Introduction
The purpose of this document is to help customers review their own support bundles without the assistance of Terraform Support. This allows the customer to self diagnose problems.
The support bundle can be generated from the Installer dashboard on port 8800 of the instance.
On the dashboard, click on the Support tab. On the next page, click the Download Support Bundle button, which downloads the support bundle directly to your web browser.
If you are unable to generate a bundle from the dashboard, connect to the instance via SSH and generate a bundle by running replicatedctl support-bundle
.
Use Case
Customer is encountering errors in TFE but the errors in the User Interface are not specific enough to allow for identification of root cause or further troubleshooting without log review.
Procedure
For v4 builds of Terraform look at the logs in [bundle]/master/app/logs. In v5, these files are in [bundle]/master/scheduler/kubernetes/logs Each of the Docker containers has its own stderr and stdout files here. Depending on where in the application flow the problem is occurring, identify the logs for the relevant container. These are the most important and cover the majority of issues:
ptfe-atlas(v4) atlas(v5) is the PTFE application itself that receives requests from users via web browser or API call.
ptfe-ingress(v4) ingress(v5) is the container that copies the files to TFE that are to be processed either from the VCS repository or local client system.
ptfe-vault(v5) vault(v5) is the internal Vault server that stores all decrypts/encrypts all sensitive data
When PTFE actually needs to execute the terraform executable like it would at a command line, it generates a new temporary Docker container to execute and do the work. In v4 these containers’ names appear to be random combinations of adjectives and proper names, so you’ll see things like “angry_boyd” or “languid_gates” In v5, these logs will start with “tf-runner” Most syntax or provider errors will appear here.