[ale] SOLVED Re: Slackware 12.0 installation scripts

[David S Jackson]

Okay, just to close this out...

I wrote a script to go into the slackware/ directories and test
installation targets on the target box.  If the install crapped out then
the target file has a zero length.  This script basically told me where
the install ran out of space.  After clearing more space, I restarted
the install with

upgradepkg --reinstall --install-new /packages/slackware/$f/*.tgz

where $f is the last good directory in a list of subsequent pkg
directories.

It worked out and that box is now purring along on Slack 12, awaiting
the next upgrade.

I will note that /sbin/ldconfig was complaining all over the place about
broken libraries.  This was confusing me, but as soon as the reinstall
did its work, those warnings went away.  

Okay, that's it. Thanks for your time!

Dave 


On Mon, 2014-10-06 at 15:13 -0400, Boris Borisov wrote:
> What about backup first. Do you have a lot of customisation of the
> slack.
> 
> On Monday, October 6, 2014, David S Jackson <deepbsd.ale at gmail.com>
> wrote:
>         Hi,
>         
>         I have a pet Slackware box (athlon 3.6 now) that I've been
>         nursing along
>         recently. It was last upgraded in 2005 (Slack 10.0 I think).
>         I've had
>         to replace/rebuild a bunch of things, but finally (up through
>         Slack
>         11.0), I was making the jump from Slack 11.0 to Slack 12.0,
>         which was
>         the first 2.6 kernel, pretty big xorg change, etc etc.  And I
>         had a
>         problem.  The /usr partition filled up sometime during the
>         installation
>         (old 80G drive, 9G /usr partition).
>         
>         I don't really know at what point the partition filled up, but
>         the
>         installation scripts just kept chugging away like nothing was
>         wrong.
>         Kinda weird.   (I was checking on another tty.)  I know I had
>         gotten
>         through the first few directories, so we might have been into
>         the x and
>         xap series, perhaps.  But I had rebooted a new 2.6.18 smp huge
>         kernel
>         prior to this part of the upgrade, so at least that was
>         working.
>         
>         So I fixed the /usr problem (housekeeping in /usr/local).  I
>         fixed the
>         *.new scripts in /etc.  But now I don't know how to check for
>         what's
>         still broke and what ain't before the reboot.  And I'd like to
>         see what
>         didn't install and what did.  I reran the upgradepkg
>         --install-new
>         scripts and they didn't find anything not installed.  But I
>         question
>         them.  Those scripts apparently weren't smart enough to stop
>         when they
>         were installing onto a full /usr partition, so how can I trust
>         them to
>         know whether they were successful or not?  Can a package
>         appear to be
>         installed in /var/log/packages yet not really be completely
>         installed
>         properly on a Slackware box?  I'd like to know before I
>         reboot.  :-)
>         
>         For all you Slackware guru's and others,, what's the best way
>         to
>         approach this problem?
>         
>         Thanks in advance!
>         
>         Dave
>         
>         
>         
>         
>         _______________________________________________
>         Ale mailing list
>         Ale at ale.org
>         http://mail.ale.org/mailman/listinfo/ale
>         See JOBS, ANNOUNCE and SCHOOLS lists at
>         http://mail.ale.org/mailman/listinfo
> 
> 
> -- 
> Sent from Gmail Mobile
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo