If a deployment fails, check your instances or deployments to identify the problem. Go to Cloud Deployment Manager to see detailed messages that indicate what caused the deployment to fail.
A deployment might fail to complete successfully for one of the following reasons:
- Deployment Manager fails to create the necessary resources, such as an instance, disk, or load-balancing forwarding rule.
- You're missing an Identity and Access Management (IAM) role.
- The software fails to install and configure.
- The deployment does not configure the firewall rules correctly on the Virtual Private Cloud (VPC) network.
Accessing your deployment
You can return to your solutions by revisiting the Cloud Marketplace. Go to the solution page where you'll see a dropdown that shows your previous deployments. For services, you can update the service options directly from that page.
Otherwise, if your deployment is running a VM instance, you can go to Deployment Manager in the Google Cloud Console. From there, you can manage your deployment and view of all the Google Cloud resources that make up the deployment.
Deployment Manager failure
Sometimes your project might not have enough CPU quota to create the deployment resources. Check the remaining quota for your project at Compute Engine Quotas, and request additional quota if necessary. Then try to deploy the software package again.
You might not have the
roles/deploymentmanager.editor role applied to
your account for the project. Review Access control for users
to ensure that your user account is added to the correct project and that you
have the correct roles applied to the account.
Software installation failure
Most solutions provide logging for your instances so you can troubleshoot software installation failures. If your deployment creates all of the requested resources but you see a deployment error message in Deployment Manager, it's possible that the software failed to install or configure successfully. Connect to the instances in the deployment and check the logs to identify the issue.
Each entry in the log corresponds to the sequenced command that was executed on that instance. Execution terminates on the first command to exit with a non-zero exit code. The specific cause of failure will be typically be found at the end of the error log.
Network and firewall issues
If your software deployment hosts content on an external IP address, the software deployment automatically configures firewall rules on the network where your deployment runs. For example, Wordpress deployments open ports so that you can access the web console and display content over HTTP and HTTPS. If the firewall configuration fails, you might receive a HTTP 404 error when you try to connect to the web console on the external IP address.
If you need to configure the firewall manually or open additional ports for other services that run on your instances, add firewall rules to the VPC network that your deployments run on.
Sending email from software deployments
Some software packages try to send email over specific ports. However, Compute Engine does not allow outbound connections on port 25. To send email from the instances that host your software packages, you must follow the tutorial on sending mail.