Fix handling of os signals in tink server

Previously tink server failed to respect os signals like SIGTERM,SIGINT etc.

This PR fixes the tink server so that os signals sent to the tink server process
will actually kill the process.

Confirmed by runnning in docker-cpompose env:

```bash
make images
make run
docker exec -it tink_tinkerbell_1 sh
ps
kill -SIGINT 1
```

Corresponding logs:

```
tinkerbell_1             | {"level":"info","ts":1630442128.6128566,"caller":"tink-server/main.go:198","msg":"signal received, stopping servers","service":"github.com/tinkerbell/tink","signal":"interrupt"}
tinkerbell_1             | {"level":"info","ts":1630442129.4250214,"caller":"metrics/metrics.go:58","msg":"initializing label values","service":"github.com/tinkerbell/tink"}
tinkerbell_1             | {"level":"info","ts":1630442129.4286315,"caller":"http-server/http_server.go:88","msg":"serving http","service":"github.com/tinkerbell/tink"}
```

Now the docker container exists and we can see the log that signal is received.

Signed-off-by: Archana krishnan <[email protected]>

cmd/tink-server/main.go

@@ -191,12 +191,12 @@ func NewRootCommand(config *DaemonConfig, logger log.Logger) *cobra.Command {
 				HTTPBasicAuthPassword: config.HTTPBasicAuthPassword,
 			}, errCh)
 
-			<-ctx.Done()
 			select {
 			case err = <-errCh:
 				logger.Error(err)
 			case sig := <-sigs:
 				logger.With("signal", sig.String()).Info("signal received, stopping servers")
+				closer()
 			}
 
 			// wait for grpc server to shutdown