Error: "Could not connect to database: Verify the database host is reachable from this machine and the provided credentials are valid" encountered when deploying an application cell to a Cloud Director server group

Products

• VMware Cloud Director

Issue/Introduction

• Adding application cell does not work
• When deploying a VMware Cloud Director appliance cell, the newly added application cell may fail to connect to the primary database cell.

In the /opt/vmware/var/log/vcd/configure-vcd.log file the following entries are seen

2026-01-08 09:58:50 | setupvcd.sh | [INFO] Response file exists in the NFS transfer storage. This assumes that this newly deployed cell will be configured as an additional cell into an already existing cluster. If you want to deploy this cell as part of new cluster deployment, ensure that NFS transfer storage is empty. For cells to be part of new cluster deployment both the NFS transfer storage and the DB should be in empty clean state.
Sleeping for 90 seconds for primary cell to enable DB access to this cell's IP.
2026-01-08 09:58:50 | setupvcd.sh | Sleeping for 90 seconds for primary cell to enable DB access to this cell's IP.
Configuring VCD from response file (/opt/vmware/vcloud-director/data/transfer/responses.properties).
2026-01-08 10:00:20 | setupvcd.sh | Configuring VCD from response file (/opt/vmware/vcloud-director/data/transfer/responses.properties).
Could not connect to database: Verify the database host is reachable from this machine and the provided credentials are valid

2026-01-08 10:00:32 | setupvcd.sh | [ERROR] Failed to configure vCD. Check /opt/vmware/vcloud-director/logs for details. Aborting.
2026-01-08 10:00:32 | configure-vcd.sh | [ERROR] setupvcd script failed to execute.
2026-01-08 10:00:32 | cleanup-vcd.sh | Starting cleanup script.

Environment

VMware Cloud Director 10.6.x

Cause

The Cloud Director appliance eth0 and eth1 interfaces are configured on different subnets but reside within the same broadcast domain (VLAN). This causes a routing issue where the OS routes traffic through the first primary entry in the routing table (eth0) even when the packet is destined for the second interface (eth1).

Resolution

Place the subnets for eth0 (Management) and eth1 (Database/Replication) on separate VLANs. This ensures proper traffic isolation.