Couchbase Server Implementation

Hello everyone,

My name is Luis and i’ve started recently learning about Couchbase, myself and a couple of friends started a project / startup company and we are developing a Saas that combine a web application and a mobile one.

Couchbase is the perfect solution for our server and mobile database and it fits extremely well in our project.

We are testing it in a old linux server we have and the performance is great.

We have just a few questions and i hope that someone can help us :

I have read that the minimum node number for production is 3 nodes, for us this is a little bit hard to acomplish because we are funding the project with our own money.

What are your opinion about start the production with only one server with this specs :

Intel Xeon E3
3,4 / 3,8 GHz
32 GB RAM DDR3 ECC 1600 MHz

And with more customers we will add more nodes.

Sorry about my english,
Best Regards,
Luis Pinto

HI, Luis,

Good to hear that you are interested in using Couchbase and I’m happy to answer this question.

The reason why we recommend a minimum of 3 nodes is because we want the system to be able to sustain node failures. Although you can run your prod system using 1 node, you will lose the High Availability / Replication benefits of the system.

There are a few things you can do to mitigate this issue:

  1. Regular backup your data
  2. Use XDCR feature to replicate your data into another data center

Please let me know if you have any more questions


1 Like

Dear Qi,

Thanks alot for your reply, couchbase is really a new think for us and we have a lot of questions ofcourse we are trying to figured out ourselfs the mosto of them.

But i have one more question if you allow me.

As i said we are devoping a Saas and we need to limit the disk space per ‘company account’, we are using one bucket and in every document we have the tenantID so we can calculate the disk space used for all documents of a spefic Tenant.

We tried the method above, but it was very slow if you have alot of documents, we are tsting with 1.000.000 :

What you think about this aproach, save in every document its size in a variable and sum all of them in a view.

Or there is some other aproach we dont know ?

Thanks alot for the feedback.
Best Regards

View is a good candidate for this. I like this approach.


1 Like