Enable Semeru Cloud Compiler for InstantOn

[leochr]

What this PR does / why we need it?:

• PR for Enable Semeru Cloud Compiler for InstantOn #438

Does this PR introduce a user-facing change?

• User guide
• CHANGELOG.md

Which issue(s) this PR fixes:

Fixes #

[tjwatson]

@anjumfatima90 can you try out the operator with InstantOn?

[tjwatson]

Fixes #438

[leochr]

We didn't feel comfortable releasing something that wasn't tested. So this enhancement is only available from a pre-release driver. Follow the instructions in our wiki to install the driver: https://github.com/OpenLiberty/open-liberty-operator/wiki

Adding the CatalogSource details here as well (for Step 2.2) in case our Wiki is changed to some other driver:

apiVersion: operators.coreos.com/v1alpha1
kind: CatalogSource
metadata:
  name: olo-catalog
  namespace: openshift-marketplace
spec:
  displayName: Open Liberty instanton
  image: 'icr.io/appcafe/open-liberty-operator-catalog@sha256:5298bed424f87f480cb0756d3dbf4cc309af1b0c9459034c814dbcc55680b808'
  sourceType: grpc

[tjwatson]

@NottyCode points out that this handles OPENJ9_JAVA_OPTIONS and OPENJ9_RESTORE_JAVA_OPTIONS values being set in kubernetes config but it doesn't handle cases where the container image has already set the env. For example, using ENV in their Dockerfile. Is this a concern? If so it has been an issue ever since the operator started auto setting the OPENJ9_JAVA_OPTIONS

[leochr]

It's not a concern. The commands will be executed at runtime inside the container. So the jitServerOptions will be appended to OPENJ9_JAVA_OPTIONS value (the value could have come from container ENV or k8s env). In fact, we took this approach in the first place to preserve the OPENJ9_JAVA_OPTIONS ENV specified in the Liberty container image:

ENV RANDFILE=/tmp/.rnd \
    OPENJ9_JAVA_OPTIONS="-XX:+IgnoreUnrecognizedVMOptions -XX:+IdleTuningGcOnIdle -Xshareclasses:name=openj9_system_scc,cacheDir=/opt/java/.scc,readonly,nonFatal -Dosgi.checkConfiguration=false"

[tjwatson]

Can we get this change merged soon?

[leochr]

@tjwatson We've been holding off on merging this until the fix for the issue mentioned here is released and validated. Has that happened?

[tjwatson]

@tjwatson We've been holding off on merging this until the fix for the issue mentioned here is released and validated. Has that happened?

Can we merge this fix ahead of the available Semeru containers with the fix available (expected mid October)? Perhaps it would be better to have that error mentioned occur now instead of silently ignoring the JIT Server configuration from the operator.

[leochr]

An Operator release prior to mid-October is not in the plan. So we'll proceed to merge. It'll still be good to validate once the Semeru fix is available.

[leochr]

@arturdzm please review and merge when you get a chance. Thanks

[leochr]

@arturdzm please review. Thanks