jordigilh commented on issue #379:
URL:
https://github.com/apache/incubator-kie-kogito-serverless-operator/issues/379#issuecomment-1926817631
> **For the record and subsequent fix:**
>
> Ok, finally I cloud get the problem.
>
> First, the tests executed one by one, or by commenting the
DescribeTable("when deploying a SonataFlow CR with PostgreSQL persistence")
works fine.
>
> So what's the problem?
>
> 1. When the tests are executed the one related with persistence is
picked up first as expected, the workflow name is callbackstatetimeouts, and
for this particular execution we are using a customized platform that
introduces the persistence related addon, which is good for that test.
> buildArgs:
>
> * name: QUARKUS_EXTENSIONS
> value:
org.kie.kogito:kogito-addons-quarkus-jobs-knative-eventing:999-SNAPSHOT,org.kie.kogito:kogito-addons-quarkus-persistence-jdbc:999-SNAPSHOT,io.quarkus:quarkus-jdbc-postgresql:3.2.9.Final,io.quarkus:quarkus-agroal:3.2.9.Final
>
> 2. A corresponding image with the name
10.104.85.50/callbackstatetimeouts is produced, and the workflow with that
image. All good, the test finalizes fine.
>
> 3. In a later point in time, another test Prod profile test, but with
NO persistence configured executes for the same workflow.
> I.e., with the same id. callbackstatetiemouts (In fact
>
>
> Same callbackstatetimeouts workflow (or with the same id) is built, and
thus, the same image name is produced 10.104.85.50/callbackstatetimeouts , this
image that is produced now, and this usecase is not prepared for persitence.
All good.
>
> However, this build is not overwritting the already existing image with
the same name 10.104.85.50/callbackstatetimeouts , and this previous image was
built with the persistence addons. So this time, when the workflow starts the
runtime complains about the missing jdbc url, since the image was build with
the persistence addon but the current test is not configuring the WF to use
persistence.
>
> A piece of investigation is needed to see
>
> 1. maybe a better images naming for the kaniko builder case
> Note that it could happen that a workflow my_workflow can be
deployed in namespace1 with persistence and in namespace2 without, etc.
>
> 2. we need to see what happens when we are in openshift too
>
> 3. any other variant to support the situation above.
Thanks so much for digging up the issue! and also it's a great thing we were
able to catch this during e2e testing and not during a customer deployment :)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
