Resources are recreated when provider Pod is restarted

[irizzant]

Hi again!
I'm using the latest Provider version 0.17.0 with the latest Crossplane release.
I noticed that when the provider Pod is restarted, some managed resources which where previously created and in Ready status are considered to recreate.
This leads to the situation where the managed resouces already exists and it errors out, and it makes our team to go hunting for them and manually fix them.
Any idea?

[irizzant]

Maybe related to crossplane/crossplane#5918 and also to #137, see this Slack thread also

[irizzant]

It looks like the problem was related to crossplane/crossplane#5918 and a missing LateInitialize in managementPolicies of the TeamMembership resource.

This in turn caused TeamMembership resources to be queued endlessly, causing very long workqueue and high load.

After I add the LateInitialize back, I noticed the workqueue became empty and became much more stable