Kickstart from satellite server version 5.6 fails for missing rpm (only @Base used in %packages of kickstart file)

Latest response

I was curious if anyone has seen a kickstart fail for gzip-1.3.12-19.el_4.x86_64.rpm being missing?
(rhel 6.5 server kickstart, 64-bit, satellite server version 5.6 running on rhel6.5 with selinux, more details below) (file is on sat server, shown through locate command, see below)

As a brief test, we included the --ignoremissing in the kickstart (but really want to find why it is failing).
We only have @Base in the %packages area of the kickstart file.

  • The system enters the kickstart fine, gets past it's partitioning, and starts the installation. It goes to about 64% of installing rpms and then hangs for a while and gives the following error:

============ begin quote of error

The file gzip=1.3.12-19.el6.x86_64.rpm cannot 
be opened.  This is due to a missing file, a 
corrupt package or ccorrupt media.  Please 
verify your installation source.

if you exit, your system will be left in an 
inconsistent state that will likely require 
reinstallation
[Reboot] [Retry]

=============== end quote of error

We have a kickstart in our satellite server for rhel6 server that fails for the above rpm after about 64% into the installation.

  • This is a relatively new satellite server. We dumped our old satellite 5.5 server on one specific network and replaced it with this new 5.6 satellite server.
  • We are running in disconnected mode, there is no chance of visibility to redhat.com for that satellite server
  • The satellite server is running RHEL 6.5, satellite version 5.6 with selinux enforcing. (we tried the kicsktart with the satellite server in permissive as well, no change)
  • We set the kickstart for a base channel of RHEL 6, 64-bit.
  • The kickstart "Available Trees" is set for ks-rhel-x86_64-server-6-5.5
  • There is only one "Child Channels" selected, "rhn-tools-rhel-x86_64-server-6
  • Base and incremental channels (iso dumps downloaded from access.redhat.com) properly ingested (we've been doing this for years) via base and incremental channel iso dumps downloaded directly from redhat.com
  • We verified all base and incremental channel isos with md5sums before proceeding
  • Just today, we reingested all base then incremental channels from the unpacked iso files. (we are highly familiar with this procedure)
    (including that one new java channel Red Hat pushed to the side, separate discussion)
  • We also tried the directive "cdrom" in the kickstart to attempt to bypass this issue and install from the cdrom, and this made no difference.
  • There is only one item in %packages section within the kickstart -> @Base
  • The file gzip=1.3.12-19.el6.x86_64.rpm is on the satellite server, but in the web gui, it says the file is "not available"
  • The file gzip=1.3.12-19.el6.x86_64.rpm is shown with the following command:
  # updatedb ; locate gzip | grep 1.3.12-19 | grep rpm$
  /var/satellite/redhat/NULL2ea/gzip/1.3.12-19.el6_4/x86_64/[VERY-LONG-STRING-HERE]/gzip=1.3.12-19.el6.x86_64.rpm

We are going to unpack our base channel isos one at a time into a directory and then reingest the base channels once more just in case one of the rsyncs failed (but then I wonder why would the file actually exists with the locate command listed earlier...?) Perhaps the file never linked... One of the last things done in a satellite-sync is the linking of files.

Some of the causes I've seen in the past for channel/rpm issues is sometimes resolved through reingesting the files.

Appreciate any help.

RJ Hinton's picture

Responses