active-orient icon indicating copy to clipboard operation
active-orient copied to clipboard

Connect to OrientDB Distributed Architecture

Open nishantkumar1292 opened this issue 8 years ago • 4 comments

Currently active-orient picks server details from a connect.yml file. The following lines takes the config in the activeorient.rb file.

username = eval(connectyml[:user])
password = eval(connectyml[:pass])
server = eval(serveryml[:server])
port = eval(serveryml[:port])

However I have setup OrientDB in a distributed mode. Is there a way active-orient can communicate with all the nodes in the distributed configuration?

nishantkumar1292 avatar Sep 13 '17 07:09 nishantkumar1292

Hi, How do you setup the distributed mode? What is required in addition? (Personally, I don't have experience with the distributed setup)

topofocus avatar Sep 13 '17 07:09 topofocus

Distributed mode is setup by running the file dserver.h and you need to mention the IPs of other nodes in the hazelcast.xml file in the config folder. More details here: http://orientdb.com/docs/2.0/orientdb.wiki/Tutorial-Setup-a-distributed-database.html

nishantkumar1292 avatar Sep 13 '17 12:09 nishantkumar1292

As I understand, that's how to setup the server.

For the purpose of ActiveOrient, the differences in terms of the Connection to the server are relevant. From the perspective of a client, the distributed setup should make no difference. Therefor I asked for necessary changes in the connect-setup.

Have you tried to connect to a distributed database cluster?

If there is an error-message, we might have a starting point.

Am 13.09.2017 um 14:14 schrieb Nishant Kumar:

Distributed mode is setup by running the file dserver.h and you need to mention the IPs of other nodes in the hazelcast.xml file in the config folder. More details here: http://orientdb.com/docs/2.0/orientdb.wiki/Tutorial-Setup-a-distributed-database.html

topofocus avatar Sep 13 '17 12:09 topofocus

I have connected to the OrientDB server and there seems to be no error. But I don't think other nodes are used much. I have been parallel reads and writes on the database. But the RAM and CPU is consumed only on the sever connected. The other node only replicates in regular intervals. No resources are used here. Should the distributed architecture not be handled on the client end to make full use of the resources.

nishantkumar1292 avatar Sep 13 '17 13:09 nishantkumar1292