Troubleshooting Failing Builds
Packer builds can fail in Atlas for a number of reasons – improper configuration, transient networking errors, and hardware constraints are all possible. Below is a list of debugging options you can use.
Verbose Packer Logging
You can set a variable in the UI that increases the logging verbosity
in Packer. Set the
PACKER_LOG key to a value of
1 to accomplish this.
After setting the variable, you'll need to rebuild.
Verbose logging will be much louder than normal Packer logs and isn't recommended for day-to-day operations. Once enabled, you'll be able to see in further detail why things failed or what operations Packer was performing.
This can also be used locally:
PACKER_LOG=1 packer build ...
Some VM builds, such as VMware or Virtualbox, may hang at various stages,
Waiting for SSH....
Things to pay attention to when this happens:
- SSH credentials must be properly configured. AWS keypairs should match, SSH usernames should be correct, passwords should match, etc.
- Any VM preseed configuration should have the same SSH configuration as your template defines
A good way to debug this is to manually attempt to use the same SSH
configuration locally, running with
packer build -debug. See
more about debugging Packer builds.
Your build may be failing by requesting larger memory or disk usage then is available. Read more about the build environment.
Typically Packer builds that fail due to requesting hardware limits that exceed Atlas's hardware limitations will fail with a The operation was canceled error message as shown below:
... ==> vmware-iso: Starting virtual machine... vmware-iso: The VM will be run headless, without a GUI. If you want to vmware-iso: view the screen of the VM, connect via VNC without a password to vmware-iso: 127.0.0.1:5918 ==> vmware-iso: Error starting VM: VMware error: Error: The operation was canceled ==> vmware-iso: Waiting 4.604392397s to give VMware time to clean up... ==> vmware-iso: Deleting output directory... Build 'vmware-iso' errored: Error starting VM: VMware error: Error: The operation was canceled ==> Some builds didn't complete successfully and had errors: --> vmware-iso: Error starting VM: VMware error: Error: The operation was canceled
Sometimes it's faster to debug failing builds locally. In this case, you'll want to install Packer and any providers (like Virtualbox) necessary.
Because Atlas runs the open source version of Packer, there should be no difference in execution between the two, other than the environment that Packer is running in. For more on hardware constraints in the Atlas environment read below.
Once your builds are running smoothly locally you can push it up to Atlas for versioning and automated builds.
This is a short list of internal errors and what they mean.
- SIC-001: Your data was being ingressed from GitHub but failed to properly unpack. This can be caused by bad permissions, using symlinks or very large repository sizes. Using symlinks inside of the packer directory, or the root of the repository, if the packer directory is unspecified, will result in this internal error. Note: Most often this error occurs when applications or builds are linked to a GitHub repository and the directory and/or template paths are incorrect. Double check that the paths specified when you linked the GitHub repository match the actual paths to your template file.
- SEC-001: Your data was being unpacked from a tarball uploaded to Atlas and encountered an error. This can be caused by bad permissions, using symlinks or very large tarball sizes.
Packer is an open source project with an active community. If you're having an issue specific to Packer, the best avenue for support is the mailing list or IRC. All bug reports should go to GitHub.
- Website: packer.io
- GitHub: github.com/mitchellh/packer
- Mailing list: Google Groups
If you believe your build is failing as a result of a bug in Atlas, or would like other support, please email us.