Troubleshooting¶
Common problems with the OpenTofu Runner AMI, what causes them and how to fix them.
The demo says first boot did not complete¶
Cause: sushi-opentofu-firstboot.service has not finished or failed, so
the workspace is missing.
Fix: Wait for cloud-init, then check the service and its log:
sudo cloud-init status --wait
systemctl status sushi-opentofu-firstboot.service
sudo journalctl -u sushi-opentofu-firstboot.service --no-pager
If it failed, keep a copy of the log, then run it again:
init cannot find a provider¶
Cause: The default CLI configuration only installs hashicorp/aws from
the local mirror and blocks registry downloads.
Fix: Point TF_CLI_CONFIG_FILE at your own CLI configuration that allows
the providers you need. See Providers.
init fails with a version constraint error¶
Cause: Your configuration asks for an AWS provider version other than 6.67.0, which is the only version in the local mirror.
Fix: Use = 6.67.0 for hashicorp/aws, or use your own CLI
configuration that can download the version you need.
init or plan runs out of memory¶
Cause: The AWS provider is large. Instances with 2 GiB of RAM are usually too small.
Fix: Use t3.medium (4 GiB) or larger.
Permission denied in the workspace¶
Cause: The workspace is owned by ubuntu. Running as another user, or
running tofu with sudo, can leave files ubuntu cannot change.
Fix: Run tofu as ubuntu, without sudo.
The demo refuses the plan¶
Cause: You changed the example workspace. The demo only applies the original synthetic example.
Fix: This is expected. Run your own validate and plan directly with
tofu -chdir=/var/lib/opentofu/workspace.