Scope inspiration:

gce_ssh.sh, gce_foreach_vm.sh.

This series follows the repo’s GCP inventory flow so every step builds a repeatable, audit-friendly picture of your environment. Part 12 closes out the series with safe, automatable VM access for operations workflows.

---

What this script does (walkthrough)

gce_ssh.sh

wraps

gcloud compute ssh

so the zone is always correct, even when your config points elsewhere.

1. Accept a VM name or IP address and resolve IPs back to instance names.
2. Unset inherited

   CLOUDSDK_COMPUTE_ZONE

   to prevent accidental mis-targeting.
3. Determine the VM zone using

   gcloud compute instances list

   before SSH.
4. Invoke

   gcloud compute ssh

   with the resolved zone to avoid interactive prompts.

---

Operational caveats and gotchas

• Inherited

  CLOUDSDK_COMPUTE_ZONE

  values can cause false “resource not found” errors if the VM is in another zone; the script clears that variable on purpose.
• If the VM isn’t found, the script prints the current project and region so you can detect misconfigured context quickly.
• You can pass standard

  gcloud compute ssh

  flags (like

  --command

  ) for automation or bulk operations.

---

Example command usage

# SSH to a VM by name with automatic zone resolution
gcp/gce_ssh.sh my-vm-name

# SSH to a VM by IP address
gcp/gce_ssh.sh 10.0.0.25

# Run a command across a fleet
for x in {1..10}; do gcp/gce_ssh.sh vm-$x --command 'sudo systemctl restart myapp'; echo; done