search cancel

Need a clarification on the Java heap memory allocation for the OpenShift Containers.

book

Article ID: 202881

calendar_today

Updated On:

Products

CA Cloud Test Mobile CA Application Test

Issue/Introduction

Our Dev team working on increasing the java heap memory for the SV Registry and Portal OpenShift containers. As per this DevTest Documentation we added -Xmx4144m in deployment config.yaml JAVA_OPTS for registry and portal containers.

But in the registry logs we are still seeing the default heap memory is used.

2020-10-30 10:51:05,085Z (10:51) [Event Sink Thread Pool Thread 4] INFO  com.itko.lisa.coordinator.TestRegistryImpl - Coordinator Servers: 0 Simulator Servers: 0 VSEs: 46 Running vusers: 0 Labs: 1 Memory used 727mb, allocated 1,003mb, max 910mb (79%) Our cpu usage 1%, system cpu used 13% GC time 1% db ping: 0 ms

We need the following clarifications in order to proceed further

  • As per this DevTest Documentation  sizing of docker containers section hardware requirements it is mentioned as 4 CPU cores minimum for docker containers. We are running registry and portal containers in a single pod. We would  like to know whether we need to allocate 4 CPU cores each for registry and portal containers?
  • Please let us know on how we can increase the Java heap memory for OpenShift Registry and Portal containers.

 

Environment

Release : 10.5

Component : CA Service Virtualization

Cause

Memory issues.

Resolution

Make these certain changes to the file "DeploymentConfig.yaml" by adding a property which should address the Xmx issue. 

Please note that the added lines are:

- name: SERVER_BASE_OPTS
value: -Xmx4144m

Here are the two paragraph stanzas where these were added.

*****************************Snippet 1*********************************************
spec:
containers:
- env:
- name: RUNTIME_ENV
value: ${RUNTIME_ENV}
- name: POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
- name: JAVA_OPTS
value: -Diam.server.url=https://testcasvdt5l.<COMPANY_NAME>.com:51111/auth -Ddevtest.enterprisedashboard.host=testcasvdt5l.<COMPANY_NAME>.com -Ddevtest.enterprisedashboard.port=1506 -Ddevtest.enterprisedashboard.https.enabled=true -Dlisa.threadDump.generate=false -Dlisa.webserver.host=$(POD_NAME) -Dlisa.webserver.port=1505 -Dlisadb.internal.enabled=true -Xmx4144m
- name: SERVER_BASE_OPTS
value: -Xmx4144m
image: docker.repo.<COMPANY_NAME>.com/<COMPANY_NAME>/grp-pi-te-sv/sv-registry:v2-latest
imagePullPolicy: Always
name: ca-service-virtualizationdevtest-registry
ports:

*****************************Snippet 1*********************************************

a little further down another 2 lines were added

*****************************Snippet 2*********************************************

- env:
- name: POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
- name: JAVA_OPTS
value: -DphoenixConfig.resHubServiceLoggingLevel=FULL -Dlisa.portal.url.prefix=https:// -Dlisa.webserver.https.enabled=false -Dhttps.protocols=TLSv1.2 -Dregistry.host=$(POD_NAME) -Dregistry.port=2010 -DlisaAutoConnect=tcp://$(POD_NAME):2010/Registry -Dregistry.portal.port=1505 -Dregistry.https.enabled=false -Xmx4144m
- name: SERVER_BASE_OPTS
value: -Xmx4144m
image: docker.repo.<COMPANY_NAME>.com/ca-broadcom/sv-devtest-10.5/portal:1.0.0
imagePullPolicy: Always
name: ca-service-virtualizationdevtest-portal
ports:

*****************************Snippet 2*********************************************

Please note that the entire yaml file could not be attached due to the proprietaries of the information.