RE: you're in a maze of twisty passages, all alike

What I have done to get around the problems of up2date updating up2date in
the middle of doing other updates is to update up2date by itself on one
command and then use up2date to do everything else on a second command. I
don't do it every night, just once a week. 
 
I have been doing this manually, but hope to make it into a cronjob soon. My
problem with automating it is reading the 'script' log of activity. All of
these hash mark activity indicators look fine in real time but make it
difficult to read after the fact. Is there some way to get rid of the
hashmark output without getting rid of all of the activity reporting?

/Thomas Kern
/301-903-2211 
-----Original Message-----
From: Denis Hennessy [mailto:[email protected]] 
Sent: Thursday, September 25, 2003 11:46
To: [email protected]
Subject: you're in a maze of twisty passages, all alike


I'm so frustrated with trying to get my test machine up to date that I'm
about to give up and wait for the next CD build. I'm wondering if I'm
missing an easier route. Here's what's happening:

1. Install from beta2 CDs a few weeks ago. Install every package.
2. Run up2date-nox -u every night.

However, this only worked for a while. Then a variety of up2date problems
stop it working, with the result that I'm now many hundreds of megabytes out
of date. The cycle is now:


Run up2date-nox -u. This will fail either at some dependency check or later
during the download (it seems to loop getting the same package). 
'Fix' the problem. Right now I simply rpm -e the problem package. 
Try again. This pretty much always fails with the 'abuse of service'
message. The one that says it can't appear with normal use... Retrying has
no effect since every use is blocked with an abuse failure.

All I'm trying to do is keep a machine up to date with the current full list
of packages. Is there a simpler way?

/dh