Re: [cobalt-developers] RaQ4 - GLIBC errors. Can't telnet, SSH etc.

[Brian Baxter <bbaxter@xxxxxxxxxxxxxxxxx>]

OK. I couldn't leave this one alone. I just reproduced this problem. It is
indeed casued by an unsuccessful migration to glibc 2.2.3. (note that
unsuccessful does not mean failed!)

If the install of 2.2.3 was not installed to /usr/lib (sometimes /lib
depending on the system) it is possible to restore access provided that you
can gain access to the shell in the first place. In /etc/ld.so.conf, rm the
line which points to the 'new' directory. Again, this will only work *if*
the attempt to move to 2.2.3 was done to a dir other than the default
/usr/lib! If the migration was done using the standard install defaults,
then the restore option is really the only one that has a chance of
succeeding. 

~~Brian
------------------------------
Brian Baxter
Executive Director of Technologies
GraffitiWorkz.com
bbaxter@xxxxxxxxxxxxxxxxx
------------------------------

At 06:26 PM 11/21/2001 -0600, you wrote:
>At 01:38 PM 11/21/2001 -0000, you wrote:
>>Our RaQ4 has shut us out. We can't telnet, FTP, SSH etc.
>>
>>Attempting an SSH results in:
>>
>>-sh: /lib/ld-linux.so.2: version `GLIBC_2.2.3' not found (required by
>>/usr/lib/libc.so.6)
>>-sh: /lib/ld-linux.so.2: version `GLIBC_2.2' not found (required by
>>/usr/lib/libc.so.6)
>>C
>>
>>Nothing has been installed or de-installed today.
>>
>>Any help greatly appreciated.
>>
>>Rgds
>>
>
>Umm. Hate to tell you this, but *something* has changed. One of the
>following things is most likely the case:
>
>1) A migration to a different version of glibc went horribly awry (been
>there and and done that, it hurts).
>
>2) Someone or something changed the symlinks in /lib and they no longer
>point where they need to.
>
>3) Someone or something mv'd or rm'd required versions of these glibc libs.
>
>4) There is physically something wrong with your HD.
>
>5) You've got a rogue program/script which likes to make your life difficult.
>
>I find it curious that you are getting error messages for two *different*
>versions of glibc though. That may be indicative of a failed migration
>between v2.2 and v2.2.3. That would be where I would start my investigation.
>
>However, you have a bigger problem to deal with first. Logging in. RaQs
>don't have SASH (StandAlone Shell) so you can't (assuming you have console
>access) just log in using SASH and debug/repair the issue. 
>
>One option left to you, and I am not saying it is a nice one, is to use the
>restore cd's or download the restore 'package' -- I use that term loosely
>as I have never had good luck with it.
>
>A second somewhat less drastic option which is inevitably more
>time-consuming, is to pull the HD from the RaQ, mount it on a different
>machine and then use the second machines shell to debug/repair your HD,
>return the HD to your RaQ, and pray you got everything fixed or prepare to
>repeat ad nauseum. This does however assume that you have direct access and
>control of the console. If this isn't the case, be prepared to pay your
>support techs plenty of money unless you can tell them what exactly is
>wrong and thus avoid hours of debugging. If your RaQ is hosted by another
>individual/firm you may want to check with them when the last attempt to do
>any 'routine maintenance' happened. This wouldn't be the first time that
>someone dropped the ball on a glibc switch.
>
>
>Good luck!
>
>Brian
>
>------------------------------
>Brian Baxter
>Executive Director of Technologies
>GraffitiWorkz.com
>bbaxter@xxxxxxxxxxxxxxxxx
>------------------------------
>
>_______________________________________________
>cobalt-developers mailing list
>cobalt-developers@xxxxxxxxxxxxxxx
>http://list.cobalt.com/mailman/listinfo/cobalt-developers
>
>