By Mark Drake
Manager, Developer Education
A managed Redis instance can provide benefits like high availability and automated updates. However, any time you make a connection to a remote database server, you run the risk of malicious actors sniffing the sensitive information you send to it.
redis-cli, the Redis command line interface, doesn’t natively support connections over TLS, a cryptographic protocol that allows for secure communications over a network. This means that without further configuration,
redis-cli is not a secure way to connect to a remote Redis server. One way to establish a secure connection to a managed Redis instance is to create a tunnel that uses the TLS protocol.
Stunnel is an open-source proxy used to create secure tunnels, allowing you to communicate with other machines over TLS. In this guide, we will walk through installing and configuring stunnel so you can connect to a managed Redis instance over TLS with
To complete this guide, you will need:
ufw. To set this up, follow our initial server setup guide for Ubuntu 18.04.
When you install a Redis server, it usually comes packaged with
redis-cli. However, you can install
redis-cli without the Redis server by installing the
redis-tools package from the default Ubuntu repositories. You can also install stunnel from the default Ubuntu repositories by downloading the
First, update your server’s package index if you’ve not done so recently:
- sudo apt update
Then install the
stunnel4 packages with APT:
- sudo apt install redis-tools stunnel4
When prompted, press
ENTER to confirm that you want to install the packages.
You can check whether stunnel was installed correctly and its systemd service is working by running the following command:
- sudo systemctl status stunnel4
Output● stunnel4.service - LSB: Start or stop stunnel 4.x (TLS tunnel for network daemons) Loaded: loaded (/etc/init.d/stunnel4; generated) Active: active (exited) since Thu 2019-09-12 14:34:05 UTC; 8s ago Docs: man:systemd-sysv-generator(8) Tasks: 0 (limit: 2362) CGroup: /system.slice/stunnel4.service Sep 12 14:34:05 stunnel systemd: Starting LSB: Start or stop stunnel 4.x (TLS tunnel for network daemons)... Sep 12 14:34:05 stunnel stunnel4: TLS tunnels disabled, see /etc/default/stunnel4 Sep 12 14:34:05 stunnel systemd: Started LSB: Start or stop stunnel 4.x (TLS tunnel for network daemons).
Here, you can see that the stunnel service is active, though the process immediately exited. This tells us that stunnel is running, but it isn’t able to actually do anything since we haven’t yet configured it.
Modern Linux systems rely on systemd for initializing and managing services and daemons. However, stunnel uses a SysV-style init script, which is based on the older UNIX System V init system, for startup. You’ll need to modify the
/etc/default/stunnel4 file to enable this init script.
Open this file with your preferred text editor. Here, we’ll use
- sudo nano /etc/default/stunnel4
ENABLED option near the top of the file. It will be set to
0 by default, but change this to
1 to enable stunnel to start at boot:
# /etc/default/stunnel # Julien LEMOINE <firstname.lastname@example.org> # September 2003 # Change to one to enable stunnel automatic startup ENABLED=1 . . .
Save and close the file. If you used
nano to edit the file, do so by pressing
Next, you will need to create a configuration file for stunnel which will tell the program where it needs to route traffic.
Open a new file called
stunnel.conf under the
- sudo nano /etc/stunnel/stunnel.conf
Add the following content to this file:
fips = no setuid = nobody setgid = nogroup pid = /home/sammy/pids/stunnel.pid debug = 7 delay = yes [redis-cli] client = yes accept = 127.0.0.1:8000 connect = managed_redis_hostname_or_ip:managed_redis_port
The first five lines in the file are global options, meaning they will apply to every service you include in this file:
fips: Enables or disables stunnel’s FIPS 140-2 mode. In this mode, stunnel will validate that the connection meets the Federal Information Processing Standard. Setting this to
nodisables this feature. Note that disabling this is not any less secure, but keeping it enabled (as it is by default) would require some extra configuration.
setuid: Defines the Unix user ID under which stunnel will run. By default, the stunnel process is owned by the root user. However, the stunnel documentation recommends that you drop administrative privileges once the tunnel starts, as failing to do so poses a security risk. Setting the
nobodywill cause nobody, an unprivileged user, to take ownership over the stunnel process once the tunnel has been established.
setgid: Defines the Unix group ID under which stunnel will run. As with
setuid, this configuration specifies a group without any special privileges — nogroup — to avoid any potential security issues.
pid: Defines a file location where stunnel will create a
.pidfile, a type of file that contains a process’s PID.
.pidfiles are typically used by other programs to find the PID of a running process. By default, stunnel creates a
.pidfile in the
/var/run/stunnel4/directory but because the nobody user doesn’t have permission to access that directory, it will prevent the tunnel from starting correctly. Instead, this line specifies a file named
stunnel.pidheld in a directory called
pidsin the Ubuntu user’s home directory. We will create this directory and the
stunnel.pidfile shortly. As you add this line, be sure to change
sammyto the name of your Ubuntu system user.
debug: Sets stunnel’s debugging level, which can range from
7. In this example we’ll set it to
7, the highest level available, as that will provide the most detailed information if stunnel runs into any issues. You can set it to any level you like, but be aware that the default setting is
delay: When set to
yes, this option causes stunnel to delay the DNS lookup for the address listed in the
connectoption. It will also prevent stunnel from caching IP addresses. This setting will help to keep the tunnel open even if the managed Redis instance goes offline, as can happen when scaling up your cluster.
The remaining lines are service-level options, and only apply to the tunnel we’ll create for
[redis-cli]: This is a service name and specifies that the following lines represent an individual service configuration for a client program. You can have more than one service in a stunnel configuration file, though each must be associated with an existing client application and you can’t have two services for the same application.
client: Setting this to
yestells stunnel to run in client mode, meaning that stunnel will connect to a TLS server (the managed Redis instance) rather than act as a TLS server.
accept: Defines the host and port on which stunnel will accept connections from the client. Here, we specify the IP address
127.0.0.1, which is an IPv4 loopback address used to represent localhost, and port
8000. This means stunnel will listen for connections originating from the Ubuntu server on port
8000and encrypt them. Note that you can set the port to any port number you like as long as it’s not already in use.
connect: Defines the remote address and port to which stunnel will make the connection. Be sure to change this parameter to align with your managed database’s port and hostname or IP address.
Note: The hostname or IP address and port you should specify in the
connect directive will be specific to your own managed Redis database. These can usually be found in your cloud provider’s database management user interface where you provisioned your Redis instance.
If you’re using a DigitalOcean Managed Redis Database, you can find this information by going to your Control Panel and clicking on Databases in the left-hand sidebar menu. Then, click on the name of the Redis instance you want to connect to and scroll down to the Connection Details section. There, you will find fields describing your database’s host and port.
This is a fairly minimal configuration that leaves many of stunnel’s default settings in place. The program has many options available for you to create tunnels that suit your particular needs. See the official documentation for more details.
After adding this content, save and close the file.
Next, make sure you’re in your Ubuntu user’s home directory:
Then create the
pids directory which will hold the
- mkdir pids
The stunnel process will create the
stunnel.pid file automatically, so you don’t need to create the file yourself. However, you do need to change ownership of the
pids directory over to the nobody user and nogroup group:
- sudo chown -R nobody:nogroup pids/
Then, restart the
stunnel4 service so stunnel will read the new configuration file:
- sudo systemctl restart stunnel4
Following that, you can test whether stunnel successfully created a tunnel by invoking
netstat, a command line utility used to display network connections. Run the following command, which pipes the
netstat output into a
grep command, which in turn searches it for every instance of
- sudo netstat -plunt | grep stunnel
Outputtcp 0 0 127.0.0.1:8000 0.0.0.0:* LISTEN 17868/stunnel
This output shows that stunnel is listening for connections on local port
You can also confirm that the nobody user has taken ownership over the stunnel process with
ps, a program that displays all currently-running processes:
- ps aux | grep stunnel
Outputnobody 15674 0.0 0.1 121912 3180 ? Ssl 19:28 0:00 /usr/bin/stunnel4 /etc/stunnel/stunnel.conf . . .
Here, you can see that nobody has indeed taken over the stunnel process.
Stunnel is now fully configured and running on your system. You’re ready to connect to your managed Redis instance and test that the tunnel is working as expected.
Now that you’ve installed
redis-cli and configured stunnel on your server, you’re ready to connect to your managed database over TLS.
Based on the settings defined in the configuration file created in Step 2, you would connect to your managed database with the following command:
- redis-cli -h localhost -p 8000
This command includes the
-h flag, which tells
redis-cli that the next argument will be the host to connect to. In this case, it’s
localhost since we’re connecting to a tunnel created locally on the server. After that is the
-p flag, which precedes the port of the local tunnel we’re connecting to, which in this case is port
After running that command, you will be connected to your managed Redis server. Your prompt will change to reflect that you’ve connected and are in
redis-cli’s interactive mode:
Note: Oftentimes, managed databases are configured to require users to authenticate with a password when they connect. If your managed Redis instance requires a password, you can include the
-a flag in your
redis-cli command, followed by your password:
- redis-cli -h localhost -p 8000 -a password
Alternatively, you can authenticate by running the
auth command followed by your password after establishing the connection:
- auth password
If you’re using a DigitalOcean Managed Database, you can find your Redis instance’s password in the same place you found its hostname and port. In your Control Panel, click on Databases in the left-hand sidebar menu. Then, click on the name of the Redis instance you’ve connected to. Scroll down to the Connection Details section, and there you’ll find a field labeled password. Click on the show button to reveal the password, then copy and paste it into either of these commands — replacing
password — in order to authenticate.
You can test whether the tunnel is working as expected by running the
ping command from Redis’s interactive mode:
If the connection is alive it will return
If, however, stunnel is not tunnelling traffic from your server to your Redis instance correctly, you may see an error message like this before being disconnected from Redis:
OutputError: Server closed the connection
If you receive this or a similar error, double check that you’ve entered your Redis instance’s hostname and port correctly in your
stunnel.conf file. Likewise, be sure that you entered the correct port number in your
You may also receive this error when the managed Redis instance resets its open connections, as can happen when you scale up your cluster’s configuration. In such cases, you may not be disconnected from Redis, although you will need to reauthenticate.
Once you’ve confirmed that the tunnel is working, go ahead and disconnect from your Redis instance:
If you ever change stunnel’s configuration, you’ll need to reload or restart the
stunnel4 service so stunnel will notice the changes:
- sudo systemctl reload stunnel4
If at any point in the future you want to close the TLS tunnel, you can also do so with
- sudo systemctl stop stunnel4
After the tunnel has ben closed, you can reopen the tunnel by starting the service again:
- sudo systemctl start stunnel4
Now that you’ve successfully configured stunnel, you’re all set to begin adding data to your managed Redis instance with
Stunnel is a handy tool for creating TLS tunnels and establishing secure connections to remote servers. This is especially useful in cases where the secure transport of information between machines is critical, as with a remote database.
From here, you can begin exploring Redis and integrating it with your next application. If you’re new to working with Redis, you may find our series on How To Manage a Redis Database useful.
Thanks for learning with the DigitalOcean Community. Check out our offerings for compute, storage, networking, and managed databases.
This textbox defaults to using Markdown to format your answer.
You can type !ref in this text area to quickly search our full set of tutorials, documentation & marketplace offerings and insert the link!
Click below to sign up and get $200 of credit to try our products over 60 days!