
As it's my friend's WL-HDD, I don't know when this occured.
Any help is highly appreciated, if you need more information, don't hesitate to ask.
CU!
F.
DF is not very reliable in my opinion.
What is the result with 'DU'
du -h gives around 890 MB which is what df -h also reports.
In dmesg I see a log entry that states the correct size of the disc (40 GB) with two partitions (swap and data).
It is also strange that when I enter "mount", it says "cannot find /etc/mtab" and an "umount /opt" gives "device or resource busy" although I've killed almost all processes manually.
Anybody another idea?
I figured out that the problem with the missing mtab does not occur if I do not use /etc/fstab for mounting. If I comment out all mount points, the non-swap partition (ext3) is still mounted, but to /tmp/mnt/disc0_2/ and "mount" displays all mountpoints.
This way, I can also unmount /tmp/mnt/disc0_2 and do an e2fsck, which runs fine, although the 251015 blocks seem to be closer to 1GB rather than 40 GB?
[root@WL-HDD root]$ e2fsck /dev/discs/disc0/part2
e2fsck 1.38 (30-Jun-2005)
/dev/discs/disc0/part2: clean, 9329/125696 files, 222840/251015 block
fdisk gives the following output:
Disk /dev/discs/disc0/disc: 40.0 GB, 40007761920 bytes
255 heads, 63 sectors/track, 4864 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/discs/disc0/part1 1 63 506016 82 Linux swap
/dev/discs/disc0/part2 64 4864 38564032+ 83 Linux
Any ideas?
I finally gave up. I realized that maybe resize2fs would have done the trick after a test run with mke2fs -n showed me a much larger number of blocks that would be created if the partition was reformatted. resize2fs works only on unmounted partitions under kernels smaller than 2.6, and copying it to flash did not help as some other programs were probably needed.
I tried some other backup superblocks, but only those within the (wrong, smaller) block range were accepted.
So I had to do a backup as the drive is on a remote location and then reinstall the backup. Time-consuming, but in the end it worked out.
Merry Christmas,
F.