Skip to content

Latest commit

 

History

History
105 lines (84 loc) · 9.52 KB

db-on-digital-ocean.md

File metadata and controls

105 lines (84 loc) · 9.52 KB
Raw

Hosting your database on Digital Ocean

Postgres on Digital Ocean Ubuntu droplet

Install Postgres
  • Run sudo apt-get update.
  • Run sudo apt-get install postgresql postgresql-contrib.
Create a database
  • Run sudo -u postgres createdb [db_name] to create a new database owned by the postgres role. For example, sudo -u postgres createdb wonderapp will create a database called "wonderapp" for the postgres user.
  • For more Postgres options and commands, including how to create and edit tables through psql commands, see these Digital Ocean Postgres docs.
  • To simply copy a pre-existing local (or hosted-elsewhere) database to your newly created Digital Ocean database, see pg_dump instructions in the Copy and transfer database section below.
See all databases on droplet
  • Inside your droplet, run psql to enter a Postgres session.
  • Run \l to list database information.

Copy and transfer database

Whether you have a local database you want to copy to your server (e.g., local to Heroku) or a hosted database you want to move to a different server (e.g., Heroku to Digital Ocean), the pg_dump command is your friend. This command is installed with Postgres and can quickly copy over the contents of one database to another.

pg_dump commands and options

The exact format you use for each part of your pg_dump command will depend a bit on where the database is and whether it is the source or the destination. Note that all these commands are done from your local computer, not from inside your server.

  • When using the pg_dump command, you might consider using the -O and -c options with it. Use -O (shorthand for --no-owner) to forego any commands to set ownership to match the original database. Use -c (shorthand for --clean) to clean (drop) the destination database's objects before dumping in the new ones. For more information about options, see the docs.
  • To copy a local database, just include the database name and any options: 
    pg_dump -Oc wonderapp
  • To copy a Digital Ocean database, use ssh to log in, -U to specify a user, and -h to specify the host: 
    ssh [email protected] "pg_dump -Oc -U anna -h localhost wonderapp"
    The command above will be followed by a prompt for the droplet user's password. Alternatively, if configured your user's SSH login and added the SSH password to the ssh-agent keychain (see the Additional SSH login options section from the README), you do not need the user or host options and can just use something like this: 
    ssh anna "pg_dump -Oc wonderapp"
  • If you enter any of the pg_dump commands above, the command line interface will output the database for you. You can use these commands to see the database right in the CLI, but this is only half of a typical pg_dump use case. Usually you want to copy that content either straight into another database or into a text file as a database backup.
Pipe contents from one db to another

Use the | (pipe) symbol to pipe one database's content into another's. Here are some examples.

  • Local to Heroku: 
    pg_dump -Oc wonderapp | heroku pg:psql -a wonderapp
    Here, -a is an option for naming your Heroku app. Before this pg_dump command works, you may need to use heroku login to log in to your Heroku account (as mentioned in the Possible issues subsection below).
  • Local to Digital Ocean (with "anna" user custom SSH login and stored SSH password (see the Additional SSH login options section from the README)): 
    pg_dump -Oc wonderapp | ssh anna "psql -d wonderapp"
  • Digital Ocean to local (logging in as root with no custom SSH login or stored SSH password): 
    ssh [email protected] "pg_dump -Oc -U anna -h localhost wonderapp" | psql wonderapp
Make database backup

It can be a good idea to make a backup file of your database that you can store on your own computer. If you ever needed it, you could use it to dump into a new hosted database. Use the > (greater than) symbol.

  • Create a blank backup SQL file on your Desktop with touch ~/Desktop/db_bkup.sql.
  • Dump the contents of a database into this file using pg_dump and >. Instead of piping the contents into another database with a |, you are using > to write the contents into your blank file, overwriting any pre-existing content in that file.
  • An example of backing up a local database:
    touch ~/Desktop/wonderapp_bkup.sql
    pg_dump -Oc wonderapp > ~/Desktop/wonderapp_bkup.sql
Possible issues
  • Heroku authentication: Before you can dump into a Heroku database, you may need to log in from the command line. Run heroku login and enter the Heroku account email and password. Once authenticated, you should be able to run the pg_dump command with the Heroku database.
  • pg_dump command not found: You may try to use the pg_dump command only to have your command line editor say something like "command not found," even though you are sure you correctly installed pg_dump. If so, you may need to find out where it was installed and point your command line editor to that path using your .bash_profile or .bashrc file. Try the following:

    • If you know exactly where pg_dump was installed, copy the full path.

    • find command: If you do not know where pg_dump is installed, use find to search for pg_dump.

      how to use find command
      • The find command includes a path, options, and a search expression. For example, if you thought pg_dump was installed in your /Applications folder and wanted to search for pg_dump by its name, you might try find /Applications -name pg_dump (where -name is the search-by-name option).
      • If you have no idea where pg_dump was installed, you might try simply find / -name pg_dump 2>/dev/null to search your entire root folder (since you are searching all your folders, use the 2>/dev/null command to suppress errors, limiting your search to more useful results).
      • Find the correct path to pg_dump in the search results and copy it. Here is an example of a search result showing where pg_dump might be installed:
          /Applications/Postgres.app/Contents/Versions/9.5/bin/pg_dump

    • Once you have copied the path to pg_dump, open the .bash_profile or .bashrc file in your home folder (e.g., if you use VS Code, type open -a Visual\ Studio\ Code ~/.bash_profile). Your command line editor uses this file to configure your command line. In the file, on a new line, type

      export PATH="/Path/To/Folder/PgDump/Is/In/:$PATH".

      Point the path to the folder pg_dump is in, not to pg_dump itself, and make sure to include export PATH= and :$PATH on either end. Once your .bash_profile has the updated PATH code, save and close this file.

    • Open a new command line window (your CLI only notices changes to the .bash_profile when opening a new window) and try the pg_dump command again.

Access Digital Ocean database from local computer using SSH tunnel

If you hosted your database on Digital Ocean and are noticing difficulty connecting to it from your local computer (e.g., through a GUI on your computer like PGAdmin or SQL Tabs), try an SSH tunnel to securely connect to the remote database.

  • To create a secure tunnel, run ssh -N -L [local_port]:[local_host]:[remote_port] [remote_host] in the command line. For example, if the droplet is "anna" and the database is a Postgres database running on 5432, we might run ssh -N -L 5555:localhost:5432 anna to create a secure tunnel and forward the local port 5555 to the remote port 5432, encrypting it in the process. For more on this, see this article or this article.
  • Connect to the database using localhost and a local port to listen for. For instance, in the situation above, we could use a GUI to connect to the database on anna using port 5555 on localhost.
  • Keep in mind the tunnel only stays open as long as the command line window is open where the tunnel command was entered. Closing the window will disconnect you from the database.

Change your user password:

Do the following to change your droplet user's password. Note this is the password for a user on the droplet, not the password to log in to the droplet using ssh. To change that password, see the Additional SSH login options section in the README.

  • Log into your droplet with ssh.
  • Run psql to start a Postgres session.
  • Run \password. You will be prompted to enter the new password twice.