Mismatched name referencing preventing tear down of Cloud instances.

[dacbd]

Currently still investigating but it feels related to #742 and #738

I can see in GCP logs a Failed:Delete VM that 404's to iterative-2r77ew2wn9qkm where the name of the instance is using a cml prefix instead.

[dacbd]

cml runner cmd:

  cml-runner \
    --single \
    --name=cli-name-opt \
    --log=debug \
    --idle-timeout=1800 \
    --token=*** \
    --cloud=gcp \
    --cloud-region=us-west \
    --cloud-type=m \
    --cloud-hdd-size=10

Service logs:

~$ journalctl --unit cml --no-pager
-- Logs begin at Mon 2021-09-27 17:51:44 UTC, end at Mon 2021-09-27 18:17:01 UTC. --
Sep 27 17:54:53 cli-name-opt systemd[1]: Started cml.service.
Sep 27 17:55:02 cli-name-opt cml.sh[22569]: {"level":"info","message":"Preparing workdir /tmp/tmp.uAzCgwV9Td/.cml/cli-name-opt..."}
Sep 27 17:55:02 cli-name-opt cml.sh[22569]: {"level":"info","message":"Launching github runner"}
Sep 27 17:55:11 cli-name-opt cml.sh[22569]: {"level":"warn","message":"SpotNotifier can not be started."}
Sep 27 17:55:12 cli-name-opt cml.sh[22569]: {"date":"2021-09-27T17:55:12.622Z","level":"info","message":"runner status","repo":"xxx"}
Sep 27 17:55:12 cli-name-opt cml.sh[22569]: {"date":"2021-09-27T17:55:12.624Z","level":"info","message":"runner status √ Connected to GitHub","repo":"xxx"}
Sep 27 17:55:13 cli-name-opt cml.sh[22569]: {"date":"2021-09-27T17:55:13.085Z","level":"info","message":"runner status Listening for Jobs","repo":"xxx","status":"ready"}
Sep 27 17:55:27 cli-name-opt cml.sh[22569]: {"date":"2021-09-27T17:55:27.393Z","job":3723710456,"level":"info","message":"runner status Running job: train-model","repo":"xxx","status":"job_started"}
Sep 27 18:00:24 cli-name-opt cml.sh[22569]: {"date":"2021-09-27T18:00:24.822Z","job":"","level":"info","message":"runner status Job train-model completed with result: Succeeded","repo":"xxx","status":"job_ended","success":true}
Sep 27 18:00:25 cli-name-opt cml.sh[22569]: {"date":"2021-09-27T18:00:25.285Z","level":"info","message":"runner status √ Removed .credentials","repo":"xxx"}
Sep 27 18:00:25 cli-name-opt cml.sh[22569]: {"date":"2021-09-27T18:00:25.286Z","level":"info","message":"runner status √ Removed .runner","repo":"xxx"}
Sep 27 18:00:25 cli-name-opt cml.sh[22569]: {"level":"info","message":"runner status","reason":"proc_exit","status":"terminated"}
Sep 27 18:00:25 cli-name-opt cml.sh[22569]: {"level":"info","message":"waiting 20 seconds before exiting..."}
Sep 27 18:00:45 cli-name-opt cml.sh[22569]: {"level":"info","message":"Unregistering runner cli-name-opt..."}
Sep 27 18:00:45 cli-name-opt cml.sh[22569]: {"level":"error","message":"\tFailed: Cannot destructure property 'id' of '(intermediate value)' as it is undefined."}
Sep 27 18:00:47 cli-name-opt systemd[1]: cml.service: Succeeded.

[dacbd]

using the following results in the same behavior:

  CML_RUNNER_NAME=cml-test-name cml-runner \
    --single \
    --log=debug \
    --idle-timeout=1800 \
    --token=*** \
    --cloud=gcp \
    --cloud-region=us-west \
    --cloud-type=m \
    --cloud-hdd-size=10

[dacbd]

The ID being used from tear down is generated here:
https://github.com/iterative/terraform-provider-iterative/blob/732d0919577867d40b7ff635dd0667caf19d3cb9/iterative/utils/helpers.go#L17-L25

Looks like the ID get pulled from here for tear down:
https://github.com/iterative/terraform-provider-iterative/blob/732d0919577867d40b7ff635dd0667caf19d3cb9/iterative/gcp/provider.go#L273-L275

Golang documentation has some not exactly clear wording that makes me think this should be the name and not the Id...
https://pkg.go.dev/google.golang.org/api/compute/v1#InstancesService.Delete

[dacbd]

iterative/terraform-provider-iterative#224

[DavidGOrtega]

related to #678

[dacbd]

Closed with upstream release of: https://github.com/iterative/terraform-provider-iterative