To help troubleshoot and resolve installation issues, you can run a diagnostic script that checks your deployment for potential problems.
Important: Do not use the script before the installation process completes. Despite a successful installation message, some processes might still need to complete, and it can take up to 10 minutes before IBM Event Streams is available to use.
To run the script:
- Download the
installation-diagnostic-script.shscript from GitHub.
- Ensure you have installed the Kubernetes command line tool and the IBM Cloud Private CLI as noted in the installation prerequisites.
- Log in to your cluster as an administrator by using the IBM Cloud Private CLI:
cloudctl login -a https://<Cluster Master Host>:<Cluster Master API Port>
The master host and port for your cluster are set during the installation of IBM Cloud Private.
- Run the script as follows:
./installation-diagnostic-script.sh -n <namespace> -r <release-name>
If you have been waiting for more than an hour, add the
--restartoldpods option to recreate lost events (by default, events are deleted after an hour). This option restarts Failed or Pending pods that are an hour old or more.
For example, the following installation has pods that are in pending state, and running the diagnostic script reveals that the issue is caused by not having sufficient memory and CPU resources available to the pods:
Starting release diagnostics... Checking kafka-sts pods... kafka-sts pods found Checking zookeeper-sts pods... zookeeper-sts pods found Checking the ibm-es-iam-secret API Key... API Key found Checking for Pending pods... Pending pods found, checking pod for failed events... ------------------ Name: caesar-ibm-es-kafka-sts-0 Type: Pod FirstSeen: 2m LastSeen: 1m Issue: 0/4 nodes are available: 1 Insufficient memory, 4 Insufficient cpu. ------------------ Name: caesar-ibm-es-kafka-sts-1 Type: Pod FirstSeen: 2m LastSeen: 1m Issue: 0/4 nodes are available: 1 Insufficient memory, 4 Insufficient cpu. ------------------ Name: caesar-ibm-es-kafka-sts-2 Type: Pod FirstSeen: 2m LastSeen: 1m Issue: 0/4 nodes are available: 1 Insufficient memory, 4 Insufficient cpu. ------------------ No failed events found for pod caesar-ibm-es-rest-deploy-6ff498d779-stf79 ------------------ Checking for CrashLoopBackOff pods... No CrashLoopBackOff pods found Release diagnostics complete. Please review output to identify potential problems. If unable to identify or fix problems, please contact support.