What causes "failed to delete logical view: %w" in googleapis/mcp-toolbox?

Treat codes.NotFound as success (idempotent delete) using errors.As/status.FromError on the wrapped error

failed to delete logical view: %w

Error message

failed to delete logical view: %w

What it means

DeleteLogicalView wraps errors from InstanceAdmin.DeleteLogicalView. Deletion fails if the logical view does not exist (NotFound), permissions are missing, or a transient gRPC error occurs. The raw API error is preserved via %w.

Source

Thrown at internal/sources/bigtable/admin_wrappers.go:232

	return map[string]string{"status": "logical view created successfully"}, nil
}

func (s *Source) UpdateLogicalView(ctx context.Context, instanceId, logicalViewId, query string) (any, error) {
	conf := bigtable.LogicalViewInfo{ // MUST be value per bigtable SDK
		LogicalViewID: logicalViewId,
		Query:         query,
	}
	err := s.InstanceAdmin.UpdateLogicalView(ctx, instanceId, conf)
	if err != nil {
		return nil, fmt.Errorf("failed to update logical view: %w", err)
	}
	return map[string]string{"status": "logical view updated successfully"}, nil
}

func (s *Source) DeleteLogicalView(ctx context.Context, instanceId, logicalViewId string) (any, error) {
	err := s.InstanceAdmin.DeleteLogicalView(ctx, instanceId, logicalViewId)
	if err != nil {
		return nil, fmt.Errorf("failed to delete logical view: %w", err)
	}
	return map[string]string{"status": "logical view deleted successfully"}, nil
}

func (s *Source) GetMaterializedView(ctx context.Context, instanceId, materializedViewId string) (any, error) {
	view, err := s.InstanceAdmin.MaterializedViewInfo(ctx, instanceId, materializedViewId)
	if err != nil {
		return nil, fmt.Errorf("failed to get materialized view: %w", err)
	}
	return view, nil
}

func (s *Source) CreateMaterializedView(ctx context.Context, instanceId, materializedViewId, query string) (any, error) {
	conf := &bigtable.MaterializedViewInfo{
		MaterializedViewID: materializedViewId,
		Query:              query,
	}
	err := s.InstanceAdmin.CreateMaterializedView(ctx, instanceId, conf)

View on GitHub (pinned to 8cc6e09de2)

Solutions

Treat codes.NotFound as success (idempotent delete) using errors.As/status.FromError on the wrapped error
Confirm the correct ID type: logical views and materialized views live in separate namespaces
Grant roles/bigtable.admin or bigtable.views.delete to the caller
Retry on Unavailable/DeadlineExceeded transient codes

Example fix

// before: non-idempotent delete breaks cleanup scripts
err := s.InstanceAdmin.DeleteLogicalView(ctx, instanceId, id)
if err != nil { return nil, fmt.Errorf("failed to delete logical view: %w", err) }
// after: tolerate already-deleted
err := s.InstanceAdmin.DeleteLogicalView(ctx, instanceId, id)
if err != nil {
    if st, ok := status.FromError(err); ok && st.Code() == codes.NotFound {
        return map[string]string{"status": "logical view already deleted"}, nil
    }
    return nil, fmt.Errorf("failed to delete logical view: %w", err)
}

Defensive patterns

Strategy: try-catch

Validate before calling

views, err := src.ListLogicalViews(ctx, instanceId)
if err != nil { return err }
if !containsViewID(views, viewId) {
    return nil // nothing to delete; treat as success
}

Type guard

func isNotFound(err error) bool {
    st, ok := status.FromError(errors.Unwrap(err))
    return ok && st.Code() == codes.NotFound
}

Try / catch

res, err := src.DeleteLogicalView(ctx, instanceId, viewId)
switch {
case err == nil:
    // deleted
case isNotFound(err):
    // already gone: treat as success (idempotent delete)
default:
    return err
}

Prevention

Write delete/cleanup scripts to treat NotFound as success so reruns are idempotent
Don't mix logical-view and materialized-view IDs — they are separate namespaces
Verify the view still exists if another process may have deleted it
Retry transient codes (Unavailable, DeadlineExceeded) with backoff

When it happens

Trigger: Calling DeleteLogicalView (bigtable_delete_logical_view tool) when: logicalViewId was already deleted or never existed, caller lacks bigtable.views.delete permission, or the InstanceAdmin RPC fails due to network/timeout.

Common situations: Double-delete in cleanup scripts expecting idempotency; deleting a logical view that was actually a materialized view (wrong ID type); IAM changes; regional API outage.

Related errors

AI-assisted analysis of googleapis/mcp-toolbox@8cc6e09de2 (2026-09-05). Data as JSON: /api/errors/5dc612e6b28f2d4e. Report an issue: GitHub.