Several instances of Couchbase on the same node

Hello Masters,

Is it possible to setup several Couchbase instances on the same node/host/machine?

Your direction would be appreciated.

Thanks,
Erwin

For development needs, you can use cluster_run script, which separates instances by using different ports.

For example, this is how you can spawn 4 nodes of branch 2.5 on ports 9000-9003:

$ repo init -u git://github.com/couchbase/manifest.git -m rel-2.5.0.xml
$ repo sync
$ make
$ cd ns_server
$ ./cluster_run --nodes 4

Another option is probably use docker to run several instances on the same physical node. Dockerfile from Dustin https://gist.github.com/dustin/6605182

In the docker index: https://index.docker.io/u/dustin/couchbase/

I don’t think it is going to be ready for production in any time. Couchbase designed to join different machine into the cluster to get more capacity/speed. On the single node, it can use whole power, and actually it is multi-threaded internally, so you don’t have to worry about consuming CPU and disk performance of the node.

Could you describe your use case BTW?

Thanks for fast turnaround response.
As I realized it’s not production ready.

Sure.
Actually, we are using cache system atop of JVM. We have multiple JVMs (cache container) on multiple nodes.
We have a very powerful machines with more than 12 cores and more than 96 physical memory.
So we would like to consume the power of the machine, that is, cores and memory.

Hi Erwin,

With our recent release Couchbase Server 2.5 we have now a new enhancement i.e. ‘installing multiple Couchbase instance on a single machine’. In our documentation (links below) you’ll find detailed instructions for installing multiple instances of Couchbase on a single dev box using user-defined ports.

Regards,

Anil Kumar

Hello Anil, I’ve tried to follow the instructions to have several instances on same node however I have a problem referenced here:

Do you have any idea how this situation can be solved?

Couchbase does not officially support multiple instances on the same host and it believes there can be issues.