vCenter Server services fail to start automatically on boot and require manual startup

Products

• VMware vCenter Server

Issue/Introduction

• Accessing the vSphere Client displays a "no healthy upstream" message.
  
  
• For vCenter Server instances configured in Linked Mode, execution of the vdcrepadmin command (usr/lib/vmware-vmdir/bin/vdcrepadmin -f showpartnerstatus -h localhost -u administrator -w <password>) on the problematic node returns the following output:

Partner: vc1_fqdn
Host available:  Yes
Status available: No

• Execution of the same command on the central Linked Mode node (star topology) returns the following output:

Partner: vc2_fqdn
Host available:  Yes
Status available: Yes

The active status of the central node VC is not available on problematic VC. But the central node can get the active status of problematic VC.

Environment

• VMware vCenter Server 7.0.x
  
  
• VMware vCenter Server 8.0.x

Cause

The Primary Network Identifier (PNID) of the vCenter Server is configured as a Fully Qualified Domain Name (FQDN). The configured Domain Name System (DNS) server fails to provide accurate domain name resolution. This failure prevents services from starting automatically on boot. The impact is categorized based on the deployment type:

• Standalone vCenter Server: Domain name resolution failure prevents the appliance from resolving its own FQDN (PNID). Core local services are unable to bind to the hostname, which stalls the startup sequence.

• Linked Mode vCenter Server: In addition to local hostname resolution failures, the domain name of the Linked Mode partner nodes cannot be resolved successfully. This causes cross-node authentication and replication services to fail during initialization.

Resolution

1. DNS connectivity verification: Verify network connectivity to the DNS server by executing a ping command to the DNS server IP address. Network routing or firewall issues must be excluded before proceeding.
   
   ping dns_server_ip

2. DNS resolution function validation: Validate the DNS resolution functionality by executing nslookup <vc_fqdn_ip> or dig <vc_fqdn_ip>. In Enhanced Linked Mode vCenter, successful resolution of both the local vCenter Server domain name and the Linked Mode partner domain name is required. If resolution fails, verification of the A records within the DNS server is necessary.

3. vCenter Server Appliance restart: Upon correction of the DNS server configuration, a reboot of the vCenter Server Appliance (VCSA) is required. Prior to rebooting, a memory snapshot or file-based backup of the vCenter Server must be captured. For Linked Mode configurations, powered-off snapshots of all participating vCenter Server nodes must be captured concurrently before any operations are performed.

Workaround

Domain name resolution records can be manually added to the /etc/hosts directory on the vCenter Server appliance until proper DNS functionality is restored.

Additional Information

For more information regarding deployment prerequisites, see DNS Requirements for the vCenter Server Appliance.