IBM Support

Unable to Install the Analytics Engine powered by Apache Spark service on Watson Studio Local 2.1

Troubleshooting


Problem

You are trying to install the Analytics Engine powered by Apache Spark service on Watson Studio Local 2.1 by following document.

Symptom

After you run the installation command for spark service, you are getting error "one or more modules have failed".
image 5161

Cause

 After exec into cpd-install-operator pod, you list the releases. You notice 0010-infra is in failed state.
[root@abdwsl21-boot ibm]# kubectl get po -n wsl | grep operator
cpd-install-operator-7fdf4b695d-xn2tx                         1/1     Running     0          48d
[root@abdwsl21-boot ibm]# kubectl exec -it -n wsl cpd-install-operator-7fdf4b695d-xn2tx sh
sh-4.2$ helm ls --tls
NAME                            REVISION        UPDATED                         STATUS          CHART                                   APP VERSION     NAMESPACE
0010-infra                      5               Fri JUN 29 16:01:38 2020        FAILED        0010-infra-2.5.0                        2.5.0.0         wsl
0015-setup                      1               Fri May 29 16:06:38 2020        DEPLOYED        0015-setup-2.5.0                        2.5.0.0         wsl
0020-core                       1               Fri May 29 16:09:04 2020        DEPLOYED        0020-zen-base-2.5.0                     2.5.0.0         wsl

Environment

Watson Studio Local Version 2.1 on IBM Cloud Private 3.2.1

Diagnosing The Problem

You don't find spark installer jobs in the namespace you are trying to install the spark service

[root@abdwsl21-boot ibm]# kubectl get jobs -n wsl | grep spark

Resolving The Problem

We need to roll back the 0010-infra release with --no-hooks flag. 

kubectl get po -n <namespace> | grep operator
kubectl exec -it -n <namespace> <pod from above result> sh
helm rollback 0010-infra 1 --no-hooks --debug

After the rollback exit out of the container and start logging the operator pod to check what it is doing, we were looking for if it started to install spark.

If It might got stuck at 0010-infra as one of the jobs was not completing - zen-metastoredb-init, delete that job. 

kubectl delete job zen-metastoredb-init -n <namespace>

Then, log the operator pod again and see whether it starts installing spark, it starts installing spark.

Document Location

Worldwide

[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSHGWL","label":"IBM Watson Studio Local"},"ARM Category":[{"code":"a8m0z000000bmO0AAI","label":"Install->Doc"}],"ARM Case Number":"TS003923930","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"2.1.x","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Document Information

Modified date:
20 July 2020

UID

ibm16250485