Although it may have been published to the npm
repo, this project is still
considered experimental. It has not been widely deployed, and I am in the
process of rolling it out for personal use and within my company.
As with any alpha-stage storage technology, you MUST expect that it will eat your data and take precautions against this. You SHOULD expect that its APIs and storage schemas will change before it is labelled stable. I MAY respond to bug reports but you MUST NOT expect that I will.
Per the MIT license, usage is entirely at your own risk.
reStore RemoteStorage server written for Node.js. It is designed to be compatible with RemoteStorage.js from version 0.6 onwards, covering versions RemoteStorage-2011.10, RemoteStorage-2012.04, and draft-dejong of the protocol.
$ npm install restore
The following Node script will run a basic server:
process.umask(077);
var reStore = require('restore'),
store = new reStore.FileTree({path: 'path/to/storage'}),
server = new reStore({
store: store,
http: {host: '127.0.0.1', port: 8000}
});
server.boot();
The host
option is optional and specifies the hostname the server will listen
on. Its default value is 0.0.0.0
, meaning it will listen on all interfaces.
The server does not allow users to sign up, out of the box. If you need to allow
that, use the allow.signup
option:
var server = new reStore({
store: store,
http: {host: '127.0.0.1', port: 8000},
allow: {signup: true}
});
If you navigate to http://localhost:8000/
you should then see a sign-up link
in the navigation.
In production, we recommend that you restrict access to the files managed by your reStore server as much as possible. This is particularly true if you host your storage on a machine with other web applications; you need to protect your files in the event that one of those apps is exploited.
You should take these steps to keep your storage safe:
- Pick a unique Unix user to run your server process; no other process on the box should run as this user
- Do not run other applications as root, or as any user that could access files owned by your reStore user
- Use
process.umask(077)
as shown above so that the server creates files that can only be accessed by the process's owner - Make sure the directory
path/to/storage
cannot be read, written or executed by anyone but this user - Do not run reStore as root; if you need to bind to port 80 or 443 use a reverse proxy like Apache or nginx
- Ideally, run your storage inside a container or on a dedicated machine
If you're using the Redis backend, apply similar access restrictions to the database and to any files containing the database access credentials.
Since RemoteStorage is a system for storing arbitrary user-specific data, and since it makes use of OAuth 2.0, we recommend you serve it over a secure connection. You can boot the server to listen for HTTP or HTTPS requests or both. This configuration boots the app on two ports, one secure and one plaintext:
var server = new reStore({
store: store,
http: {
host: '127.0.0.1',
port: 8000
},
https: {
force: true,
host: '127.0.0.1',
port: 4343,
key: 'path/to/ssl.key',
cert: 'path/to/ssl.crt',
ca: 'path/to/ca.pem' // optional
}
});
server.boot();
Note that you should not run reStore as root. To make it available via port 80 or 443, use Apache, nginx or another reverse proxy.
The force: true
line in the https
section means the app will:
- Return HTTPS URLs in WebFinger responses
- Force sign-up and OAuth login pages onto an HTTPS connection
- Refuse to process POST authentication requests over insecure connections
- Block insecure storage requests and revoke the client's access
reStore considers a request to be secure if:
- reStore itself acts as an SSL terminator and the connection to it is encrypted
- The
X-Forwarded-SSL
header has the valueon
- The
X-Forwarded-Proto
header has the valuehttps
- The
X-Forwarded-Scheme
header has the valuehttps
So you can have an SSL-terminating proxy in front of reStore as long as it sets
one of those headers, and does not let external clients set them. In this
setup, you can set https.force = true
but omit https.port
; this means
reStore itself will not accept encrypted connections but will apply the above
behaviour to enforce secure connections.
reStore supports pluggable storage backends, and comes with two implementations out of the box:
reStore.FileTree
- Uses the filesystem hierarchy and stores each item in its own individual file. Content and metadata are stored in separate files so the content does not need base64-encoding and can be hand-edited. Must only be run using a single server process.reStore.Redis
- Stores data in a Redis database, and all stored data is base64-encoded. It can be run with any number of server processes.
All the backends support the same set of features, including the ability to store arbitrary binary data with content types and modification times.
They are configured as follows:
// To use the file tree store:
var store = new reStore.FileTree({path: 'path/to/storage'});
// To use the Redis store:
var store = new reStore.Redis({
host: 'redis.example.com', // default is 'localhost'
port: 1234, // default is 6379
database: 2, // default is 0
password: 'unhosted' // default is no password
});
// Then create the server with your store:
var server = new reStore({
store: store,
http: {port: process.argv[2]}
});
server.boot();
(The MIT License)
Copyright (c) 2012-2015 James Coglan
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.