Bug: parks-app crashes after being restored when hitting the api

[weshayutin]

Contact Details

whayutin@redhat.com

Describe bug

parks-app is OLD... afaict it's only value is mongodb.
https://github.com/openshift/oadp-operator/tree/master/tests/e2e/sample-applications/parks-app

Let's replace parks app w/ the same app we use for mariadb (todolist).
This will allow for data validation across db backends

We have the todolist app working with MongoDB.
https://github.com/weshayutin/todolist-mongo-go

What happened?

A bug happened!

OADP Version

1.0.1 (Stable)

OpenShift Version

4.10

Velero pod logs

No response

Restic pod logs

No response

Operator pod logs

No response

New issue

• This issue is new

[weshayutin]

@deepakraj1997 @hhpatel14 FYI

[deepakraj1997]

Further debug of this issue:

after i did a backup and restore, we have the same pods running initially but after i hit the parks REST API, it doesnt work, however, another deployment get triggered due to a change in build config: https://github.com/openshift/oadp-operator/blob/master/tests/e2e/sample-applications/parks-app/manifest4.8.yaml#L69-L77

After restore

Hitting the API

New Deployment trigger

After this it works.

[kaovilai]

I think the verification function should wait a few seconds and only check the last deployment (restify-2).

[kaovilai]

Whatever replacement is should remains a deploymentconfig with buildconfig which is a usecase that OADP currently requires a workaround on to restore successfully.

[weshayutin]

oh @kaovilai so just to make sure I 100% understand what you are saying...
Replace the container definition in the manifest and have the app built via an openshift build-config.
Is that correct? Happy to do it if so.

[kaovilai]

Yes. That would be great!