Temperature Database Prototyping – Hardware, Postgresql and HTSQL


In this post I am going to detail prototyping a test Postgresql database server, installing Postgres support on the Raspberry PI (for HTSQL) and then serving up the contents of the DB server via HTSQL over http.

The purpose of this task is to familiarise ourselves with each of the components in the system before we design how the actual database server will be setup and built out.

Throw away test machine

I like to keep a couple of old Windows machine on hand for building out quick test servers that I can use to prototype a system and then wipe out ready for my next project. I find this saves a lot of cost of going out and buying hardware, only to discover your original plan has changed for some technical/design reason and then having to purchase more equipment.

For this task I am using an old Pentium 4 machine running XP with about 60 Gigs of HDD space. I’ve hooked it up with a monitor and keyboard and plugged it into my router via a network cable. At this point it is worth grabbing the IP address of the machine you have just plugged in. This should be available via your routers DHCP table.

Consult your router documentation on how to access this information if you are not familiar with the process. Note: I checked my DHCP table after installing Ubuntu – since this is the only Ubuntu machine on my network currently, it appeared as “ubuntu” in the table

Now my first step is going to be to set it up as a dual-boot machine and load Ubuntu desktop edition onto it.

Ubuntu setup

Depending on the type of machine you use for prototyping the instructions below will obviously be different, for those of you with an old XP machine at hand, the steps below should work for you.

  1. Downloaded Ubuntu’s latest version windows installed here.
  2. Run the installer and follow the on-screen instructions – Note: select your OS type as Ubuntu.
  3. Once your machine has rebooted into Windows follow the next steps:
  4.  Right click My computer and select Properties
  5. Select the Advanced tab from the pop-up box
  6. Under Start-up and Recovery select Settings
  7. Select Time to display list of operating systems checkbox – change the drop-down to 10 seconds
  8. Select the Default operating system drop-down and change this to “Ubuntu”
  9. Save your changes and re-boot your machine into Ubuntu.

Setting up Postgresql

Our Ubuntu machine is going to act as a database server, a Postgresql one in fact. For those of you not familiar with Postresql, it is a powerful, free, open source relational database and has been around for 15+ years. You can read more about it here.

Your first step will be to install the necessary software on Ubuntu.

Follow the instructions below to set up Postgresql

1. Open Terminal in Ubuntu and type the following command:

sudo apt-get install postgresql postgresql-client postgresql-contrib

2. The above command once it has completed running will have installed all the necessary components to start a Postgresql server on your Ubuntu machine.

3. Navigate via the command line to the following directory:


Note: depending on the version of Postgres you have installed the directory 9.1 may be different. As of August 26, 2012 9.1 is the latest version installed via the apt-get command above.

4. You will now need to edit the postgresql.conf in order to tell Postgresql to listen on the IP address assigned to your new dev machine via DHCP. Using vi type:

sudo vi postgresql.conf

Once you have the file open, edit the following line (you may have to un-comment it, and it will probably be set to localhost):

listen_addresses = 'ip address from dhcp table'

Save the file and exit.

5. We can now login to the “template1” DB that comes setup by default with pgsql

sudo -u postgres psql template1

From here for example we can change the password of the postgres account:

ALTER USER postgres with encrypted password 'my_new_password';

6. We now want to create a test database to serve via HTSQL, to do this we type the following:


7. With our new database created we can add a dummy table and a dummy row of data:

CREATE TABLE temperature (code char(5) );
INSERT INTO temperature (code) VALUES ('12345');

Feel free to add some more test rows.

8. Now we can quit postgresql and edit the configuration in order to allow remote access to the DB. In order to do this, you will need to edit the pg_hba.conf file

Change the configuration to include the following (you can use Vi):

local   all         postgres                          md5
host   all         postgres              samenet            md5

Then save and exit

9. We now need to restart our postgresql instance to pick up the changes, this can be done by:

sudo /etc/init.d/postgresql restart

Once the instance has restarted, we are now ready to make some configuration changes on the Raspberry PI.

Raspberry PI configuration

In order to connect our HTSQL server to our newly built Postgresql DB we need to install some extra packages on the Raspberry PI. First of all you will need ssh into the RPI.

Once logged in, follow the steps below:

1. Install python developer packages and the HTSQL postgres package as follows:

sudo apt-get install libpq-dev python-dev
pip install HTSQL-PGSQL

Once these have successfully been installed you should now see:

Successfully installed HTSQL-PGSQL psycopg2

Now we can connect to the postgresql instance.

2. We can test our connection to the testtemp database we created as follows:

htsql-ctl shell -p pgsql://postgres@host:port/database

The host above should be replaced with the IP address of the Ubuntu server you setup with postgresql installed on it. Also remember to include the port number after the IP address, by default postgresql runs on 5432. The final option is the database we created – temptest.

Once you can log in via the HTSQL shell to the DB instances then you are good to go with actually running a server.

3. Exit out of the HTSQL shell and then create a HTSQL server as follows:

htsql-ctl server 'engine://username:password@host:port/database'

engine should be replaced with pgsql, username and password are those that you set on the DB server, host and port are the DB servers IP address and postgresql port (as mentioned above this will probably be 5432) and finally database is our temptest database.

Once you have started the server you will see:

Starting an HTSQL server on raspberrypi:8080 over temptest

4. We can now check that everything is running as expected. Load up your web browser and in the URL bar type:  http://<ip of raspberry pi>:8080

If everything is working as expected you will see

Welcome to HTSQL!

Please enter a query in the address bar.

5. You can then hit the test table we created as follows:

http://<ip of raspberry pi>:8080/temperature

And see some test data (note I have added a few extra rows to my temperature table of random test data)

1 12345
2 1245
3 56789

6. And that’s it, you now have HTSQL running on the Raspberry PI, pointing to a database server.


The above demonstrates the process of setting up a database server and using HTSQL on the Raspberry PI to communicate with it.  However this is only the beginning of what is required in order to build a system that can record our thermostat temperature data.

Using a prototype is useful for the purposes of ensuring we are building a modular system, where each component can be swapped out as necessary. Now we have demonstrated that we can plug a database into the system, we can concentrate on sourcing the right hardware that will replace the prototype.

Of course whilst our Raspberry PI now runs HTSQL, there is still a lot of work to be done on that end of things. We need to configure the RPI so that it is secure and also build out our controller software – which will use the data stored in the Postgres DB to tune our heating system.

The next steps then are going to be to complete the setup of the Raspberry PI and use our prototype DB server to test this configuration against. Once this is complete we will then concentrate on designing a relational database for the heating data, and building out the server hardware.

Installing HTSQL on RPI


As I have mentioned in earlier posts I am planning on using HTSQL as part of my heating system. It will primarily provide me with a way to query data stored from the controller and thermostats over HTTP.

Eventually it will be pointing to a PostgreSQL DB which will act as my area for data storage.

Below are instructions on how to install HTSQL on your Raspberry PI.

Installation instructions

Note: The instructions presented are for installing HTSQL on a RPI running Debian squeeze. If you are looking for the Mac instructions they are located here. For other flavours of Linux available on the RPI the instructions should work, I make no promises though!

As always, you can check http://htsql.org/ for up-to-date information on operating systems supported and instructions.

Anyway on with the installation process:

The Raspberry PI runs on an ARM based architecture. On last checking the htsql website, there is no support for this via apt-get.

You can check the architecture of your chipset (which of course will be ARM if you are following the steps for your RPI) as follows:

apt-config dump

You will see in the returned information:

APT::Architecture "armel"

If you attempt to follow the standard apt-get procedure as presented on the htsql website for Debian, you will run into the following error message:

E: Unable to locate package htsql

This is because there is currently no armel distro.

So in order to install HTSQL, you will need to install via source.

Install pip

In order to run pip (a python package installer) you will need to add it to your RPI.

Run the following command :

sudo apt-get install python-pip

Once this is complete you can then run:

pip install HTSQL

Once this is complete you can now type:

htsql-ctl version

You should now see:

HTSQL 2.3.1 (or whatever the latest version is you downloaded).

HTSQL is now ready to go!


Out of the box HTSQL works with sqlite, if you want to run it against other databases you can follow the instructions here.

Follow up posts to this will detail setting up a PostgreSQL DB server and then connecting to it via HTSQL.

Raspberry PI – Network access


One of the benefits of having a machine like the RPI is that you can set it up on your home network and access it without having to worry about a keyboard, mouse or monitor.

Combined with the RPIs small size, and the ability to place it in a custom case with a wall mount, these features are what make it one of the most exciting tools for home embedded/smart systems.

The steps below are an introduction on getting your RPI to boot up with ssh access, and allow you to connect to it from another machine on your home network.


So what is ssh? In brief ssh (secure shell) is a method that allows us to log into other machines over the network, most commonly through a command line tool (for those of you with a Mac you can do this via the Terminal). Once logged in, depending on our permissions we can then perform actions on the machine we are connected to.

For a deeper more complex technical evaluation of ssh you can read the wiki article here.

For the purposes of configuring our RPI (and later using it as a controller) we are going to be setting up ssh access so we can work on the machine remotely.

In my specific case, this is born from the fact that my router is nowhere near my TV set so trying to configure the RPI with a monitor, but without network access is going to be impossible in some cases i.e. installing HTSQL.

Steps to set up 

Fist of all connect your RPI up to its power supply, monitor and keyboard. Then log in.

Raspberry PI login

Raspberry PI login

At this point you may want to set the date of your RPI if you haven’t done so already.

By default the keyboard is set up to be UK format, so the @ is actually the ”

To set the date:

sudo date –set “23 APR 2012 16:01”

Next you will need to set the RPI to start ssh when it boots. To do this follow the next steps:

  1. cd /boot/
  2. sudo cp bootenable_ssh.rc boot.rc

The above copies bootenable_ssh.rc and creates a second version called boot.rc. The boot.rc version is then loaded when you start your RPI up.

Now follow the next steps:

  1. cd /etc/network
  2. vi interfaces

By default vim isn’t installed on the RPI, so if you haven’t taken the step to install it, you’ll need to use vi to edit the interfaces file.

You’ll need to edit the settings in the interfaces file as follows:

  1. Add eth0 to this line: insert eth0 so : auto lo eth0
  2. Change iface eth0 inet dhcp  to  iface th0 inet static
  3. Now add:

Finally save the file  – press  : then type x and hit enter to save and exit
The above steps have now added configuration to tell the RPI to run on a static IP address on your local network.

Testing the above worked

You can now log out of your RPI and power it down.

Next, boot the PI back up with your network jack attached. Your PI should now boot with ssh enabled.

Raspberry PI hooked up

Raspberry PI hooked up

I plugged my PI directly into my Cisco E4200 router. I then logged into it via in my web browser (type into the address bar) – note this may be different on your home network depending on how you have configured it.

If you have a Windows machine for example and your router is plugged into that, you can run:


Now look for Default Gateway, this will give you the IP address that the router should be running on.

Once your PI is powered up and booted (you should also see a yellow light, indicating you are on the network) you can shell (ssh) into your server.


The fixed IP address you assigned may not work. To test if it did try:

 ssh pi@

If this does not work, log into your router and check the DHCP Client table. Most of the devices I run on my network have a client name e.g. kindle-xxxxxxxxxx.

You may now see a device appear in the list with no name. To confirm which device is your RPI, you can try disconnecting it, refreshing the DHCP client table, reconnecting it and then refreshing again. The new device added to the list is probably your RPI.

Next, take the new IP address from the table and shell into the machine (via Terminal, console, Cygwin or whatever you are using):

ssh pi@<new ip>

This gave me the authenticity message which I accepted. I then entered the default password for the PI, and presto! I am now logged in.


So I need to investigate how to ensure the static IP is always picked up, I’ll blog a separate post on that for those of you having problems.

For the moment we now have an easy way to get access to the PI without a monitor and can go ahead and install whatever packages we need.

Quick update – number 2

With summer being here, things have been a little slower on the development/soldering front. However as we approach fall this will change.

Here is an update of what is happening.

3D printing co-operative

This evening the first meeting of our 3D printing co-operative will take place. Now we have ironed out the issues with the MakerBot, the rest of the co-operative is ready to start using it. I’ll be posting the occasional picture up here of any cool objects my colleagues make.

Instructional Videos

As part of the inauguration of the 3D printing co-op we will doing some training sessions on the Replicator. I will try to film some of the important aspects of this (changing material spools, changing the base plate heat etc.) and adding links to these on the blog.

Ordering Materials

I still have some outstanding components to order for my Thermostats, I’ll be sorting these out next week. Once I have them all, expect a post listing each piece, what it does and how it fits together as part of the overall system.

Raspberry PI remote access/ssh

Expect two separate posts, one on setting up the Raspberry PI to be accessed over your home network (so you don’t need a TV/Monitor) and the second on installing the latest version of HTSQL. (Update: these can be found here)