4.3 KiB

Raw Blame History

New Features in ArangoDB 2.0

@NAVIGATE_NewFeatures20 @EMBEDTOC{NewFeatures20TOC}

Features and Improvements

The following list shows in detail which features have been added or improved in ArangoDB 2.0. ArangoDB 2.0 also contains several bugfixes that are not listed here.

Sharding

ArangoDB 2.0 can be run in a clustered mode, allowing to use multiple database servers together as if they were a single database.

This is an optional feature that must be turned on and configured explicitly. More information about the setup and the design decisions of ArangoDB's sharding feature can be found in the @ref Sharding "Sharding manual".

ArangoDB can still be used as a single server database.

New Web Interface

The web interface has been refurbished in ArangoDB 2.0. It now also allows planning and managing clusters visually, and comes with built-in user management, available from via the "Tools" menu item.

Changed shape storage

Prior to ArangoDB 2.0, the structure of documents stored (called "shapes" in ArangoDB lingo) was saved in separate collections (the "shape collections"). The separate storage of structural information and the actual document data caused overhead when a document with a new shape was saved. In this case, two writes were required (one into the shape collection, and for the actual document data).

ArangoDB 2.0 now saves the shape information inside the same collection as the document data goes into, requiring less writes and potentially less space.

This saves up to 2 MB of memory and disk space for each collection. The actual savings depend on the number and size of shapes per collection, but the following rule of thumb applies: the less different shapes there are in a collection, the higher the savings. Additionally, one less file descriptor will be used per opened collection.

AQL Improvements

The following list functions have been added to AQL:

NTH: returns the nth element from a list
POSITION: returns the position of an element in a list
SLICE: extracts a slice of a list

Miscellaneous Improvements

the db object on the server now allows direct access to collections whose names start with an underscore. This makes accessing system collections on the server easier, for example from a Foxx action:

In 1.4, accessing a system collection on the server required this:
```
db._collection("_user").<whatever>
```
whereas in 2.0, a system collection works like this everywhere:
```
db._users.<whatever>
```
Added coffee script support for require:

Scripts written in CoffeeScript can be required and executed in arangosh and arangod. Scripts must reside inside the modules path in order to be found, e.g.:
```
unix> cat js/common/modules/add-example.coffee
add = (a, b) -> a + b
exports.add = add

arangosh> require("add-example");
{ 
    "add" : [Function "(a, b)"] 
}
```

Added JSON support for require:

Files containing JSON objects can be required from arangosh and arangod. Files must reside inside the modules path in order to be found, e.g.:

unix> cat js/common/modules/my-config.json
{
  "app": {
    "name": "foo"
  }
}

arangosh> require("my-config");
{ 
  "app" : { 
    "name"  : "foo"
  }
}

Added startup option --server.reuse-address for arangod:

This option can be used to control whether sockets should be acquired with the SO_REUSEADDR flag.
Added startup option --javascript.gc-interval for arangosh:

This option can be used to control the amount of garbage collection that is performed by arangosh. For example, to trigger the garbage collection inside arangosh (not arangod) after each 5 commands, use

unix> arangosh --javascript.gc-interval 5

@BNAVIGATE_NewFeatures20

4.3 KiB Raw Blame History