When performing a Helm upgrade for a CNF, the following symptoms are observed:
values.yaml files.3.4
This issue is caused by the processing overhead of large or numerous values.yaml files. During the upgrade process, TCA base64 encodes these files and includes them in log statements. When a CNF has many Helm charts, the cumulative size of these encoded strings creates significant messaging overhead and resource contention, leading to GUI unresponsiveness and initiation delays.
This issue is resolved in TCA 3.4.0.1. Please refer 3.4.0.1 release notes
If an immediate upgrade is not possible, the following workaround can alleviate the impact:
Suppress Verbose Logging: Reduce the logging level for CNF operations to minimize the processing of large base64 strings in log outputs.
The following steps must be executed within the TCA-M and all relevant TCA-CPs to effectively stabilize the environment.
kubectl exec -it tca-mgr tca-app-<podname> -- bashcd configsed -i 's#<root level="INFO">#<logger name="com.vmware.hybridity.service.nfv.vim.workflows.orch.CnfReconfigureJob" level="OFF" /><root level="INFO">#g' app-logback.xmlsed -i 's#<root level="INFO">#<logger name="com.vmware.hybridity.service.nfv.vim.workflows.orch.ComponentReconfigureJob" level="OFF" /><root level="INFO">#g' app-logback.xmlsed -i 's#<root level="INFO">#<logger name="com.vmware.hybridity.service.nfv.vim.workflows.orch.CnfVduUpgradeJob" level="OFF" /><root level="INFO">#g' app-logback.xml