Troubleshooting Guide | Hybrid Cloud Patterns

Understanding the Makefile

The Makefile is the entrypoint for the pattern. We use the Makefile to bootstrap the pattern to the cluster. After the initial bootstrapping of the pattern, the Makefile isn’t required for ongoing operations but can often be useful when needing to make a change to a config within the pattern by running a make upgrade which allows us to refresh the bootstrap resources without having to tear down the pattern or cluster.

make install / make deploy

Executing make install within the pattern application will trigger a make deploy from <pattern_directory>/common. This initializes the common components of the pattern framework and will install a helm chart in the default namespace. At this point cluster services such as Red Hat Advanced Cluster Management and OpenShift Gitops are deployed.

Once common completes, the remaining tasks within the make install target will execute.

make vault-init / make load-secrets

This pattern is integrated with HashiCorp Vault and External Secrets services for secrets management within the cluster. These targets install vault from a Helm chart and the load the secret (values-secret.yaml) you created during Getting Started.

If values-secret.yaml does not exist, make will exit with an error saying so. Furthermore, if the values-secret.yaml file does exist but is improperly formatted, ansible will exit with an error about being improperly formatted. If you are not sure how format the secret, please refer to Getting Started.

make bootstrap / make upgrade

make bootstrap is the target used for deploying the application specific components of the pattern. It is the final step in the initial make install target. Running make bootstrap directly should typically not be necessary, instead you are encouraged to run make upgrade.

Generally, executing make upgrade should only be required when something goes wrong with the application pattern deployment. For instance, if a value was missed, and the chart wasn’t rendered correctly, executing make upgrade after fixing the value would be necessary.

If you have any further questions, please, feel free to review the Makefile for the common and Medical Diagnosis components. It is located in common/Makefile and ./Makefile respectively.

Troubleshooting the Pattern Deployment

Occasionally the pattern will encounter issues during the deployment. This can happen for any number of reasons, but most often it is because of either a change within the operator itself or something has changed/happened to the Operator Lifecycle Manager (OLM) which determines which operators are available in the operator catalog. Generally, when an issue occurs with the OLM, the operator is unavailable for installation. To ensure that the operator is in the catalog:

oc get packagemanifests | grep <operator-name>

When an issue occurs with the operator itself you can verify the status of the subscription and make sure that there are no warnings.An additional option is to log into the OpenShift Console, click on Operators, and check the status of the operator.

Other issues encounter could be with a specific application within the pattern misbehaving. Most of the pattern is deployed into the xraylab-1 namespace. Other components like ODF are deployed into openshift-storage and the OpenShift Serverless Operators are deployed into knative-serving, knative-eventing namespaces.

Use the grafana dashboard to assist with debugging and identifying the issue

Problem: No information is being processed in the dashboard

Solution: Most often this is due to the image-generator deploymentConfig needing to be scaled up. The image-generator by design is scaled to 0;

oc scale -n xraylab-1 dc/image-generator --replicas=1

Or open the openshift-console, click on workloads, then click deploymentConfigs, click image-generator, and scale the pod to 1 or more.

Problem: When browsing to the xraylab grafana dashboard and there are no images in the right-pane, only a security warning.

Solution: The certificates for the openshift cluster are untrusted by your system. The easiest way to solve this is to open a browser and go to the s3-rgw route (oc get route -n openshift-storage), then acknowledge and accept the security warning.

Problem: In the dashboard interface, no metrics data is available.

Solution: There is likely something wrong with the Prometheus DataSource for the grafana dashboard. You can check the status of the datasource by executing the following:

oc get grafanadatasources -n xraylab-1

Ensure that the prometheus datasource exists and that the status is available. This could potentially be the token from the service account (grafana-serviceaccount) that is provided to the datasource as a bearer token.

Problem: The dashboard is showing red in the corners of the dashboard panes.

Solution: This is most likely due to the xraylab database not being available or misconfigured. Please check the database and ensure that it is functioning properly.

Step 1: Ensure that the database is populated with the correct tables:


oc exec -it xraylabdb-1-<uuid> bash

mysql -u root

USE xraylabdb;

SHOW tables;

The expected output is:


Welcome to the MariaDB monitor.  Commands end with ; or \g.
Your MariaDB connection id is 75
Server version: 10.3.32-MariaDB MariaDB Server

Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.

MariaDB [(none)]> USE xraylabdb;
Database changed
MariaDB [xraylabdb]> show tables;
+---------------------+
| Tables_in_xraylabdb |
+---------------------+
| images_anonymized   |
| images_processed    |
| images_uploaded     |
+---------------------+
3 rows in set (0.000 sec)

Step 2: Verify the password set in the values-secret.yaml is working

oc exec -it xraylabdb-1-<uuid> bash

mysql -u xraylab -D xraylabdb -h xraylabdb -p
<provide your password at the prompt>

If you are able to successfully login then your password has been configured correctly in vault, the external secrets operator and mounted to the database correctly.

Problem: The image-generator is scaled correctly, but nothing is happening in the dashboard.

Solution: This could be that the serverless eventing function isn’t picking up the notifications from ODF and therefore, not triggering the knative-serving function to scale up. In this situation there are a number of things to check, the first thing is to check the logs of the rook-ceph-rgw-ocs-storagecluster-cephobjectstore-a-<podGUID> pod in the openshift-storage namespace.

oc logs -n openshift-storage -f <pod> -c rgw

You should see the PUT statement with a status code of 200

Next ensure that the kafkasource and kafkservice and kafka topic resources have been created:

oc get -n xraylab-1 kafkasource

NAME          TOPICS            BOOTSTRAPSERVERS                                      READY   REASON   AGE
xray-images   ["xray-images"]   ["xray-cluster-kafka-bootstrap.xraylab-1.svc:9092"]   True             23m

oc get -n xraylab-1 kservice
NAME              URL                                                  LATESTCREATED           LATESTREADY             READY   REASON
risk-assessment   https://risk-assessment-xraylab-1.apps.<SUBDOMAIN>   risk-assessment-00001   risk-assessment-00001   True

oc get -n xraylab-1 kafkatopics
NAME                                                                                               CLUSTER        PARTITIONS   REPLICATION FACTOR   READY
consumer-offsets---84e7a678d08f4bd226872e5cdd4eb527fadc1c6a                                        xray-cluster   50           1                    True
strimzi-store-topic---effb8e3e057afce1ecf67c3f5d8e4e3ff177fc55                                     xray-cluster   1            3                    True
strimzi-topic-operator-kstreams-topic-store-changelog---b75e702040b99be8a9263134de3507fc0cc4017b   xray-cluster   1            1                    True
xray-images                                                                                        xray-cluster   1            1                    True