tiwillia / openshift_lab_automation Goto Github PK

When creating or editing virtual machines, all images from the OpenStack backend are included in the list. The images that are not available to the new tenants should not be shown in the list.

Problem Recreation Steps
-Instance is Deployed
-While Deployed it is removed from project
-Add new instance with the same IP
-Deploy instance
-Instance never spins up

I was able to see the old instance was still up in the Open Stack environment.

FIx:
I would think adding the undeploy instance call when removing from project

I would be good if we could add some color or highlighting to what items are deployed / un-deployed!
(example: green when deployed | blue when building | red when un-deployed, etc)
It would also be nice to see a live update on the status check (example: flash green when check completes | change background to yellow when checking | change back ground to orange when check fails)

If the root password, hostname, domain, or gear size of an instance includes a ':', cloud-init will fail to shellify.

There are other instances where this may also occur. Any variable in the cloud init data in app/models/instance.rb that could contain and illegal cloud-init character should be validated before submitting cloud-init data.

Seemingly randomly, the asset precompilation will fail when pushing to openshift:

remote: /opt/rh/ruby193/root/usr/bin/ruby /var/lib/openshift/54c16f893b696e355c000963/app-root/runtime/repo/vendor/bundle/ruby/1.9.1/bin/rake assets:precompile:all RAILS_ENV=production RAILS_GROUPS=assets
remote: rake aborted!
remote: Command failed with status (): [/opt/rh/ruby193/root/usr/bin/ruby /var/lib...]
remote: /var/lib/openshift/54c16f893b696e355c000963/app-root/runtime/repo/vendor/bundle/ruby/1.9.1/gems/actionpack-3.2.16/lib/sprockets/assets.rake:12:in `ruby_rake_task'
remote: /var/lib/openshift/54c16f893b696e355c000963/app-root/runtime/repo/vendor/bundle/ruby/1.9.1/gems/actionpack-3.2.16/lib/sprockets/assets.rake:21:in `invoke_or_reboot_rake_task'
remote: /var/lib/openshift/54c16f893b696e355c000963/app-root/runtime/repo/vendor/bundle/ruby/1.9.1/gems/actionpack-3.2.16/lib/sprockets/assets.rake:29:in `block (2 levels) in <top (required)>'
remote: Tasks: TOP => assets:precompile
remote: (See full trace by running task with --trace)
remote: -------------------------
remote: Git Post-Receive Result: failure
remote: Activation status: failure
remote: Activation failed for the following gears:
remote: 54c16f893b696e355c000963 (Error activating gear: CLIENT_ERROR: Failed to execute: 'control deploy' for /var/lib/openshift/54c16f893b696e355c000963/ruby)
remote: Deployment completed with status: failure
remote: postreceive failed

When a user un-checks out an environment, a dialog should appear that will ask if the environment should be re-deployed or not. If the user answers yes, the environment should then be re-deployed.

Have not investigated into how possible this is.

Instead of creating a new connection to the openstack backend in every method where it is needed, find a way to cache or store the necessary connection information to avoid authenticating so often.

If an openstack server is overseas, it could add quite a bit of time to authenticate on just about every action.

When an instance is removed, it is currently not removed on the backend.