Aria automation patch pre-validation fails with " LCMVRAVACONFIG590047 VMware Aria Automation Upgrade/Patch Pre-validation failed"

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

Below observations are seen while running the below commands in Aria automation via SSH session 

/opt/health/run.sh on Aria automation threw "make: *** [/opt/health/Makefile:73: memory-usage] Error 1"

Free -g shows memory usage is more than 90 percent 

Top command describes postgres memory as a high memory usage contender.

Log snippet from Aria suite lifecycle manager: /var/log/vrlcm/vmware_vrlcm.log

####-##-##T##:##:##Z INFO vrlcm[1511] [pool-3-thread-32] [c.v.v.l.p.a.s.Task]  -- Injecting task failure event. Error Code : 'LCMVRAVACONFIG590047', Retry : 'false', Causing Properties : '{ CAUSE ::  }'
com.vmware.vrealize.lcm.common.exception.EngineException: VMware Aria Automation Upgrade/Patch pre-validation failed. Check pre-check results from upgrade wizard for detailed report.
        at com.vmware.vrealize.lcm.plugin.core.vra80.task.upgrade.VraVaUpgradePreValidationTask.execute(VraVaUpgradePreValidationTask.java:247) [vmlcm-vrapre
ludeplugin-core-8.18.0-SNAPSHOT.jar!/:?]

Environment

Aria automation 8.18.1

Cause

There was a change introduced in 8.18.1.This change reconfigured a shared buffer used by the postgres database which increased the value from ~1.5 GiB to 12 GiB on XL environments (and from 1 GiB to 4 GiB on medium environments) without recommendation of bumping the VA memory.

Resolution

Broadcom is aware of this issue and working internally for a solution.

Workaround:

1.As best practice, Take a snapshot of all the Aria Automation appliance(s) before proceeding

2.Take SSH session to Aria automation and restart the pods using /opt/scripts/deploy.sh 

3.Re-run the health-check - /opt/health/run.sh