In About Aerospike it writes "The free Aerospike Community Edition manages up to 200G of unique data and can be upgraded to the Aerospike Enterprise Edition…" So, what does “200G of unique data” mean? Here is comparison between Community and Enterprise versions, so it seems the Community edition is just for try-out-and-decide, it is stripped off all key functionalities a big data management system must have. I just wonder if there are use cases with Community edition. What can the Community edition be used for?
You have a link for that quote?
That is from 2015. You can have upto 32 nodes on CE. Differences between EE and CE: EE has following additional features which really give a enhanced/better cluster management in production:
- 24/7 Support
- Fast Restart
- Rapid Rebalance
- Cross Datacenter Replication (XDR)
- IPv6 Support
- Durable Deletes
This is stale, and thanks for pointing that out. You have the link to the product matrix, which includes the feature comparison. The effective limits on CE cluster sizes is 32 nodes. There is no size limit imposed on CE. There is currently an effective limit of 2^32 objects per-namespace, per-node. Depending on your object size, you can figure out what that means in terms of GB. There’s no license limit on the data size you put in CE.
Thanks for answers. I am new to Aerospike (but have a strong cloud-computing background) .Currently I have Installed an instance of Aerospike on a Ubuntu server using Virtual Box. I am just experimenting, so I want to create a small cluster of four or five nodes and try out the Aerospike as a distributed system. I want to play around with namespace configurations like persist data to SSD, clustering, replica configurations, and so on… Only then I am planning to explore APIs. What do you think, is it possible with a CE and Virtual Box nodes? Any recommendations? BBTW with CEPH it worked.
I have a 2-3 node cluster running on my laptop over VMWare Fusion. I use Vagrant and Vagrant Manager to manage the cluster. Do not expect any realistic performance from such a virtual environment.
With VMs, you will have to use “file” device for persistence. Don’t think you can do SSD raw block storage which is what Aerospike gives best “hybrid memory” performance with. Also, you can run multiple Aerospike processes on the same VM and form a cluster as long as you separate out the ports in the config file - just for testing. 300x on first, 600x on next, 900x on third etc. (See Aerospike Command Line Options and run multiple asd’s with different config files from command line. ) So for getting started … you can do all this, don’t try to infer performance numbers though!
Sure, I am not going to do any performance measuring, I just want to learn Aerospike basics and try APIs.