CentOS - Apr 2009 - Yum update corrupted something badly - can only get access through SSH terminal or Webmin

Hey there -

This morning I ran the yum updater through Webmin as I do every month  
or so - after about two hours I realized that I still had the same  
updating screen going - and no response.   Seemed strange.

Simple attempts to kill it didn't work so I sent a command to reboot  
the system - that seemed to work, till I turned on the monitor  
directly connected to the server -

I saw this error go by - and then only a spinning cursor - never get  
to a login screen.  (I power cycled again and saw it go by again - and  
recognized it when I tried again at Webmin and through a terminal  
connection).

Traceback (most recent call last):
   File "/usr/bin/yum", line 28, in ?
     import yummain
   File "/usr/share/yum-cli/yummain.py", line 29, in ?
     import cli
   File "/usr/share/yum-cli/cli.py", line 30, in ?
     import output
   File "/usr/share/yum-cli/output.py", line 26, in ?
     from i18n import _
ImportError: No module named i18n

I deleted the yum cash and no improvement.  I tried the yum clean and  
no imporvement
[root at trailrunner yum]# yum clean all
Traceback (most recent call last):
   File "/usr/bin/yum", line 28, in ?
     import yummain
   File "/usr/share/yum-cli/yummain.py", line 29, in ?
     import cli
   File "/usr/share/yum-cli/cli.py", line 30, in ?
     import output
   File "/usr/share/yum-cli/output.py", line 26, in ?
     from i18n import _
ImportError: No module named i18n
[root at trailrunner yum]#

Not at all sure what happened or why, or where I even start trying to  
straighten this mess out.

How do I blow away the apparent yum junk in the hope of getting a  
local login capability back?

How or what should I do to reset yum - seems reinstalling the latest  
version is probably the route.  What is the version that CentOS  
uses?   Should I go the RPM route or try something more specific?

Totally unexpected problem and leaving me feeling very uncertain about  
things.

On Wed, Apr 22, 2009 at 11:47 AM, Dan Roberts <dan at jlazyh.com>
wrote:
> This morning I ran the yum updater through Webmin as I do every month
> or so - after about two hours I realized that I still had the same
> updating screen going - and no response. ? Seemed strange.
No idea why you are getting this problem or using Webmin to update
your box, but I suggest you rethink the idea of only doing updates
"every month or so". You could have the box exposed to an exploit for
a long time, with that schedule. Hopefully someone else on the list
will have a solution for your immediate problem. GL

On Wed, Apr 22, 2009 at 12:47 PM, Dan Roberts <dan at jlazyh.com>
wrote:
> Hey there -
>
> This morning I ran the yum updater through Webmin as I do every month
> or so - after about two hours I realized that I still had the same
> updating screen going - and no response. ? Seemed strange.
So you probably missed the notes and announcements about the 5.3
release, as well as the release notes about upgrade hiccups.
There are also quite a few updates here, so the update will take some
time to run.
> Simple attempts to kill it didn't work so I sent a command to reboot
> the system - that seemed to work, till I turned on the monitor
> directly connected to the server -
Updating via webmin may not be the best approach, as some things may
update which could impact webmin.
> I saw this error go by - and then only a spinning cursor - never get
> to a login screen. ?(I power cycled again and saw it go by again - and
> recognized it when I tried again at Webmin and through a terminal
> connection).
Power cycling a system should be considered a last-resort option, not
a method for troubleshooting.
> Traceback (most recent call last):
> ? File "/usr/bin/yum", line 28, in ?
> ? ? import yummain
> ? File "/usr/share/yum-cli/yummain.py", line 29, in ?
> ? ? import cli
> ? File "/usr/share/yum-cli/cli.py", line 30, in ?
> ? ? import output
> ? File "/usr/share/yum-cli/output.py", line 26, in ?
> ? ? from i18n import _
> ImportError: No module named i18n
>
> I deleted the yum cash and no improvement. ?I tried the yum clean and
> no imporvement
> [root at trailrunner yum]# yum clean all
> Traceback (most recent call last):
> ? File "/usr/bin/yum", line 28, in ?
> ? ? import yummain
> ? File "/usr/share/yum-cli/yummain.py", line 29, in ?
> ? ? import cli
> ? File "/usr/share/yum-cli/cli.py", line 30, in ?
> ? ? import output
> ? File "/usr/share/yum-cli/output.py", line 26, in ?
> ? ? from i18n import _
> ImportError: No module named i18n
> [root at trailrunner yum]#
>
> Not at all sure what happened or why, or where I even start trying to
> straighten this mess out.
For systems which experienced issues moving from 5.2 to 5.3, a 'yum
clean all' resolved the majority of issues.
> How do I blow away the apparent yum junk in the hope of getting a
> local login capability back?
Yum doesn't have anything to do with logging into the system at all.
I'm not sure why you think this would help. However as previously
stated a 'yum clean all' should 'blow away' the yum bits.
Further
nuking yum transactions is not to be done lightly, and requires the
yum-utils package to be installed. We'll go over that command later if
we have to.
> How or what should I do to reset yum - seems reinstalling the latest
> version is probably the route. ?What is the version that CentOS
> uses? ? Should I go the RPM route or try something more specific?
>
> Totally unexpected problem and leaving me feeling very uncertain about
> things.
Please subscribe to the Centos-Announce list. You got blind-sided by a
very public release because you only update every month or so, and
don't generally follow updates in a manner most admins would consider
'good practice'. I believe that had you been aware of the 5.3 release,
you would have been a bit more careful in things.


-- 
During times of universal deceit, telling the truth becomes a revolutionary act.
George Orwell

On Wed, Apr 22, 2009 at 12:47 PM, Dan Roberts <dan at jlazyh.com>
wrote:
> Hey there -
>
> This morning I ran the yum updater through Webmin as I do every month
> or so - after about two hours I realized that I still had the same
> updating screen going - and no response. ? Seemed strange.
>
> Simple attempts to kill it didn't work so I sent a command to reboot
> the system - that seemed to work, till I turned on the monitor
> directly connected to the server -
My first approach would be to use the lower level rpm to determine
what packages may be in error. Start with yum:

   rpm -qV yum

It will list files that have changed from the initial package installation:
  [u108752 at linbox ~]$ rpm -qV yum
  S.5....T  c /etc/yum.conf

If you see any files besides the user editable ones (such as the
yum.conf, for example) it could mean you have a corrupted binary.  If
so, re-install the package:

  rpm -ivh yum-x.x.x..rpm

You may need to use the --oldpackage flag  or even --force.

If you don't see anything that jumps out as wrong, try running the rpm
verify against all packages;

  rpm -qaV

Ignore things like conf files and logs and look for packages with a S
(checksum change) in the results colums..

Once you can get yum reinstalled, update glibc first, then the rest of
the system.


> I saw this error go by - and then only a spinning cursor - never get
> to a login screen. ?(I power cycled again and saw it go by again - and
> recognized it when I tried again at Webmin and through a terminal
> connection).
>
> Traceback (most recent call last):
> ? File "/usr/bin/yum", line 28, in ?
> ? ? import yummain
> ? File "/usr/share/yum-cli/yummain.py", line 29, in ?
> ? ? import cli
> ? File "/usr/share/yum-cli/cli.py", line 30, in ?
> ? ? import output
> ? File "/usr/share/yum-cli/output.py", line 26, in ?
> ? ? from i18n import _
> ImportError: No module named i18n
>
> I deleted the yum cash and no improvement. ?I tried the yum clean and
> no imporvement
> [root at trailrunner yum]# yum clean all
> Traceback (most recent call last):
> ? File "/usr/bin/yum", line 28, in ?
> ? ? import yummain
> ? File "/usr/share/yum-cli/yummain.py", line 29, in ?
> ? ? import cli
> ? File "/usr/share/yum-cli/cli.py", line 30, in ?
> ? ? import output
> ? File "/usr/share/yum-cli/output.py", line 26, in ?
> ? ? from i18n import _
> ImportError: No module named i18n
> [root at trailrunner yum]#
>
> Not at all sure what happened or why, or where I even start trying to
> straighten this mess out.
>
> How do I blow away the apparent yum junk in the hope of getting a
> local login capability back?
>
> How or what should I do to reset yum - seems reinstalling the latest
> version is probably the route. ?What is the version that CentOS
> uses? ? Should I go the RPM route or try something more specific?
>
> Totally unexpected problem and leaving me feeling very uncertain about
> things.
>
>
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> http://lists.centos.org/mailman/listinfo/centos
>