Performing post-upgrade steps (introduced in v6.3.0)

Before performing the post-upgrade steps, ensure you have completed the upgrade steps.

Step 1: (Required) Run a full data load

The cache pods restart during the upgrade, which wipes the Ignite in-memory cache. Until a full loader run repopulates it, the trafficmanager readiness probe reports loaders NOTREADY, and the trafficmanager and configui rollouts don't complete.

Trigger the full-load CronJob manually, since it otherwise runs only once a day (at midnight):

kubectl create job --from=cronjob/loader-job-full loader-job-full-manual -n <namespace>
kubectl wait --for=condition=complete job/loader-job-full-manual -n <namespace> --timeout=300s

After the job completes, wait for the rollouts to finish:

kubectl rollout status deploy/trafficmanager-deploy -n <namespace>
kubectl rollout status deploy/configui-deploy -n <namespace>

Step 2: Verify the upgrade

Verify the Helm release using this command: helm list -n <namespace>. The new revision should show the chart apim-6.3.0 with status deployed.

Verify if the pre-upgrade hook got updated with these values:

Table	Column	6.2.x value	6.3.0 value
`service_provider`	`config` (JSON `targets` field)	`platformapi.<domainSuffix>:7000`	`platformapi-svc:7000`
`system_domain_endpoints`	`domain`	`platformapi.<domainSuffix>:7000`	`platformapi-svc:7000`

Also, confirm that the updated columns on migrated rows carry the upgrade timestamp.

Verify if all pods are running and ready on the 6.3.0 images:
```
kubectl get pods -n <namespace> \
  -o custom-columns="NAME:.metadata.name,READY:.status.containerStatuses[*].ready,IMAGES:.spec.containers[*].image"
```
note
If new pods show ImagePullBackOff, the 6.3.0 images aren't present, or apimImageRegistry or imagePullPolicy is incorrect.
Test an API via the trafficmanager service port and verify it returns a 200 or the expected response.

tip

If you are facing any issues upgrading to 6.3.0, consider rolling back to 6.2.x. Refer to Rollback 6.3.0 → 6.2.x for more information.