Error
While uninstalling the IBM Operations Dashboard capability from the Cloud Pak for Integration the NGINX pods from Foundational Services started to fail with the following error:
2021/12/01 17:35:38 [emerg] 33#33: host not found in upstream "ops-dashboard-ibm-integration-operations-dashboard-fe-ui-svc.cp4i.svc" in /user-home/_global_/nginx-conf.d/ops-dashboard-ibm-integration-operations-dashboard-fe-zen--7707-cp4i.conf:4
nginx: [emerg] host not found in upstream "ops-dashboard-ibm-integration-operations-dashboard-fe-ui-svc.cp4i.svc" in /user-home/_global_/nginx-conf.d/ops-dashboard-ibm-integration-operations-dashboard-fe-zen--7707-cp4i.conf:4Solution
The easiest way to fix this issue is to make sure that the configuration that is now missing is removed from the NGINX config file by deleting in the pod.
Since the pods are in a CrashLoopBackOff state you won’t be able to do a simple oc exec -it ibm-nginx-f98c84888-77pgb -n cp4i -- bash to do, but instead, you should start a debug pod and remove the entry.
Here is what I did:
$ oc debug pod ibm-nginx-f98c84888-77pgb
$ cd /user-home/_global_/nginx-conf.d/
$ rm ops-dashboard-ibm-integration-operations-dashboard-fe-zen--7707-cp4i.conf
$ exit
$ oc delete pods -n cp4i -l app.kubernetes.io/component=ibm-nginx