9.10 upgrade failure

[parkavenue]

When upgrading from 9.04 to 9.10 I received the following error:

"The upgrade is now aborted. Your system could be in an unusable state. A recovery will run now (dpkg --configure -a). Please report this bug against the 'update-manager' package and include the files in /var/log/dist-upgrade/ in the bug report.
installArchive() failed"

9.04 is listed in the boot menu - booting into 9.04 or recovery mode fails. I see "One or more of the mounts listed in /etc/fstab cannot yet be mounted" and I'm left at the command line.

Help please!

[prshah]

"The upgrade is now aborted. Your system could be in an unusable state. A recovery will run now (dpkg --configure -a)

booting into 9.04 or recovery mode fails. I see "One or more of the mounts listed in /etc/fstab cannot yet be mounted" and I'm left at the command line.


Try the command
dpkg --configure -a at the command prompt. For a more detailed explanation, please see Hardy is wonderful... (http://ubuntuforums.org/showthread.php?t=780531).

If that doesn't work, try following the tips in this post: Re: System will not boot (http://ubuntuforums.org/showthread.php?p=8286981&highlight=%2Fetc%2Ffstab#post8286981) to disable the offending /etc/fstab entries (provided they are not related to your linux filesystem); then try a recovery console boot and then try the dpkg command as above, over again.

Short of these, all I can suggest is that perhaps you can use a live CD to pull out your important files, and then start over? I usually avoid advising reinstalls.

[parkavenue]

The dpkg command didn't work for me. I get:
sudo: unable to resolve host (none)
dpkg: unable to access dpkg status area: Read-only file system

so I went with the livecd to check out /etc/fstab. I'm not sure what to comment out but this is what I see:
unionfs / unionfs rw 0 0
tmpfs /tmp tmpfs nosuid,nodev 0 0
/dev/sda5 swap swap defaults 0 0

Also, I notice that the command line (which I get to when attempting to boot into recovery mode) says "root@(none). I'm not sure why my name is not listed there as before.

And finally, the original error that I got occurred during the beginning of the 9.10 installation, after the packages were downloaded. Not sure if that helps.

[parkavenue]

Okay, that was a hasty reply. Turns out I looked in the wrong /etc/fstab file. Perhaps it had to do with being on the livecd? Will try prshah's suggestion again and report back.

[prshah]

I looked in the wrong /etc/fstab file. Perhaps it had to do with being on the livecd?

Yes, on the live CD, the /etc/fstab is a "virtual" one. There are more details in my post on how to access the "actual" one.

You seem to be missing a hostname file (and possibly others). Here's something else to try in recovery mode.
cat yourhostname > /etc/hostname Replace "yourhostname" with whatever it was before the upgrade. reboot, and try the dpkg command again. In Recovery mode, at a root prompt, ("#" instead of "$"), you do not need to use sudo, so run the dpkg command without sudo.

[parkavenue]

The correct /etc/fstab file contains:
# /etc/fstab: static file system information.
#
# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc defaults 0 0
# /dev/sda6
#UUID=02b374cb-05e0-4f54-a22a-8c1657567820 / ext3 relatime,errors=remount-ro 0 1
# /dev/sda7
UUID=af84d309-0966-4725-b895-34e4665b18b3 /home ext3 relatime 0 2
# /dev/sda5
UUID=ce2569b5-cfd9-4349-84dc-54f31ea645e1 none swap sw 0 0
/dev/scd0 /media/cdrom0 udf,iso9660 user,noauto,exec,utf8 0 0

So it looks like /dev/sda6 has errors? So I commented out its UUID line but didn't see a difference.
When I boot into recovery mode I see "/", /tmp, /home and swap are not mounting, rather they are all listed as waiting for <something>.

I tried the following command:


cat yourhostname > /etc/yourhostname

but it returned "bash: /etc/parkavenue: Read-only file system"

[prshah]

So it looks like /dev/sda6 has errors? So I commented out its UUID line but didn't see a difference.
but it returned "bash: /etc/parkavenue: Read-only file system"

Ok you can't comment out anything in that /etc/fstab (All are your linux install related partitions). Please reverse the process.

If you have a problem with /dev/sda6, perhaps a diskcheck can fix it; please boot off a live CD, then open a terminal (Applications-Accessories-Terminal) and give the command
sudo umount /dev/sda6 #ignore errors about "not mounted"
sudo e2fsck -pcvf /dev/sda6 Post back results and errors if any. If you are asked to run the scan "manually", ie, without the -p switch, then please use
sudo e2fsck -ycvf /dev/sda6 Warning! Small chance of data loss, so please do so at your own risk.

[parkavenue]

Okay... I had ran fsck on sda6 and sda7 before seeing your last reply.
I ran it differently than was suggested in your post, so I will try your code tomorrow. I had ran this:


fsck /dev/sda6

I answered yes when it asked "Fix?" (many times). I rebooted, but I'm still seeing "One or more of the mounts listed in /etc/fstab cannot yet be mounted."

I know that sda7 has had "issues" (such as failing the disk check) prior to attempting to install 9.10 but I ignored them. The computer ran with those issues for a while, on 8.04, 8.10 and 9.04.

Will try your code tomorrow. Thanks for your replies.

[prshah]

I answered yes when it asked "Fix?" (many times). I rebooted, but I'm still seeing "One or more of the mounts listed in /etc/fstab cannot yet be mounted."

Sorry, but my "code" will not do anything more than you have already done. The "-y" switch says "answer yes to questions". You can feel free to try it, but keep your expectations low :)

Can you post back the output of
sudo blkid # or just blkid if sudo fails from recovery mode (or equivalent, _not_ live CD).

[parkavenue]

blkid returns the following:
/dev/sda5: UUID="ce25..." TYPE="swap"
/dev/sda6: UUID="02b3..." SEC_TYPE="ext2" TYPE="ext3"
/dev/sda7: UUID="af84..." SEC_TYPE="ext2" TYPE="ext3"

[parkavenue]

I ran the e2fsck command on /dev/sda6. It returned FILE SYSTEM WAS MODIFIED. It listed a number of files including inodes used, non-contiguous inodes, etc. I also noticed "21 bad blocks." Should I write any of this down?
I'm awaiting the results of /dev/sda7 (this is the one that is known to have errors.)

[parkavenue]

Running /dev/sda7 was taking an extremely long time (over an hour for ~20%). So I canceled the error check (Ctrl+D?) but now I cannot access /dev/sda7. I commented it out from the /etc/fstab file but I still boot into recovery mode to see the mounting issue.
So I'm still stuck on the original issue (can't boot, waiting for mounting) but in the meantime I suppose I should try to remount /dev/sda7.
Is this the point at which I should do a fresh re-install of 9.04?

[prshah]

noticed "21 bad blocks."


Running /dev/sda7 was taking an extremely long time

Is this the point at which I should do a fresh re-install of 9.04?

badblocks are a sign of physical damage to the hard disk drive. I usually replace HDDs with bad sectors (badblocks) as soon as possible.

So, yes, I guess you should back up what data of yours is still accessible using a live CD, and then take the plunge for a re-installation.

With badblocks on the HDD, it _may_ fail.

Good luck!

[parkavenue]

Thanks again for your reply. I found this post with a few different suggestions. It looks like this helped one of the users so I will try this next.
http://ubuntuforums.org/showthread.php?t=1308671

[parkavenue]

mount -o remount,rw /
dpkg --configure -a

seems to be doing something... hopefully fixing things. [-o<

[parkavenue]

Not sure anyone is following this but I'm just replying for my own documentation purposes...

After failing a 9.04 to 9.10 install, I rebooted to find that the system was waiting for the partitions to mount. I was impatient but they wouldn't mount if I waited ~1minute.

Eventually I ran disk-checks on each partition. The erroneous partition took an extremely look time (>24hours) and then started running again once it finished (so there's definitely something wrong with the harddrive). So I stopped it and rebooted.


umount /dev/sda7
e2fsck /dev/sda7 -pcvf # or ycvf


Then ran the following and I saw '9.10' at the command line when I rebooted into recovery mode.

dpkg --configure -a
sudo apt-get upgrade
sudo apt-get dist-upgrade


I went in to /etc/fstab (as described in above replies) using the live CD and I un-commented the /dev/sda7 UUID-xxx line. (I think this was necessary).

I can currently boot into 9.10 recovery mode and everything looks fine (networking works, etc.). Regular 9.10 has some issues (with Nautilus?) but I'll start a new thread on that.

Ultimately, after gra,bing my stuff (whatever I haven't already grabbed), a new harddrive will be in order... Thanks again for the help prshah!