What causes "build failed, not uploading artifacts" in hashicorp/packer?

Read the build log above this error for the actual builder/provisioner failure and fix that root cause

build failed, not uploading artifacts

Error message

build failed, not uploading artifacts

What it means

completeBuild is called when a build finishes. If the build itself errored (buildErr != nil), Packer marks the build FAILED (or CANCELLED if the context was cancelled) in HCP Packer and returns this error to signal that no artifacts will be uploaded to the registry. It is the user-facing consequence of the build failing, not a registry failure itself.

Source

Thrown at internal/hcp/registry/types.bucket.go:744

	if !ok {
		log.Print("[ERROR] done build does not have an entry in the heartbeat table, state will be inconsistent.")
	} else {
		log.Printf("[TRACE] signal stopping heartbeats")
		// Stop heartbeating
		doneCh <- struct{}{}
		log.Printf("[TRACE] stopped heartbeats")
	}

	if buildErr != nil {
		status := hcpPackerModels.HashicorpCloudPacker20230101BuildStatusBUILDFAILED
		if ctx.Err() != nil {
			status = hcpPackerModels.HashicorpCloudPacker20230101BuildStatusBUILDCANCELLED
		}
		err := bucket.UpdateBuildStatus(context.Background(), buildName, status)
		if err != nil {
			log.Printf("[ERROR] failed to update build %q status to FAILED: %s", buildName, err)
		}
		return packerSDKArtifacts, fmt.Errorf("build failed, not uploading artifacts")
	}

	artifacts, err := bucket.doCompleteBuild(ctx, buildName, packerSDKArtifacts, ui, buildErr)
	if err != nil {
		err := bucket.UpdateBuildStatus(ctx, buildName, hcpPackerModels.HashicorpCloudPacker20230101BuildStatusBUILDFAILED)
		if err != nil {
			log.Printf("[ERROR] failed to update build %q status to FAILED: %s", buildName, err)
		}
	}

	return artifacts, err
}

func (bucket *Bucket) doCompleteBuild(
	ctx context.Context,
	buildName string,
	packerSDKArtifacts []packerSDK.Artifact,
	ui packerSDK.Ui,

View on GitHub (pinned to eb36e3c3e4)

Solutions

Read the build log above this error for the actual builder/provisioner failure and fix that root cause
If cancelled, re-run the build with adequate timeouts
Verify communicator connectivity (SSH/WinRM credentials, security groups) if provisioning failed
Re-run packer build after fixing; artifacts will upload once the build succeeds

Defensive patterns

Strategy: try-catch

Try / catch

// Go
artifacts, err := bucket.CompleteBuild(ctx, buildName, arts, ui, buildErr)
if err != nil {
	if strings.Contains(err.Error(), "build failed, not uploading artifacts") {
		// root cause is in buildErr — surface it, don't retry the upload
		return fmt.Errorf("build failed upstream: %w", buildErr)
	}
	return err
}

Prevention

Fix the underlying builder/provisioner error reported before this message
Set generous CI timeouts to avoid context cancellation mid-build
Validate provisioner scripts locally before registry-backed runs
Check communicator connectivity (SSH keys, security groups) before long builds

When it happens

Trigger: Bucket.CompleteBuild -> completeBuild is invoked with a non-nil buildErr — i.e. the underlying builder/provisioner failed or the run context was cancelled — so completeBuild skips doCompleteBuild and returns "build failed, not uploading artifacts".

Common situations: Provisioner script failure; builder could not reach the source image or SSH; user pressed Ctrl-C (ctx cancelled -> CANCELLED status); CI timeout killing the run; any plugin error during the build.

Related errors

AI-assisted analysis of hashicorp/packer@eb36e3c3e4 (2026-09-05). Data as JSON: /api/errors/00a178ef67263998. Report an issue: GitHub.