osu-itis/ansible-github-enterprise

Ansible playbook for GitHub Enterprise

ansible-github-enterprise's Issues

Add "no_log" to set production state in zenoss task

no_log: true is missing from, and needs to be added to, the "set production state in zenoss" task.

Add support for rebooting after upgrade

Certain hotpatch upgrades require a reboot of the server. Add an option to force a reboot after the upgrade is complete.

Fix "waiting for server restart" timeouts

Recent feature releases (3.6, 3.7, and in particular, 3.8) now take far longer to upgrade than ever before. A few timeouts need to be increased to accommodate.

Parameterize zenoss and vmware steps

This playbook currently updates maintenance status in Zenoss and takes a snapshot of the VM in VMware whether one needs it or not. It would be nice to pull those steps out and make them optional for those who don't use Zenoss or VMware in their infrastructures.

Check for configuration run in progress

Upgrades to 2.14 fail with the following error:

TASK [upgrade-ghe : disable maintenance mode in GHE] ***************************
fatal: [github-stage.someplace.edu]: FAILED! => {"changed": false, "cmd": ["ghe-maintenance", "-u"], "delta": "0:00:00.096351", "end": "2018-07-12 17:29:12.469892", "msg": "non-zero return code", "rc": 1, "start": "2018-07-12 17:29:12.373541", "stderr": "", "stderr_lines": [], "stdout": "A configuration run is currently in progress. Unable to unset maintenance mode. Please try again after the configuration run is completed.", "stdout_lines": ["A configuration run is currently in progress. Unable to unset maintenance mode. Please try again after the configuration run is completed."]}
	to retry, use: --limit @/opt/ansible/ansible-github-enterprise/upgrade_ghe.retry

Add a check before this task that determines if the configuration run is complete or not before proceeding.

Do not fail if GHE is already at the latest version

If this playbook is run on a GHE instance that is already at the latest version, the playbook run results in a hard failure. We should check for this condition and handle it gracefully.

Add support for GitHub Enterprise 2.21 ghe-upgrade-check changes

In GitHub Enterprise 2.21, the ghe-upgrade-check command now returns two different strings depending on if the -i flag is given. This causes the current playbook to handle the version check incorrectly and attempt to upgrade even if GHE is already running the latest version.

Update vmware module for python 3

Update zenoss module for python 3

Make GHE sign-in page the default when checking if the application has come back up

This playbook checks for our CAS auth login page to determine if the GHE application has come back up properly after upgrading. Let's make the GHE sign-in page the default instead, and allow users to customize the string to search for.

Warnings thrown when running playbook with ansible 2.3.x

A few warnings are now being thrown when running this playbook in ansible 2.3.x.

TASK [upgrade-ghe : verify ghe-update-check results] ***************************
 [WARNING]: when statements should not include jinja2 templating delimiters

TASK [upgrade-ghe : set maintenance state in zenoss] ***************************
 [WARNING]: Module did not set no_log for password

etc.

Whatever ansible is complaining about here should probably be fixed.

Remove root environment workaround

The issue where ghe-update-check did not honor the system proxy setting was fixed in 2.6.4, so the workaround that preserved the root environment via a modified version of the script (ghe-update-check-OSU) can be removed.

Support hotpatches for patch release upgrades

https://help.github.com/enterprise/2.11/admin/guides/installation/upgrading-github-enterprise/#upgrading-a-single-appliance-using-a-hotpatch