problems with 10.04 upgrade to 12.04 (procps)

NoOp glgxg at sbcglobal.net
Sat Sep 8 23:24:33 UTC 2012


On 09/08/2012 08:07 AM, Liviu Andronic wrote:
> On Fri, Sep 7, 2012 at 9:58 PM, NoOp <glgxg at sbcglobal.net> wrote:
>> On 09/07/2012 11:43 AM, Liviu Andronic wrote:
>>> However, I'm still unable to boot Precise. Using default kernel
>>> parameters (for 3.2.0) I get the following messages:
>>> [..]
>>> Begin: Running /scripts/init-bottom ... done.
>>> could not log bootup: Address already in use
>>> [ 10.1 ] init: plymouth main process (327) terminated with status 69
>>> mount: Too many levels of symbolic links
>>> mountall: mount /run [349] terminated with status 32
>>> mountall: Filesystem could not be mounted: /run
>>> An error occurred while mounting /run.
>>
>> Are you able to boot to recovery mode?
>>
> No, not really. If I try to boot with default kernel options (the
> 'recovery' variant), then the booting gets stuck as described earlier
> at mounting the various partitions. The only way to get myself on a
> console is to use the 'rw init=/bin/bash' kernel parameters which
> let's me fall back to a console with 'rw' permission on /. This is not
> quite the recovery mode.
> 
> 
>> If so, can you try the dpkg option?
>>
> I'm not sure that I follow. How do I do that? Is this a kernel
> parameter, or some other option?
> 

It's a menu option in Recovery mode. Basically it runs a script that
attempts to fix broken packages & update the system. Given that you are
at a cli console, you can actually run the same script from there. To
see what it does:

$ cat /lib/recovery-mode/options/dpkg

To run directly from the cli:

$ sudo /lib/recovery-mode/options/dpkg

Recovery mode options:

/lib/recovery-mode/recovery-menu
/lib/recovery-mode/recovery-mode
/lib/recovery-mode/options/apt-snapshots
/lib/recovery-mode/options/clean
/lib/recovery-mode/options/dpkg
/lib/recovery-mode/options/failsafeX
/lib/recovery-mode/options/fsck
/lib/recovery-mode/options/grub
/lib/recovery-mode/options/network
/lib/recovery-mode/options/root
/lib/recovery-mode/options/system-summary

To see & actually use the recovery-mode menu from the cli:
$ sudo /lib/recovery-mode/recovery-menu

Forgot to ask & forgive me if you've mentioned this in the thread: are
you able to boot to an older kernel (i.e., 2.6.32-42-generic)?
If so, I would boot to recovery mode in that kernel, try to clean up
(using the dpkg option on the recovery menu (or cli), and then try the
upgrade again.

> 
>>> Then I press S to skip mounting and I get the following:
>>> The disk drive for / is not ready yet or not present.
>>> Skipping /run at user request
>>> udevd[359]: error: runtime dir '/run/dev/' not writable, for now
>>> falling back to '/dev/.udev'
>>>
>>>
>>> Next I press M for manual recovery and get this:
>>> Root filesystem check failed.
>>> A maintenance shell will now be started.
>>
>> That sounds like the cryptswap issue:
>> <https://bugs.launchpad.net/ubuntu/+source/cryptsetup/+bug/874774>
>> if so, then generally it will time out after a moment or two.
>>
> Hmm, this didn't work for me. I waited for ~30min and the screen was
> still stuck at 'continue to wait' while trying to mount the /
> partition.
> 
> But I'm not sure that I'm seeing the same issue. For one, the first
> time such an issue arises is while running 'mount /run'; I don't have
> such a device in /etc/fstab. And the second thing is that I tried
> booting up without swap, commenting out the /dev/sda7 and
> /dev/mapper/cryptoswap related items in /etc/fstab (
> http://paste.ubuntu.com/1191361/ ), and still got the same booting
> trouble.
> 
> 
>>> All this looks strange to me, as I force checked / (and all other FS)
>>> and they're all clean. I'm wondering though if after installing all
>>> the packages there were some Precise-specific scripts that didn't get
>>> executed to prepare the new OS. Among other things, perhaps /etc/fstab
>>> didn't get updated? (Here's my current /etc/fstab:
>>> http://paste.ubuntu.com/1191361/ )
>>
>> Do the UUID's match against:
>>
>> $ sudo blkid
>>
> Yes. The only difference that I can spot is that the 'swap' partition
> doesn't come up in the blkid output.
>  /etc/fstab: http://paste.ubuntu.com/1191361/
>  blkid: http://paste.ubuntu.com/1192838/
>  fdisk -l: http://paste.ubuntu.com/1192840/
> 
> 
>> And  does
>> $ sudo fdisk -l
>> show the cryptswap?
>>
> The /dev/sda9 partition is in the output, but it is marked as 'linux',
> not 'cryptswap'.
> 
> 
>>>
>>> Should I run 'do-release-upgrade' again?
>>
>> Try the recovery dpkg option first.
>>
> I'm not sure how exactly I should use the recovery dpkg option.
> 
> Thanks
> Liviu
> 






More information about the ubuntu-users mailing list