Create a file that pretends to be a database query
Project description
Create a file that pretends to be a MongoDB collection. If you try to open the file, it will JSONify the collection and dump it back to you.
Why?
Because our enterprise database is file based, but our cloud based database isn’t.
…no really why?
Files are simpler and don’t break. The database operations we need to perform aren’t complicated and don’t require high performance. By using files there isn’t a database that can fail. There’s the added benefit that if customers want to use different types of databases for distributed systems, we can plug into any of them using a filedb (we just have to make something that dumps out JSON .
Quickstart
Install FUSE
Make sure you have FUSE installed.
Ubuntu: sudo apt-get install g++ libfuse-dev
Get mongo running
And you’ll need a collection with some data in it.
$ mongod$ mongo
> use test
> db.people.insert({"username": "greg"})
> db.people.insert({"username": "phil"})
> db.people.insert({"username": "bob"})
> db.people.insert({"username": "samantha"})Install filedb
$ pip install filedbRunning via python
$ filedb /tmp/tutorial/mnt/ mongodb://localhost:27017/test peopleOpen the file
$ cat /tmp/tutorial/mnt/db
[{ "_id": "52f104a7a0fb769e0cd0d1d4", "username": "greg" }
{ "_id": "52f104a9a0fb769e0cd0d1d5", "username": "phil" }
{ "_id": "52f104aca0fb769e0cd0d1d6", "username": "bob" }
{ "_id": "52f104aea0fb769e0cd0d1d7", "username": "samantha" }]Running as a service
Upstart job.
# install the job
$ sudo cp -R overlay/* /
# or
$ wget https://raw2.github.com/yhat/filedb/master/overlay/etc/init/filedb.conf
$ sudo mv filedb.conf /etc/init/filedb.conf
# start the job
$ sudo start filedb
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
