Rootfs twice

techux · 07-10-2013, 11:38 AM

Hi guys.

I have a server at running debian 6.0.7, after a while I noticed when I run df -h / is twice as you can see below. After the installation it didnt show any UUID nor rootfs

Is this critical or there isnt nothing to worry about,

Code:

rootfs 4.6G 1.3G 3.1G 30% /
udev 10M 128K 9.9M 2% /dev
tmpfs 101M 84K 101M 1% /run
/dev/disk/by-uuid/7ce388cb-e37a-490f-8de9-e2d78c09345454
4.6G 1.3G 3.1G 30% /
tmpfs 5.0M 0 5.0M 0% /run/lock
tmpfs 393M 0 393M 0% /run/shm

PS: not sure if this post should be here or at debian section.sorry for the double post

eSelix · 07-10-2013, 12:10 PM

Check your "/etc/fstab". You probably mounted it twice. It should not be a problem.

techux · 07-10-2013, 02:03 PM

No, I havent changed or touch /etc/fstab.

eSelix · 07-10-2013, 03:51 PM

Can you show files '/etc/mtab' and '/proc/mounts'?

techux · 07-10-2013, 04:15 PM

/etc/mtab is a symlink to /proc/mounts.

I think it could be a debian bug due the symlink.

I really wont care so much about the UUID if this is not big deal and wont cause any problem in the system, is it?

Here is /proc/mounts

Quote:

rootfs / rootfs rw 0 0
sysfs /sys sysfs rw,nosuid,nodev,noexec,relatime 0 0
proc /proc proc rw,nosuid,nodev,noexec,relatime 0 0
udev /dev devtmpfs rw,relatime,size=10240k,nr_inodes=128098,mode=755 0 0
devpts /dev/pts devpts rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000 0 0
tmpfs /run tmpfs rw,nosuid,noexec,relatime,size=103420k,mode=755 0 0
/dev/disk/by-uuid/7ce388cb-e37a-490f-8de9-e2d78c09fda7 / ext3 rw,relatime,errors=remount-ro,data=ordered 0 0
tmpfs /run/lock tmpfs rw,nosuid,nodev,noexec,relatime,size=5120k 0 0
tmpfs /run/shm tmpfs rw,nosuid,nodev,noexec,relatime,size=401980k 0 0
rpc_pipefs /var/lib/nfs/rpc_pipefs rpc_pipefs rw,relatime 0 0

Quote:

fstab:

# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc defaults 0 0
# / was on /dev/md0 during installation
UUID=7ce388cb-e37a-490f-8de9-e2d78c09fda7 / ext3 errors=remount-ro 0 1
# swap was on /dev/md1 during installation
UUID=ba9b5c46-6a30-4264-aa35-22a7ce36f28c none swap sw 0 0
/dev/scd0 /media/cdrom0 udf,iso9660 user,noauto 0 0
/dev/scd1 /media/cdrom1 udf,iso9660 user,noauto 0 0
/dev/fd0 /media/floppy0 auto rw,user,noauto 0 0

Firerat · 07-10-2013, 05:05 PM

/etc/mtab is probably a symlink to /proc/mounts

And this is the problem..

Here is some background I came across

https://bugzilla.redhat.com/show_bug.cgi?id=709351
( opened 2011-05 )

coreutils (8.20-3), in sid does not resolve the issue.

In your case, you can resolve the issue by excluding file system type 'rootfs'

Code:

df -h -x rootfs

In the case of bind mounts, well.. some script to clean up the duplicates, far from ideal.

techux · 07-10-2013, 05:20 PM

I added to /etc/profile those lines to hide the rootfs but it is only for rootfs. Still display the UUID for /

Quote:

Originally Posted by Firerat

/etc/mtab is probably a symlink to /proc/mounts

And this is the problem..

Here is some background I came across

https://bugzilla.redhat.com/show_bug.cgi?id=709351
( opened 2011-05 )

coreutils (8.20-3), in sid does not resolve the issue.

In your case, you can resolve the issue by excluding file system type 'rootfs'

Code:

df -h -x rootfs

In the case of bind mounts, well.. some script to clean up the duplicates, far from ideal.

Firerat · 07-10-2013, 05:56 PM

Quote:

Originally Posted by techux

I added to /etc/profile those lines to hide the rootfs but it is only for rootfs. Still display the UUID for /

just done a little reading around the subject ( my Debian Jessie does not have the UUID issue )

It seems that the initramfs *might* be using busybox to mount the root is causing the UUID to be used in proc/mounts

I have 'emulated' this by using busybox to mount via UUID, and indeed I have a mixed df output now.

As you are otherwise booting fine and this is a cosmetic issue, I would be reluctant to fix the initramfs ( it isn't very broken ).
Instead I would look at using the output of either ; /sbin/blkid or ls -l /dev/disk/by-uuid/ to substitute the UUID with the block device using a script
Guess it depends on how important an issue it is.

For reference here is what pointed me to busybox
http://lwn.net/Articles/549363/
It might turn out not to be the culprit, I didn't do too much work here.

I guess another option is to not use UUID in /etc/fstab or grub.. depends if you are in the habit of adding/moving drives around ( physically )

Firerat · 07-10-2013, 07:40 PM

as an exercise for myself I came up with this

Code:

df -h -x rootfs $@ \
| awk '{if ($1 ~ /by-uuid/) "ls -l "$1"|sed -e s[.*./[/dev/["|& getline $1;printf "%-11s%9s %9s %9s %4s %-10s%s\n",$1,$2,$3,$4,$5,$6,$7}'

The formatting is a little screwy if you pass things like -T ( the extra field messes everything up )

Not that the formatting is great to begin with ( I need to learn how to manipulate $1's length without causing the rest to 're-evaluate' )

But hey Botox is Botox