Troubleshoot the Service Manager Service Portal installation

This section provides troubleshooting information that can assist you in installing Service Manager Service Portal.

There are unfinished transactions remaining during glibc upgrade

Symptom:

The following message appeared in the TASK [propel-dependencies : Install Propel dependencies] section of the propel_prerequisites_installation.log file:

There are unfinished transactions remaining. You have an upgrade for glibc which is missing some dependency that another package requires.

Fix:

1. Run the following command to see what versions of glibc are already installed on the machine:

   yum list glibc

2. If the latest version of glibc is already installed, do the following:
   1. Open the /opt/hp/propel_complete_installer/ansible_content/roles/propel-dependencies/tasks/main.yml file.

   2. Comment out the following line:

      #- glibc.i686

The [Update Everything] task failed when attempting to download the package

Symptom:

The [Update Everything] task failed with the following error:

[Errno -1] Package does not match intended download.

Fix:

Clean up the rpm packages in the yum cache and then rerun the shell script (propel_install.sh).

HPE Operations Orchestration installation failed

Symptom:

HPE OO re-install is not successful when re-running the Ansible playbook.

Cause:

HPE OO installation failed and residual files and the central service are present on the Service Manager Service Portal host.

Fix:

1. Try running the HPE OO uninstaller:

   # cd /opt/hp/oo
   # ./uninstall --silent central
   # cd opt/hp/
   # rm -rf oo/

2. The central service might still be present:

   # systemctl status central
   # systemctl disable central
   # systemctl reset-failed central
   # rm -rf /etc/rc.d/init.d/central

OO configuration tasks fail due to proxy settings

Symptom:

OO configuration tasks that need to perform an HTTPS request fail.

Cause:

The tasks fail due to proxy settings.

Fix:

Add the following lines to ansible_content/roles/oo/main.yml for each of the tasks that need to perform an HTTPS request:

environment:
no_proxy: <myVM.domain>

These tasks include:

• name: "Setup the admin user in OO"

• name: "Enable authentication in OO"

• name: "Create deployment number in OO"

• name: "Upload the base content pack to OO"

• name: "Deploy base content pack in OO"

The following is an example:

- name: "Deploy base content pack in OO"
  # This takes a while (in the background)
  uri:
    url: https://{{ ansible_fqdn }}:8443/oo/rest/latest/deployments/{{ deployment.json.deploymentProcessId }}?force=false
    method: PUT
    body_format: json
    validate_certs: no
    user: admin
    password: changeit
    force_basic_auth: yes
    use_proxy: no
status_code: "200,204"
  environment:
     no_proxy: <myVM.domain>

Analytics service fails to start

Symptom:

The following error occurs:

Failed to start analytics.service: Unit analytics.service failed to load: No such file or directory.

Cause:

The analytics service failed to start.

Fix:

Run the following commands to solve this problem:

#cd /etc/init.d
#rm –f analytics
# cat /opt/hp/propel/etc/services.d/analytics.daemon.sh >analytics
#chmod 777 analytics
#service analytics start

Installing Service Manager Service Portal on a different disk partition fails to create hardlinks for services

Symptom:

When installing Service Manager Service Portal services, if the /opt/hp directory points to a system on a different disk partition than /etc, the setup.sh setup utility will fail to install Service Manager Service Portal services.

Cause:

The Service Manager Service Portal setup utility tries to create hardlinks between systemd services under /etc/systemd/system and Service Manager Service Portal service definition files under /opt/hp/propel/etc/system/system. However, Red Hat Enterprise Linux only supports hardlinks for files under the same disk partition (because they share inodes), hence, the failure to install Service Manager Service Portal services.

Fix:

1. Copy the files physically from /opt/hp/propel/etc/systemd/system to /etc/systemd/system:

   #cp -a /opt/hp/propel/etc/systemd/system/* /etc/systemd/system/

2. Reload the systemctl daemon:

   # systemctl daemon-reload

3. Enable the HPE Propel services:

   # systemctl enable <Services_Names>