[rhn-users] fatal unpack error during up2date -u

Jim Pelton Jim.Pelton at noaa.gov
Tue May 24 16:28:31 UTC 2005


Christoph,

    Thanks for the reply, however the problem is not that the rpm 
database and package profile on the satellite are un-synced.  We did do 
an up2date -p and up2date -l but with no resolve.  The problem is that 
the packages that where not installed, but only downloaded, completely 
disappeared from the rpm database on the client system!  The database 
does not seem corrupted, but incomplete or rather just missing about 30 
entries.  Again it is sinked with the package profile on the satellite 
system.  So I guess my question is, does up2date remove packages from 
the client rpm database as soon as they are downloaded to the client 
system, then re-add them after/during rpm installation?

Thanks again!

--Jim

Doerbeck, Christoph wrote:

>You might try "up2date -p" which (I think) should refresh the system 
>profile on the  rhn server.  Follow-up with a "up2date --list" to see 
>if the missing packages are now flagged as requiring updates.
>
>What version of RHN proxy are you using?
>
>Christoph
>
>-----Original Message-----
>From: Jim Pelton [mailto:Jim.Pelton at noaa.gov] 
>Sent: Monday, May 23, 2005 6:33 PM
>To: rhn-users at redhat.com
>Subject: [rhn-users] fatal unpack error during up2date -u
>
>
>Hello List!
>
>Here's a tricky one for ya!  We've got a RHN proxy server, through which
>
>we communicate with a RHN satellite server.  Now the proxy is working 
>great.  It seems to be communicating correctly with the satellite server
>
>and downloading rpm patches to its database.  I have a test machine 
>running which applies updates downloaded from the proxy via up2date.
>
>The issue is this:  we ran up2date -u (from the command line) on a 
>client machine and during the install phase of the update, the process 
>died completely.  This error was issued: "There was a rpm unpack error 
>installing the package: gnome-applets-2.2.2-2.1E."  Normally one would 
>restart the process, assuming that it would pick up where it left off.  
>RPM is supposed to be quite robust correct?  Well, up2date -u (as well 
>as up2date -l) claims that the system is up to date but something like 
>30 packages where not installed--they where d/led however as they where 
>confirmed in /var/spool/up2date.  If we log onto the satellite server 
>the system is claimed to be up2date there as well.  The output of "rpm 
>-qa | nl" shows 507 packages, the same as the satellite server.  The 
>interesting thing is that if we "grep" the output of "rpm -qa" with a 
>package name that was not installed before the fatal rpm unpack error, 
>not even the old package is shown as being installed!
>
>We wiped out /var/spool/up2date in an effort to "trick" up2date into 
>downloading and installing the packages again but we now know that 
>up2date does not even look there for confirmation of what it has
>downloaded.
>So what could the problem be?
>
>Thanks is advance, I understand this is probably more of an RPM issue 
>but since it does involve a lot of RHN stuff I thought I'd post here
>anyway.
>
>--Jim
>
>_______________________________________________
>rhn-users mailing list
>rhn-users at redhat.com
>https://www.redhat.com/mailman/listinfo/rhn-users
>
>_______________________________________________
>rhn-users mailing list
>rhn-users at redhat.com
>https://www.redhat.com/mailman/listinfo/rhn-users
>  
>




More information about the rhn-users mailing list