Provisioning Failures
This article covers issues that occur when Thinfinity® Cloud Manager attempts to provision a new instance but the operation fails or hangs. Symptoms include deployments stuck in the Running state, Terraform Apply errors returned in the deployment log, and instances that provision but never register with the broker.
Common Symptoms
| Symptom | Likely Cause |
|---|---|
| Deployment status: Failed; error mentions “quota”, “limit”, “capacity” | Cloud provider quota exceeded |
| Deployment status: Failed; error mentions “image not found”, • “AMI not found”, “template not found” | Image ID / Golden Image reference invalid for the region or account |
| Deployment status: Failed; error mentions “subnet”, “VPC”, “network” | Network configuration in the model template is invalid |
| Deployment succeeds, instance runs, but “Instance Registered” event never occurs | Virtualization Agent not installed, not running, or cannot reach the broker |
| Deployment stuck in Running for > 30 minutes | Provider is slow, or Terraform Apply is blocked on a resource that will never reconcile |
Cloud Provider Quota
Quota errors are among the most common provisioning failures. The deployment log will contain a provider error message referring to a limit or quota — for example, “You have requested more instances than your current instance limit allows”, “VcpuLimitExceeded”, “Operation could not be completed as it results in exceeding approved quota”, or “SkuNotAvailable”.
Quota increases are requested in the cloud provider’s console, not in Cloud Manager. See your provider’s documentation for the quota request workflow. After the increase is approved, retry the deployment.
| Provider | Console area |
|---|---|
| AWS | Service Quotas |
| Azure | Usage + quotas |
| GCP | IAM & Admin > Quotas |
| Oracle Cloud | Governance > Limits, Quotas, and Usage |
| IONOS | Contact IONOS account team |
| VMware / Hyper-V / Proxmox / Verge.IO | Finite host capacity — add hosts or reduce instance count |
Image / Template Not Found
Error messages vary by provider but generally indicate that the image specified in the model is not accessible in the target region or account.
Cloud Provider Images
| Cause | What to check / fix |
|---|---|
| The image ID is from a different region than the target region | Verify the image exists in the target region via the provider console. Update the model’s General Settings with the correct image ID. |
| The image has been deprecated or deleted by the provider | Select a current image in the target region and update the model. |
| The image is private and not shared with the target account | Share the image with the account, or use a different image. |
| Account or subscription doesn’t have rights to use the image | Grant the credential account permission to use the image, or use a different image. |
When testing image updates, validate in a Testing or Staging model version before promoting to Production.
Golden Images (On-Premises)
Error messages typically mention the template or VM name not being found.
| Cause | What to check / fix |
|---|---|
| The Golden Image template was deleted or renamed in the hypervisor | Verify the Golden Image exists in the target hypervisor environment via the management console. Update the model’s General Settings with the correct reference. |
| The Golden Image exists but is on a different datacenter / cluster / node • than the credential set targets | Move the Golden Image to a location the credential set can reach, or update the credential set to target the correct scope. |
| The Golden Image was converted back to a VM and is no longer a template | Convert the VM back to a template in the hypervisor, or capture a new Golden Image. See Create a Golden Image. |
Network / Subnet Errors
| Error pattern | What to check / fix |
|---|---|
“Subnet <id> not found” |
Verify the subnet exists in the target region and account. |
| “No IP addresses available in subnet” | The subnet’s CIDR range is exhausted. Enlarge the subnet if possible, or reconfigure the model to use a different subnet. |
| “Virtual network has no available address space” | The parent VPC/VNet is full. Use a different network, or expand the address space. |
| “Security group not in same VPC as subnet” | Ensure the security group and the subnet are in the same VPC. |
For Terraform-authored models, check main.tf for hardcoded resource references that may not exist in the target account. Use variables for environment-specific values.
Instance Runs But Doesn’t Register
The deployment succeeds, the VM is running in the provider, but no “Instance Registered” entry appears in the instance’s View Log dialog. This means the Virtualization Agent did not check in with the broker.
| Cause | How to Verify |
|---|---|
| Virtualization Agent not installed on the image | RDP / SSH to the VM using provider console or direct IP; check for the agent installation |
| Virtualization Agent service not running | On the VM, check Services (Windows) or systemctl (Linux) for the agent service |
| Agent cannot reach broker (network) | From the VM, test connectivity to the broker’s FQDN on the agent port |
| Agent configured with wrong broker URL | Check the agent’s configuration file on the VM |
| Agent certificate issue | If the broker uses TLS with a private CA, the VM must trust that CA |
If the agent is misconfigured or missing, update the Golden Image with a correctly configured agent and redeploy. See Create a Golden Image.
Deployment Stuck in “Running”
A deployment that has been in the Running state for more than 30 minutes is likely stuck. Open the deployment’s View Log in the Deployments tab and match what you see to the table below.
| Log signal | Likely cause | Fix |
|---|---|---|
| Log shows a specific resource it’s waiting on | Terraform is blocked on a resource that will never reconcile (for example, waiting for an AMI to become available that was never made available in the region) | Resolve the resource at the provider, or cancel the deployment and fix the template |
| Log shows an unrecoverable error | Terraform Apply failed and cannot recover | In the Deployments tab, use the Destroy action on the affected row, then Delete to remove the deployment record |
| Log appears truncated or stopped abruptly with no error | Terraform process may have crashed but the deployment state wasn’t updated | Same Destroy → Delete sequence as above |
| Provider status page reports a regional incident | Provider-side outage affecting deployments in the region | Wait for incident resolution and re-attempt once the provider reports the issue resolved |