Feisty Ubuntu: The UUID problem
Jeremy Boden
correct.
In Edgy a change was made to use a UUID instead of constructs like /
dev/sda1
History:-
In Dapper I had LVM using /dev/mapper - purely because I wanted to
see what it could do.
Upgrade to Edgy (& these UUID's) - seemed OK
After upgrade to Feisty, I managed to cause occasional breakages in /
dev/mapper
I have, I think successfully, removed /dev/mapper
However, if I force a restart - with fsck checking I get messages
like:
/dev/sdb1: 41/526320 files (22.0% non-contiguous), 153744/1052226
blocks
Failed to open the device 'UUID=7b3d8e1c-2080-4016-a672-1b2fed6ce7c8':
No such file or directory
Eventually fsck "dies" and I get a single user CTRL-D option
If I take this and reboot (no fsck checks) the PC boots normally.
I run commands like:-
jeremy@HECTOR:~$ udevinfo -q env -n /dev/sdb2
ID_VENDOR=ATA
ID_MODEL=SAMSUNG_SP2504C
ID_REVISION=VT10
ID_SERIAL=1ATA_SAMSUNG_SP2504C_S09QJ1HL916745
ID_SERIAL_SHORT=ATA_SAMSUNG_SP2504C_S09QJ1HL916745
ID_TYPE=disk
ID_BUS=scsi
ID_PATH=pci-0000:00:07.0-scsi-1:0:0:0
ID_FS_USAGE=filesystem
ID_FS_TYPE=ext3
ID_FS_VERSION=1.0
ID_FS_UUID=064f1127-b29f-4146-8ffd-181c4ac9f1ee
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
But that UUID is just a copy of what I have in my fstab
Yet another reboot with fsck would only fail.
Should I
a) revert all those UUID's to /dev/sda1 style definitions OR
b) Is there a definitive way of finding out the UUID of a partition/
filesystem which doesn't just involve a lookup of fstab.
BTW I have 2 disks - one SATA & one SCSI
Thanks in advance...
Robert Citek
> Should I
> a) revert all those UUID's to /dev/sda1 style definitions OR
No. Once you get used to UUIDs or Labels, I think you'll like them.
It allows you (instead of the kernel) to control the naming of the
partitions. This is extremely useful if you move drives around or use
external drives.
> b) Is there a definitive way of finding out the UUID of a partition/
> filesystem which doesn't just involve a lookup of fstab.
Yes, use tune2fs, assuming the fs is ext2/3. For example:
tune2fs -l /dev/sda2 | grep -i uuid
To change it, use tune2fs -U ${new_UUID} /dev/sda2
To generate a random UUID, use uuidgen.
> BTW I have 2 disks - one SATA & one SCSI
That is the perfect reason to love UUIDs and Labels. With UUIDs and
Labels you can add, remove, or change the position of those drives and
not have to worry about redoing all the system settings, e.g.
/etc/fstab.
Personally, I like Labels better than UUIDs. To discover the label of
a partition:
tune2fs -l /dev/sda2 | grep -i volume
To change the label to foobar:
e2label /dev/sda2 foobar
To use the label in /etc/fstab, replace /dev/sda2 or UUID= with LABEL=.
To use the label in /boot/grub/menu.lst, replace root=UUID= with
root=LABEL=. Do that everywhere, including the # kopt= line.
Good luck and let us know how things go.
Regards,
- Robert
Jeremy Boden
On 16 Oct, 19:45, "Robert Citek" <robert.ci...@gmail.com> wrote:
Thanks Robert,
So far so good.
Yes - I will go the UUID route - although I'm not in the habit of
swapping hard disks...
BTW tune2fs requires sudo.
Partial success - I've been able to confirm that all ext3 partitions
are correct.
All partitions in GRUB & / are ext3 (already using UUID's) and working
- so that's OK!
However, /home & a mount created by me (/filestore) are reiserfs.
I've looked around for something like tune2fs for reiser and all I can
find is reiserfstune, which only allows setting parameters, not
interrogating them.
I appreciate I could generate a UUID & then set the filesystem - but
that seems to be living dangerously!
That /dev/mapper left comments in fstab which interchanged sda? with
sdb? - very confusing.
Can you help with reiserfs?
Thanks & Regards,
Robert Citek
> Thanks Robert,
>
> So far so good.
That's great to hear.
> Yes - I will go the UUID route - although I'm not in the habit of
> swapping hard disks...
> BTW tune2fs requires sudo.
Yes, it does. Good catch.
> I've looked around for something like tune2fs for reiser and all I can
> find is reiserfstune, which only allows setting parameters, not
> interrogating them.
It does both. Unfortunately, it doesn't work if the filesystem is
mounted. From the bit of testing I did, typing 'reiserfstune /dev/sda2'
seems to interrogate the fs without modifying it. But perhaps someone
more familiar with reiserfs can comment.
> Can you help with reiserfs?
Looks like you'll have to boot into recovery mode, probably unmount the
reiser filesystems, interrogate them, and then remount them before
continuing the normal boot process.
Again, good luck and let us know how it goes.
Regards,
- Robert
Jeremy Boden
Sorry to be a pain!
You are right - I booted using the recovery option.
"Luckily" I had preset my root password so I can sign-on as root.
If I enter reiserfstune /dev/sdaX then I get a full description
(including UUID).
However, booting into recovery mode seems to swap sda & sdb.
Here is a list of all my file systems (Partial /dev contents)
sda1 /boot (ext3)
sda2 /backup (ext3)
sda3 /filestore (reiserfs)
sdb2 /swap
sdb5 / (ext3)
sdb6 /usr (ext3)
sdb7 /var (ext3)
sdb8 /home (reiserfs)
There are no other /dev/sd* entries.
[sdb has a small unused partion, but no file system in what would be
sdb1]
There are no other entries in /dev
But when I boot into recovery mode
I see only sda8 & sdb3 as reiserfs (AND the UUID's correspond with the
"real" sda3 & sdb8).
Since pressing CTRL-D on a normal boot allows /filestore & /home to be
mounted correctly I'm reluctant to change fstab.
Is it possible that this is a non-UUID problem?
Robert Citek
> If I enter reiserfstune /dev/sdaX then I get a full description
> (including UUID).
>
> However, booting into recovery mode seems to swap sda & sdb.
> Here is a list of all my file systems (Partial /dev contents)
> sda1 /boot (ext3)
> sda2 /backup (ext3)
> sda3 /filestore (reiserfs)
> sdb2 /swap
> sdb5 / (ext3)
> sdb6 /usr (ext3)
> sdb7 /var (ext3)
> sdb8 /home (reiserfs)
> There are no other /dev/sd* entries.
How do you know the above information? I suspect something is doing
some sleight-of-hand. To figure out what's going on, boot normally
and type (cut-n-paste) the following commands into a terminal:
true ; (
sudo echo -n
set -x
cat /etc/issue.net
uname -va
sudo fdisk -l
cat /etc/fstab
awk 'NR >1 && $2 > 0 {print "/dev/" $4}' /proc/partitions |
xargs -n 1 sudo file -s
) >& output.txt
Please post the contents of output.txt to this list.
> Is it possible that this is a non-UUID problem?
For now, anything is possible.
Regards,
- Robert
Jeremy Boden
On Oct 17, 2:08 pm, "Robert Citek" <robert.ci...@gmail.com> wrote:
I ran those commands - please don't be misled by comment lines in
fstab - I typed/amended them myself.
BTW The partition information is from my recollection + what "System
Monitor" GUI says.
=======================START OF
OUTPUT=================================
+ cat /etc/issue.net
Ubuntu 7.04
+ uname -va
Linux HECTOR 2.6.20-16-generic #2 SMP Sun Sep 23 18:31:23 UTC 2007
x86_64 GNU/Linux
+ sudo fdisk -l
Disk /dev/sda: 250.0 GB, 250059350016 bytes
255 heads, 63 sectors/track, 30401 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/sda1 1 131 1052226 83 Linux
/dev/sda2 132 6505 51199155 83 Linux
/dev/sda3 6506 30401 191944620 83 Linux
Disk /dev/sdb: 73.4 GB, 73407865856 bytes
255 heads, 63 sectors/track, 8924 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/sdb2 132 164 265072+ 82 Linux swap /
Solaris
/dev/sdb3 165 8924 70364700 5 Extended
/dev/sdb5 165 674 4096543+ 83 Linux
/dev/sdb6 675 2203 12281661 83 Linux
/dev/sdb7 2204 2713 4096543+ 83 Linux
/dev/sdb8 2714 8924 49889826 83 Linux
+ cat /etc/fstab
# /etc/fstab: static file system information.
#
# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc defaults 0 0
# /dev/sda5 -- converted during upgrade to edgy
UUID=233565b8-f1c9-412d-add7-848988ee52d3 / ext3
defaults,errors=remount-ro 0 1 # OK
# /dev/sdb2 -- converted during upgrade to edgy
UUID=064f1127-b29f-4146-8ffd-181c4ac9f1ee /backup ext3 defaults 0 2
# OK
# /dev/sdb3 -- converted during upgrade to edgy
UUID=feb26ef0-be8e-49ac-8fbe-2c443549c553 /filestore reiserfs defaults
0 2 # ?
# /dev/sdb1 -- converted during upgrade to edgy
UUID=403d16ca-ce7b-44f4-bcb3-87654deb5466 /boot ext3 defaults 0 2 #
OK
# /dev/sda8 -- converted during upgrade to edgy
UUID=7b3d8e1c-2080-4016-a672-1b2fed6ce7c8 /home reiserfs defaults 0
2 # ?
# /dev/sda6 -- converted during upgrade to edgy
UUID=247b2017-4636-40c5-8c7d-82721dd0caf5 /usr ext3 defaults 0 2 #
OK
# /dev/sda7 -- converted during upgrade to edgy
UUID=2ff24943-33a8-43e9-97da-73982cf8845b /var ext3 defaults 0 2 #
OK
# /dev/sda2 -- converted during upgrade to edgy
UUID=936ee5a4-2f4f-48c6-ba36-894426deb357 none swap sw 0 0 # ?
### /dev/sda1 /media/disk ext3 rw,noexec,nosuid,nodev,noauto 0 0
/dev/cdrom /media/cdrom0 udf,iso9660 user,noauto 0 0
+ awk 'NR >1 && $2 > 0 {print "/dev/" $4}' /proc/partitions
+ xargs -n 1 sudo file -s
/dev/sda1: Linux rev 1.0 ext3 filesystem data (needs journal recovery)
/dev/sda2: Linux rev 1.0 ext3 filesystem data (needs journal recovery)
(large files)
/dev/sda3: ReiserFS V3.6 block size 4096 (mounted or unclean) num
blocks 47986144 r5 hash
/dev/sdb: x86 boot sector; partition 2: ID=0x82, starthead 0,
startsector 2104515, 530145 sectors; partition 3: ID=0x5, starthead 0,
startsector 2634660, 140729400 sectors
/dev/sdb2: Linux/i386 swap file (new style) 1 (4K pages) size 66267
pages
/dev/sdb5: Linux rev 1.0 ext3 filesystem data (needs journal recovery)
/dev/sdb6: Linux rev 1.0 ext3 filesystem data (needs journal recovery)
/dev/sdb7: Linux rev 1.0 ext3 filesystem data (needs journal recovery)
/dev/sdb8: ReiserFS V3.6 block size 4096 (mounted or unclean) num
blocks 12472448 r5 hash
===========END OF OUTPUT=====================
Can I point out that the 250G disk is SATA and apart from /boot is a
non-system disk.
The SCSI disk is 73G & should be plain Linux partitions.
Thanks & regards
Jeremy
Robert Citek
Those two partitions are reiserfs. Boot into recovery mode, find the
UUID, and add them to the /etc/fstab in the appropriate place.
BTW, I think you swapped sda and sdb in your notes in the /etc/fstab.
Regards,
- Robert
Jeremy Boden
Thanks Robert,
However - this is where things get really bad!
I sorted out the comments in fstab.
When I checked the UUID's (in recovery mode) for the two ReiserFS's
they matched the two values of UUID's in fstab - although I was
confused about the disk names, I did write the two UUID's down.
However, I didn't swap the two UUID's over - because the correct
partitions are always being loaded for /home & /filestore
I have done several sudo shutdown -rF now commands to reboot with
forced fsck's:-
When fully started from recovery mode I have run "mount" - #1
When this restarted in normal mode I have run "mount" - #2
Rebooted into normal mode - I get errors (UUID's not found)
I still ran "mount" - #3
====BUT all the mount outputs - 1, 2 & 3 are identical!!!!
I will include the contents of /var/log/fsck/checkfs as it gets wiped
on every boot:-
It starts here:--------------
Log of fsck -C -R -A -a -f
Thu Oct 18 12:28:43 2007
fsck 1.40-WIP (14-Nov-2006)
/dev/sda2: 471/6406144 files (9.6% non-contiguous), 6575990/12799788
blocks
/dev/sda1: 41/526320 files (22.0% non-contiguous), 153744/1052226
blocks
/dev/sdb6: 189331/1537088 files (2.1% non-contiguous), 1444187/3070415
blocks
Failed to open the device 'UUID=feb26ef0-be8e-49ac-8fbe-2c443549c553':
No such file or directory
Failed to open the device 'UUID=7b3d8e1c-2080-4016-a672-1b2fed6ce7c8':
No such file or directory
/dev/sdb7: 12877/513024 files (6.8% non-contiguous), 239317/1024135
blocks
fsck died with exit status 8
Thu Oct 18 12:30:23 2007
----------------
(end of the log file)
My problem is that those two UUID's are definitely the ReiserFS
partitions and that even though it says they don't exist, I was able
to locate them using reiserfstune and Feisty was able to locate them
to mount them correctly.
Would it help to admit defeat and call them /dev/sdb8 & /dev/sda3 in
fstab...
I'd like to know why fsck says it has died...
Thanks for your help,
- Jeremy
Robert Citek
> Would it help to admit defeat and call them /dev/sdb8 & /dev/sda3 in
> fstab...
Quite possibly. Or change the ending "2" to a "0" on those lines. It
looks like you have tripped upon a bug similar to these:
https://bugs.launchpad.net/ubuntu/+source/e2fsprogs/+bug/66032
https://bugs.launchpad.net/ubuntu/+source/e2fsprogs/+bug/110138
> I'd like to know why fsck says it has died...
I suspect you have a buggy blkid or vol_id. To confirm, try this bit of code:
true ; (
sudo echo -n
set -x
blkid -v
for i in $(awk 'NR >1 && $2 > 0 {print "/dev/" $4}' /proc/partitions) ; do
echo $i
sudo blkid $i
sudo vol_id $i
done
) >& output.txt
Please post the contents of output.txt
Regards,
- Robert
Jeremy Boden
> On 10/18/07, Jeremy Boden <JeremyBo...@gmail.com> wrote:
>
> > Would it help to admit defeat and call them /dev/sdb8 & /dev/sda3 in
> > fstab...
>
> Quite possibly. Or change the ending "2" to a "0" on those lines. It
> looks like you have tripped upon a bug similar to these:
>
>
> > I'd like to know why fsck says it has died...
>
> I suspect you have a buggy blkid or vol_id. To confirm, try this bit of code:
>
> true ; (
> sudo echo -n
> set -x
> blkid -v
> for i in $(awk 'NR >1 && $2 > 0 {print "/dev/" $4}' /proc/partitions) ; do
> echo $i
> sudo blkid $i
> sudo vol_id $i
> done
> ) >& output.txt
>
> Please post the contents of output.txt
>
> Regards,
> - Robert
Thanks, once again Robert!
Here is the output - rather voluminous I'm afraid:-
+ blkid -v
blkid 1.0.0 (12-Feb-2003)
++ awk 'NR >1 && $2 > 0 {print "/dev/" $4}' /proc/partitions
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sda2
/dev/sda2
+ sudo blkid /dev/sda2
/dev/sda2: TYPE="swap" UUID="936ee5a4-2f4f-48c6-ba36-894426deb357"
+ sudo vol_id /dev/sda2
ID_FS_USAGE=other
ID_FS_TYPE=swap
ID_FS_VERSION=2
ID_FS_UUID=936ee5a4-2f4f-48c6-ba36-894426deb357
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sda3
/dev/sda3
+ sudo blkid /dev/sda3
+ sudo vol_id /dev/sda3
/dev/sda3: unknown volume type
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sda5
/dev/sda5
+ sudo blkid /dev/sda5
/dev/sda5: UUID="233565b8-f1c9-412d-add7-848988ee52d3" SEC_TYPE="ext2"
TYPE="ext3"
+ sudo vol_id /dev/sda5
ID_FS_USAGE=filesystem
ID_FS_TYPE=ext3
ID_FS_VERSION=1.0
ID_FS_UUID=233565b8-f1c9-412d-add7-848988ee52d3
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sda6
/dev/sda6
+ sudo blkid /dev/sda6
/dev/sda6: UUID="247b2017-4636-40c5-8c7d-82721dd0caf5" SEC_TYPE="ext2"
TYPE="ext3"
+ sudo vol_id /dev/sda6
ID_FS_USAGE=filesystem
ID_FS_TYPE=ext3
ID_FS_VERSION=1.0
ID_FS_UUID=247b2017-4636-40c5-8c7d-82721dd0caf5
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sda7
/dev/sda7
+ sudo blkid /dev/sda7
/dev/sda7: UUID="2ff24943-33a8-43e9-97da-73982cf8845b" SEC_TYPE="ext2"
TYPE="ext3"
+ sudo vol_id /dev/sda7
ID_FS_USAGE=filesystem
ID_FS_TYPE=ext3
ID_FS_VERSION=1.0
ID_FS_UUID=2ff24943-33a8-43e9-97da-73982cf8845b
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sda8
/dev/sda8
+ sudo blkid /dev/sda8
/dev/sda8: UUID="7b3d8e1c-2080-4016-a672-1b2fed6ce7c8"
TYPE="reiserfs"
+ sudo vol_id /dev/sda8
ID_FS_USAGE=filesystem
ID_FS_TYPE=reiserfs
ID_FS_VERSION=3.6
ID_FS_UUID=7b3d8e1c-2080-4016-a672-1b2fed6ce7c8
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sdb
/dev/sdb
+ sudo blkid /dev/sdb
+ sudo vol_id /dev/sdb
/dev/sdb: unknown volume type
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sdb1
/dev/sdb1
+ sudo blkid /dev/sdb1
/dev/sdb1: UUID="403d16ca-ce7b-44f4-bcb3-87654deb5466" SEC_TYPE="ext2"
TYPE="ext3"
+ sudo vol_id /dev/sdb1
ID_FS_USAGE=filesystem
ID_FS_TYPE=ext3
ID_FS_VERSION=1.0
ID_FS_UUID=403d16ca-ce7b-44f4-bcb3-87654deb5466
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sdb2
/dev/sdb2
+ sudo blkid /dev/sdb2
/dev/sdb2: UUID="064f1127-b29f-4146-8ffd-181c4ac9f1ee" SEC_TYPE="ext2"
TYPE="ext3"
+ sudo vol_id /dev/sdb2
ID_FS_USAGE=filesystem
ID_FS_TYPE=ext3
ID_FS_VERSION=1.0
ID_FS_UUID=064f1127-b29f-4146-8ffd-181c4ac9f1ee
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
+ for i in '$(awk '\''NR >1 && $2 > 0 {print "/dev/" $4}'\'' /proc/
partitions)'
+ echo /dev/sdb3
/dev/sdb3
+ sudo blkid /dev/sdb3
/dev/sdb3: UUID="feb26ef0-be8e-49ac-8fbe-2c443549c553"
TYPE="reiserfs"
+ sudo vol_id /dev/sdb3
ID_FS_USAGE=filesystem
ID_FS_TYPE=reiserfs
ID_FS_VERSION=3.6
ID_FS_UUID=feb26ef0-be8e-49ac-8fbe-2c443549c553
ID_FS_LABEL=
ID_FS_LABEL_SAFE=
Note that it says /dev/sda3: unknown volume type - but gparted says
that's my extended partition.
Unless you can see a quick fix, I think I will follow your suggestion
to change that 2 to a 0 and try to remember to do the occasional
manual fsck.
It does look awfully like the first launchpad error.
Many thanks for all your efforts - especially for those two scripts.
Regards,
Jeremy
Robert Citek
> that's my extended partition.
Which makes sense, since /dev/sda3 doesn't contain a filesystem.
> Unless you can see a quick fix, I think I will follow your suggestion
> to change that 2 to a 0 and try to remember to do the occasional
> manual fsck.
Let us know if that works.
> It does look awfully like the first launchpad error.
You might want to file a bug on launchpad and let those more
in-the-know decide if it is the same bug. In your report include the
error message along with the output of this script:
true ; (
sudo echo -n
set -x
cat /etc/issue.net
uname -va
blkid -v
cat /etc/fstab
sudo fdisk -l
for i in $(awk 'NR >1 && $2 > 0 {print "/dev/" $4}' /proc/partitions) ; do
echo $i
sudo blkid $i
done
) >& output.txt
> Many thanks for all your efforts - especially for those two scripts.
No problem. Again let us know how it goes and post the url to the bug
report here. That way we can all follow it.
BTW, the above launchpad link mentions that the bug was fixed in
Gutsy, which was released yesterday. May be worth exploring on an
experimental machine or VM.
Regards,
- Robert
Jeremy Boden
Thanks for your efforts.
I've added a comment (with all the details) to
https://bugs.launchpad.net/ubuntu/+bug/106782/comments/1
("checkfs & fsck command options don't work with fsck.jfs at boot")
I am now going to try changing the 0 2 in fstab to 0 0.
I will post if anything serious happens!
Thanks & Regards,
Jeremy