3.12. Installer Issues

This section describes known issues with the Installer component of Satellite 6.1.
BZ 1171841: ProxyAPI::ProxyException: ERF12-2749 [ProxyAPI::ProxyException]
Issue

If you install Satellite 6 with a signed CA certificate and then change to a self-signed certificate, it can cause issues with SSL and Foreman. This is the result of the previous certificates still existing in the /etc/pki/katello-certs directory.

Workaround

Refer to https://access.redhat.com/solutions/1311844 for information on how to work around this issue.

BZ 1230393: katello-installer saves the password in clear text and echoes it on subsequent runs
Issue

When you run katello-installer, it saves the password in plain text. Further, every time you run katello-installer successfully after that it is echoed to the console.

Workaround

To prevent this issue, change the password using the Satellite UI/API. The new password is not displayed as the installer output; only the old, invalid password is shown.

BZ 1122623: Installation fails if host puppet certificates have already been generated
Issue

If Puppet has been run at any time on a host prior to the installation of Satellite 6 (such as in an environment where the normal bootstrap process involves running Puppet), the host-specific certs are generated, but the CA certificate is not generated. When you run the Satellite 6 installation program, it attempts to generate the CA certificates but fails, because the host .PEM files already exist. The installation does not appear to error out in any way. The result is that the installation fails because HTTPD does not start. You might see the following error message: 

Could not start Service[httpd]: Execution of '/usr/share/katello-installer/modules/service_wait/bin/service-wait httpd start' returned 1: Starting httpd: [Tue Jul 22 12:49:31 2014] [warn] module passenger_module is already loaded, skipping
/Stage[main]/Foreman::Database/Foreman::Rake[db:seed]/Exec[foreman-rake-db:seed]: Failed to call refresh: /usr/sbin/foreman-rake db:seed returned 1 instead of one of [0]
Workaround

To work around this issue, use the following procedure:

  1. Delete or move all of the contents of /var/lib/puppet/ssl.
  2. Generate a puppet certificate for the host system: 
    # puppet cert generate $(hostname -f)
  3. Clear all subsequent back-end data stores: 
    # katello-installer --reset

Warning

The --reset option drops the entire database; all existing data will be lost. Only use this option prior to initial installation when there is no previous data accumulated. To preserve any existing data, do not use the --reset option.
BZ 1181283: error when changing fqdn and updating certificates
Issue

Attempting to change the FQDN and update the certificates on a Satellite 6 system currently does not work as expected, and as described in https://access.redhat.com/documentation/en-US/Red_Hat_Satellite/6.0/html/User_Guide/sect-Troubleshooting.html#idp27144192. Following this procedure does not correctly update the server certificates and results in an error.

Workaround

Updating the Satellite server host name is currently not supported, and consequently there is no workaround for this issue. This is scheduled to be addressed in an upcoming release.

BZ 1242264: Default configuration only supports up to 225 Content Hosts
Issue

The default Satellite 6 installation settings allow a maximum of 225 connected Content Hosts.

Workaround

See https://access.redhat.com/solutions/1375253 for instructions on how to configure your Satellite 6 deployment to support more Content Hosts.

BZ 1125414: Rerunning katello-installer still shows the "seed" password even if it has been changed
Issue

The first time you run katello-installer it issues a randomly-generated "seed" password. You can change this password from the web UI. However, if you run the katello-installer command again, it issues another random password, which does not work. The password change made in the web UI is required.

Workaround

There is no actual workaround for this issue. If you forget the password change that you made in the web UI you can reset the password from the command line, as follows: 

# foreman-rake permissions:reset
  Reset to user: admin, password: qwJxBpxb7Gfcjj5
BZ 11929256: The capsule-certs-generate command always suggests registering the capsule to ACME_Corporation
Issue

The capsule-certs-generate command always suggests using the "ACME_Corporation" organization with the subscription-manager command, whether that organization exists or not. For example: 

# capsule-certs-generate --capsule-fqdn mycapsule.example.com --certs-tar ~/mycapsule.example.com-certs.tar

# rpm -Uvh http://myserver.example.com/pub/katello-ca-consumer-latest.noarch.rpm
  subscription-manager register --org "ACME_Corporation"
  capsule-installer --parent-fqdn          "myserver.example.com"\
Workaround

Ensure that you replace "ACME_Corporation" with a valid organization.

BZ 1126759: ISO installer "--help" output should not contain "--enhanced_reporting" option
Issue

The ISO installer --help output contains the --enhanced_reporting option, which attempts to install the splice, ruby193-rubygem-splice_reports, and spacewalk-splice-tool packages. These packages do not exist on the ISO.

Workaround

None at this time.

BZ 1127307: Installation program fails if host name does not contain a top-level domain
Issue

If the host name does not contain a top-level domain, the katello-installer command fails with an error similar to the following: 

/Stage[main]/Foreman_proxy::Register/Foreman_smartproxy[sat6.test.example-domain]: Could not evaluate: 500 Internal Server Error
Installing             Done            [100%]
Workaround

Either update the host name to include a top-level domain, or pass a valid email address to the katello-installer command. For example: 

# katello-installer --foreman-admin-email=root@localhost.example.com
BZ 1221749: Puppet content promotion fails if there is no reverse DNS entry
Issue

If you create a content view with at least one module and attempt to promote it on a system that does not have a reverse DNS record, the action will fail with an error similar to the following: 

ERROR -- : Unable to resolve hostname for connecting client - 10.20.30.256. If it's to be a trusted host, ensure it has a reverse DNS entry.
Workaround

There is currently no workaround for this issue. Red Hat Satellite 6 requires that both forward and reverse DNS be configured and working correctly.

BZ 1221217: Rerunning installer can overwrite manual changes
Issue

The Satellite 6 installation program is based on Puppet, which means that any manual configuration changes might be overwritten if you run the installation program more than once.

Workaround

You can use the --noop argument when you run the installation program to determine what changes would be applied. This argument ensures that no actual changes are made. Potential changes are written to the /var/log/katello-installer.log file.

Files are always backed up and so you can revert any unwanted changes. For example, in the katello-installer logs you can see an entry similar to the following about Filebucket: 
/Stage[main]/Dhcp/File[/etc/dhcp/dhcpd.conf]: Filebucketed /etc/dhcp/dhcpd.conf to puppet with sum 622d9820b8e764ab124367c68f5fa3a1
You can restore the previous file as follows: 
puppet filebucket -l restore /etc/dhcp/dhcpd.conf 622d9820b8e764ab124367c68f5fa3a1
BZ 1226708: Running katello-remove does not remove puppet packages or directory
Issue

If you install katello and run the katello-installer installation program, and then use the katello-remove command, various packages and directories remain. This may cause issues with any future installations.

For example, the following files may remain: 
$ ls -la /etc/puppet
drwxr-xr-x.   3 root          root    4096 Jun  1 06:10 .
drwxr-xr-x. 105 root          root   12288 Jun  1 06:12 ..
-rw-r--r--.   1 root          root    4179 May 26 02:23 auth.conf.rpmsave
-rw-rw-r--.   1 foreman-proxy puppet     0 May 26 02:23 autosign.conf
-rw-r--r--    1 root          root    5529 Jun  1 03:49 client_cert.pem
-r--------    1 puppet        root    1679 Jun  1 03:49 client_key.pem
-r-xr-x---.   1 puppet        puppet 10641 May 26 02:23 node.rb
-rw-r--r--.   1 root          root    2456 May 26 02:23 puppet.conf.rpmsave
drwxr-xr-x.   4 puppet        root    4096 May 26 02:23 rack
-rw-r--r--    1 root          root    5469 Jun  1 03:49 ssl_ca.pem
The following package may also remain: 
$ rpm -qa | grep puppet
sat6home.example.com-puppet-client-1.0-1.noarch
BZ 1217821: System may not boot if Satellite services are configured incorrectly
Issue

If Satellite 6.1 services are configured to start at the wrong runlevel, it could result in the system failing to boot. Satellite 6.1 services require that networking be enabled, and so all Satellite 6.1 services should be off for runlevels 0, 1, 2, and 6, and on for runlevels 3, 4, and 5.

Workaround

Use the chkconfig command on Red Hat Enterprise Linux 6 or the systemctl command on Red Hat Enterprise Linux 7 to ensure that all Satellite services are enabled for the correct runlevels.