Mishap upgrading from Woody to Etch

[Chris Austin]

Hi,

I have had a Debian system working very well since 2003, when I installed it
from the Stable system, which was Woody at the time.  Recently I wanted to
install the gap package, (Groups, Algorithms and Programming computer
algebra system), and for this I had to start upgrading packages to the
current Stable system, which is Etch.  However I hit a circular dependency
problem trying to upgrade libc6, so started upgrading packages with the
Check Dependencies checkbox deselected, in the KPackage windows that were
started by Konqueror, when I downloaded packages from the Debian website.
At some stage e2fsprogs and e2fslibs were upgraded in this way, but I always
hit an error trying to upgrade libc6, even after I exited KDE and tried to
stop kdm with the kill signal in top.  I then found that kdm no longer
restarted KDE, so to access the Internet I had to dual boot to MS Windows.
Trying to reboot in Linux now starts a file system that is Read Only, with
lots of warnings and errors.  It seems from some of the error messages that
I now need to install the libuuid1 package, but I cannot do this because the
file system is started Read Only.  I can get a writable file system by
booting from the old Woody rescue.bin floppy, but file locations then need
to be prepended by /target, and dpkg gives an error message: relocation
error, symbol _obstack_begin, version GLIBC_2.0 not defined in file
libc.so.6 with link time reference.

Best regards,
Chris Austin.



--
To UNSUBSCRIBE, email to debian-user-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Andrew Sackville-West]

On Sun, Sep 09, 2007 at 07:34:22PM +0100, Chris Austin wrote:
> Hi,
>
> I have had a Debian system working very well since 2003, when I installed it
> from the Stable system, which was Woody at the time.  Recently I wanted to

... snipped woes of a partial upgrade ...

> hit an error trying to upgrade libc6, even after I exited KDE and tried to
> stop kdm with the kill signal in top.  I then found that kdm no longer
> restarted KDE,

...

Chris, I'm quite sure that this kind of partial upgrade is completely
unsupported. It also appears that you've borked the system
sufficiently well that you're better off starting from
scratch. sorry. Debian went through some serious changes between
woody and etch -- a whole new X system, a whole new libc, among
others. Your would be better spent backing up what you need and
reinstalling etch. It is surely possible to save your system, but it
will probably be very time-consuming and you will end up with an etch
system anyway.

.02

A

[Matus UHLAR - fantomas]

On 09.09.07 19:34, Chris Austin wrote:

> I have had a Debian system working very well since 2003, when I installed it
> from the Stable system, which was Woody at the time.  Recently I wanted to
> install the gap package, (Groups, Algorithms and Programming computer
> algebra system), and for this I had to start upgrading packages to the
> current Stable system, which is Etch.  However I hit a circular dependency
> problem trying to upgrade libc6, so started upgrading packages with the
> Check Dependencies checkbox deselected, in the KPackage windows that were
> started by Konqueror, when I downloaded packages from the Debian website.
> At some stage e2fsprogs and e2fslibs were upgraded in this way, but I always
> hit an error trying to upgrade libc6, even after I exited KDE and tried to
> stop kdm with the kill signal in top.  I then found that kdm no longer
> restarted KDE, so to access the Internet I had to dual boot to MS Windows.
> Trying to reboot in Linux now starts a file system that is Read Only, with
> lots of warnings and errors.  It seems from some of the error messages that
> I now need to install the libuuid1 package, but I cannot do this because the
> file system is started Read Only.  I can get a writable file system by
> booting from the old Woody rescue.bin floppy, but file locations then need
> to be prepended by /target, and dpkg gives an error message: relocation
> error, symbol _obstack_begin, version GLIBC_2.0 not defined in file
> libc.so.6 with link time reference.

That's why yyou should never upgrade across two versions. You still can
upgrade to sarge and then to etch.

--
Matus UHLAR - fantomas, uhlar@... ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
"Two words: Windows survives." - Craig Mundie, Microsoft senior strategist
"So does syphillis. Good thing we have penicillin." - Matthew Alton


--
To UNSUBSCRIBE, email to debian-user-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Adam Hardy (debian)]

>On Sun, Sep 09, 2007 at 07:34:22PM +0100, Chris Austin wrote:
>> Hi,
>>
>> I have had a Debian system working very well since 2003, when I installed it
>> from the Stable system, which was Woody at the time.  Recently I wanted to
>
>... snipped woes of a partial upgrade ...
>
>> hit an error trying to upgrade libc6, even after I exited KDE and tried to
>> stop kdm with the kill signal in top.  I then found that kdm no longer
>> restarted KDE,
>
>...
>
>Chris, I'm quite sure that this kind of partial upgrade is completely
>unsupported. It also appears that you've borked the system
>sufficiently well that you're better off starting from
>scratch. sorry. Debian went through some serious changes between
>woody and etch -- a whole new X system, a whole new libc, among
>others. Your would be better spent backing up what you need and
>reinstalling etch. It is surely possible to save your system, but it
>will probably be very time-consuming and you will end up with an etch
>system anyway.

I have a 5 year-old system hosted on Xen by a hosting company, which I only use for Java and mysql - currently it's running Woody and being slap-dash, I tried to install a piece of software using apt-get, which promptly failed.

Seeing the conversation above, it sounds bad - but I'm going to try something otherwise I have to ask the hosting co to wipe my Xen slice for me.

If I put 'sarge' in all the sources.list urls, I should be able to upgrade to sarge OK? Or is it too late?

I just want to install galrey to generate some image galleries for some images I scp'd over. I guess I could find a java webapp which could do this too if this fails.

Thanks!
Adam

[Andrew Sackville-West]

On Fri, Jun 20, 2008 at 04:51:25AM -0700, Adam Hardy (debian) wrote:

[snipped tales of woe regarding mixed systems]

>
> I have a 5 year-old system hosted on Xen by a hosting company, which I only
> use for Java and mysql - currently it's running Woody and being slap-dash, I
> tried to install a piece of software using apt-get, which promptly failed.
>
> Seeing the conversation above, it sounds bad - but I'm going to try
> something otherwise I have to ask the hosting co to wipe my Xen slice for
> me.
>
> If I put 'sarge' in all the sources.list urls, I should be able to upgrade
> to sarge OK? Or is it too late?

you should read the release notes for sarge first:

http://www.debian.org/releases/sarge/i386/release-notes/ch-upgrading.en.html

and you might need to use "oldstable" as the identifier in
sources.list. But yes, you should be able to upgrade to sarge. Once
you're in the think of it, you might want to take the next step and
move up to etch since it may go to old stable in the next few
months. Then when that happens, you can move up to lenny when it's
fairly fresh. At that point, you can let it go for another 5 years or
so... ;-)

A

[Adam Hardy (debian)]

Andrew Sackville-West on 20/06/08 14:38, wrote:

> On Fri, Jun 20, 2008 at 04:51:25AM -0700, Adam Hardy (debian) wrote:
>
> [snipped tales of woe regarding mixed systems]
>
>> I have a 5 year-old system hosted on Xen by a hosting company, which I only
>> use for Java and mysql - currently it's running Woody and being slap-dash, I
>> tried to install a piece of software using apt-get, which promptly failed.
>>
>> Seeing the conversation above, it sounds bad - but I'm going to try
>> something otherwise I have to ask the hosting co to wipe my Xen slice for
>> me.
>>
>> If I put 'sarge' in all the sources.list urls, I should be able to upgrade
>> to sarge OK? Or is it too late?
>
> you should read the release notes for sarge first:
>
> http://www.debian.org/releases/sarge/i386/release-notes/ch-upgrading.en.html
>
> and you might need to use "oldstable" as the identifier in
> sources.list. But yes, you should be able to upgrade to sarge. Once
> you're in the think of it, you might want to take the next step and
> move up to etch since it may go to old stable in the next few
> months. Then when that happens, you can move up to lenny when it's
> fairly fresh. At that point, you can let it go for another 5 years or
> so... ;-)

OK!

Those upgrade instructions warn:

"Important! You should not upgrade using telnet, rlogin, rsh, or from an X
session managed by xdm, gdm or kdm etc on the machine you are upgrading. That is
because each of those services may well be terminated during the upgrade, which
can result in an inaccessible system that is only half-upgraded."

Do you think ssh is deliberately excluded from that list of remote shells?

I can only get onto my server via ssh since the machine is at the hosting co's
site (and indeed it's only a Xen slice, not a whole OS).


Adam



--
To UNSUBSCRIBE, email to debian-user-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Daniel Burrows]

On Fri, Jun 20, 2008 at 06:38:59AM -0700, Andrew Sackville-West <andrew@...> was heard to say:

> On Fri, Jun 20, 2008 at 04:51:25AM -0700, Adam Hardy (debian) wrote:
> > If I put 'sarge' in all the sources.list urls, I should be able to upgrade
> > to sarge OK? Or is it too late?
>
> you should read the release notes for sarge first:
>
> http://www.debian.org/releases/sarge/i386/release-notes/ch-upgrading.en.html
>
> and you might need to use "oldstable" as the identifier in
> sources.list. But yes, you should be able to upgrade to sarge. Once
> you're in the think of it, you might want to take the next step and
> move up to etch since it may go to old stable in the next few
> months. Then when that happens, you can move up to lenny when it's
> fairly fresh. At that point, you can let it go for another 5 years or
> so... ;-)

  You can use "sarge" in sources.list.  If you use "oldstable" you'll be
rudely upgraded to etch when lenny is released. ;-)

  Daniel


--
To UNSUBSCRIBE, email to debian-user-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Sven Joachim]

On 2008-06-20 15:58 +0200, Adam Hardy wrote:

> Those upgrade instructions warn:
>
> "Important! You should not upgrade using telnet, rlogin, rsh, or from
> an X session managed by xdm, gdm or kdm etc on the machine you are
> upgrading. That is because each of those services may well be
> terminated during the upgrade, which can result in an inaccessible
> system that is only half-upgraded."
>
> Do you think ssh is deliberately excluded from that list of remote shells?

Yes, see the same section:

,----
| The distribution upgrade should be done either locally from a textmode
| virtual console (or a directly connected serial terminal), or remotely
| via an ssh link.
|
| In order to gain extra safety margin when upgrading remotely, we suggest
| that you run upgrade processes in the virtual console provided by the
| screen program, which enables safe reconnection and ensures the upgrade
| process is not interrupted even if the remote connection process fails.
`----

Sven


--
To UNSUBSCRIBE, email to debian-user-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Adam Hardy (debian)]

Sven Joachim on 20/06/08 15:06, wrote:

> On 2008-06-20 15:58 +0200, Adam Hardy wrote:
>
>> Those upgrade instructions warn:
>>
>> "Important! You should not upgrade using telnet, rlogin, rsh, or from
>> an X session managed by xdm, gdm or kdm etc on the machine you are
>> upgrading. That is because each of those services may well be
>> terminated during the upgrade, which can result in an inaccessible
>> system that is only half-upgraded."
>>
>> Do you think ssh is deliberately excluded from that list of remote shells?
>
> Yes, see the same section:
>
> ,----
> | The distribution upgrade should be done either locally from a textmode
> | virtual console (or a directly connected serial terminal), or remotely
> | via an ssh link.
> |
> | In order to gain extra safety margin when upgrading remotely, we suggest
> | that you run upgrade processes in the virtual console provided by the
> | screen program, which enables safe reconnection and ensures the upgrade
> | process is not interrupted even if the remote connection process fails.
> `----

What's the 'virtual console provided by the screen program'? Some sort of
bolt-on to ssh?


--
To UNSUBSCRIBE, email to debian-user-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Steve Kemp]

On Fri Jun 20, 2008 at 15:13:59 +0100, Adam Hardy wrote:

> What's the 'virtual console provided by the screen program'? Some sort of
> bolt-on to ssh?

  "screen" is a program which allows you to run commands in "windows"
 even on a text console.

  See here for a reasonable introduction:

    http://www.debian-administration.org/articles/34

Steve
--
Managed Anti-Spam Service
http://mail-scanning.com/


--
To UNSUBSCRIBE, email to debian-user-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...