Hi.

Our customer TecBan - A large Brazilian banking technology company (and controlled by other large private banks) - is migrating from Websphere to JBoss EAP.

The team in contact with both TAM (MW and OCP) is currently going to migrate around 20 applications.

In the local environment of the developer who tested the MTA CLI/Docker, the report generation is working without any problems.

Our services team (SA) preferred to migrate the report generation to OpenShift to decentralize the report generation from the hands of the developers in each area.

However, when trying to generate the report (in Microsoft Edge), the MTA UI keeps sending 3 requests in a loop and does not start analyzing/generating the report.

Primarily we suspect that it could be something related to the very strict security policy applied in the browsers.

The customer + TAM MW performed new tests in an environment with direct access to the OCP, but without success and with the same problem.

The same applications were used in the tests in the developer environment and in the OCP.

An application on GitHub [1] was also used without success in the OCP, while it worked in the developer environment.

The same application on GitHub [1] was used in the tests in the internal environment at Red Hat and it works.

HAR was requested during the tests in OCP to check if there is anything that can be analyzed from the point of view of the requests.

According to the HAR file [2], the `https://mta-openshift-mta.apps.osccorpdev.tecban.com/hub/tasks` requests are returning 200, but it is returning "error" in the content. It seems to be failing to start the image registry.redhat.io/mta/mta-analyzer-addon-rhel9, it is terminated with "Pod failed: Error". 

Question 1: Is it an expected error?

From the openshift-mta namespave inspect [3], there are two task pods, one failed and one still running for 2min. The task that failed is failing with the deps not found, however the analyzer-dep worked.

The entire installation of Operator 7.1 has now been redone and any references to cluster-wide objects created by it have been removed.

Question 2: What can be done or reviewed by the engineering team to help us to solve this problem at the customer environment?

At this moment, TAM team installed the MTA CLI in their environment (a RHEL machine) only to their team use for a while until we discovery what happen with the OCP and MTA in this case. 

Thanks.

[1] https://github.com/Azure-Samples/websphere-cafe

[2] https://access.redhat.com/support/cases/#/case/03956394/discussion?attachmentId=a09Hn00000YYcA9IAL

[3] https://access.redhat.com/support/cases/#/case/03956394/discussion?attachmentId=a09Hn00000YYiJ9IAL