Under /sos-####/VCFMgmtComponents/vcf-management-services/services-logs/ you may observe entries similar to the following within vcf-fleet-upgrade-service.log:
org.springframework.web.client.HttpServerErrorException$BadGateway: 502 Bad Gateway: "connection failed: Get "https://#########:443/PROD/metadata/productVersionCatalog/v1/productVersionCatalog.json": unexpected EOF<EOL>"
VMware Cloud Foundation Operations 9.1.1
The issue is caused due to network timeouts when downloading the large master catalog manifest (productVersionCatalog.json) from the offline depot. Intermediate network devices terminate the connection after ~9–10 seconds, triggering a 502 Bad Gateway (unexpected EOF) that the vcf-fleet-upgrade-service fails to handle, leaving the parent upgrade task stuck in "In Progress" in the VCF Operations UI.
vcf-fleet-upgrade-service microservices, effectively clearing stagnant TCP socket states and resetting task database locks.