Installation troubleshooting

Note

This page goes through common encountered problems during installation and how to deal with these.

For troubleshooting problems after installation, please refer to Operational troubleshooting

For release-specific bugs and their workarounds, please refer to Hotfixes / Known issues

Installation fails due to repo or meta data

Installation fails with a message that a repo or meta data could not be reached. As a result a package could not be installed

TASK [trinity/openldap : Install OpenLDAP packages] *****************************************************************************************************************************************
fatal: [controller1]: FAILED! => {"changed": false, "msg": "Failed to download metadata for repo 'plus': Cannot download repomd.xml: Cannot download repodata/repomd.xml: All mirrors were tried", "rc": 1, "results": []}

This tends to happen on occasion where we rely on external distribution repositories to be available. When one of the mirrors is temporarily unavailable, it causes the errors seen. A rerun of the playbook is the best approach; a simple retry.

TASK [trinity/chrony : Install chrony packages] *********************************************************************************************************************************************
fatal: [compute-rocky8.osimages.luna]: FAILED! => {"changed": false, "failures": [], "msg": "Unknown Error occurred: Could not run transaction.", "rc": 1, "results": []}

Same for the above. Just rerun the playbook again.

Could not bind to openldap

TASK [trinity/sssd : Adding access controler groups to the system] **********************************************************************************************************************
failed: [controller1] (item=admins) => {"ansible_loop_var": "item", "changed": true, "cmd": "/usr/local/sbin/obol group list | grep admins || /usr/local/sbin/obol group add admins", "delta": "0:00:00.195545", "end": "2025-02-10 13:25:41.418735", "item": "admins", "msg": "non-zero return code", "rc": 1, "start": "2025-02-10 13:25:41.223190", "stderr": "[ConnectionError] Failed binding to ldap\n[ConnectionError] Failed binding to ldap", "stderr_lines": ["[ConnectionError] Failed binding to ldap", "[ConnectionError] Failed binding to ldap"], "stdout": "", "stdout_lines": []}

This almost certainly is caused by having a previous openldap configuration in place where a certificate mismatch prevents connecting to the openldap backend. Make sure that there is no legacy in place while trying a complete (re)install of TrinityX.

Details: The symlink to an existing openldap is not present and will be placed. However no files/config is being regenerated. Also see the Installation notes

# ls -l /etc/openldap/
total 20
# the clashing certificates:
drwxr-xr-x. 2 root root 4096 Feb 11 01:44 certs
-rw-r--r--. 1 root root  121 Jul 26  2024 check_password.conf
-rw-r--r--. 1 root root 1545 Feb 11 01:45 ldap.conf
-rw-r--r--. 1 root root  900 Apr 30  2024 ldap.conf.ipabkp
drwxr-xr-x. 2 root root 4096 Feb 11 01:48 schema
# this link:
lrwxrwxrwx. 1 root root   35 Feb 11 01:44 slapd.d -> /trinity/local/etc/openldap/slapd.d

Ansible

When following the main branch, there may be updates to the variables in group_vars/all.yml.example which may not be incorporated with your group_vars/all.yml. A message such as the following may appear if this is the case:

# ansible-playbook controller.yml

PLAY [controllers] *****************************************************************************************************

TASK [Gathering Facts] *************************************************************************************************
fatal: [controller1]: FAILED! => {"msg": "The field 'environment' has an invalid value, which includes an undefined variable. The error was: 'trix_external_fqdn' is undefined. 'trix_external_fqdn' is undefined. 'trix_external_fqdn' is undefined. 'trix_external_fqdn' is undefined"}

To see what needs to be adjusted a diff can be done:

# diff group_vars/all.yml group_vars/all.yml.example

Please refer to Updating and Upgrading for more information.

Luna role fails when downloading iPXE kernels

TASK [trinity/luna2 : Download X86 64-bit iPXE EFI image to tftpboot] **********
FAILED - RETRYING: [controller1]: Download X86 64-bit iPXE EFI image to tftpboot (3 retries left).
FAILED - RETRYING: [controller1]: Download X86 64-bit iPXE EFI image to tftpboot (2 retries left).
FAILED - RETRYING: [controller1]: Download X86 64-bit iPXE EFI image to tftpboot (1 retries left).
fatal: [controller1]: FAILED! => {"attempts": 3, "changed": false, "dest": "/tftpboot/luna_ipxe.efi", "elapsed": 0, "msg": "Request failed", "response": "HTTP Error 404: Not Found", "status_code": 404, "url": "https://boot.ipxe.org/ipxe.efi"}

This is due to ipxe.org having changed the download locations.

Fix:

while being in the 'site' location, edit roles/trinity/defaults/main.yml
change the content to read the following for the ipxe_files variables:

luna2_aarch64_ipxe_files: "https://updates.clustervision.com/trinityx/{{ trix_version }}/additional/aarch64/ipxe"
luna2_x86_64_ipxe_files: "https://updates.clustervision.com/trinityx/{{ trix_version }}/additional/x86_64/ipxe"

write/quit
retry the playbook