Remove deployment updates from autodetect loop #869
Conversation
Signed-off-by: Ruben Vargas <[email protected]>
rubenvp8510
force-pushed
the
Issue-855
branch
from
5f62402 to
f014b52
Compare
There was a problem hiding this comment.
LGTM. Just need one clarification before we merge it:
because for some reason (I'm still not sure why) the reconciliation run on each deployment at bootstrap time (see logs). Not sure when this behaviour changed.
It would be good to check the changelogs and/or ask the Operator SDK team (Slack or mailing list) about this. Otherwise, we might end up depending on a behavior that might be a bug, to be fixed in a future version.
|
@objectiser @jpkrohling @rubenvp8510 What is the status of this? E2E tests are broken on OpenShift without this. |
|
@rubenvp8510 please, find evidence that this new behavior is the intended one, rather than a bug that has been introduced in the latest Operator SDK versions. |
|
I asked on kubernetes-operator slack channel.
|
There was a problem hiding this comment.
@jpkrohling As the question seems to have been answered, and you previously indicated LGTM (subject to that answer) I will merge.
Signed-off-by: Ruben Vargas [email protected]
In #454 we introduced the detection/update of deployments to address #442, But after seen this closely, I think this is not the right approach. We should run the scan only at start.
More than that, In my latest tests debugging for fix #855. I see that we don't need to run the scan any more (operator-sdk 0.12.0) , because for some reason (I'm still not sure why) the reconciliation run on each deployment at bootstrap time (see logs). Not sure when this behaviour changed.
bootstrap-operator.txt