Will the Below Configuration is a ideal configuration

Hi all,

Can Any one of you tell if the below configuration is an ideal configuration

we are Storing 3 tb data in 7 nodes having each node having 8 cores and 32 gb ram and 1 tb sdd.Will this be an ideal configuration to support 50 tps reads and writes and updates in cock roach db

As it suggested to use 500 gb hard disk per node.What happens if we use more.And Is the No of cores and ram size is ideal for the above scenario to support 50 tps

Hey @satya

We define the recommended hardware for CockroachDB in general on our documentation page here. We recommend 300-500 GB size storage in the documentation, as it’s more ideal to have more nodes to do distributed processing, spreading the data and the work load around, then to have very large stores on fewer nodes.

Please use this as a guide in order to determine your specific needs, and let me know if there are any other questions.

Cheers,
Ricardo

Thanks Ricardo for the quick response,
I have one question on your comment very large stores on fewer nodes means can i configure larger memory on few nodes for example can i configure few nodes with memory >=1 tb and remaining nodes having 500 gb. Will the performance be efficient.What will be the ideal cpu cores and ram configurations for same 500 gb ssd which is not provided in the documentation.What if i use hdd instead of ssd.Please clarify.

Hey @satya

We go through the required hardware specs for a node on our documentation, as well as best practices. It’s more ideal to have a higher number of smaller sized nodes then it is to have a few very large-sized nodes.

Unfortunately, we can’t go into specific configurations because workloads and demand could vary significantly, so this is something that needs to be determined by the user. We do state in the documentation that SSDs and NVMes should be used for your storage.

Let me know if you have any other questions.

Cheers,
Ricardo

Hi Richardo,

I have one query related to replication.We have 2 data centers.We want to maintain 10 nodes in first data center with replication factor 3.And want to maintain the replica for first data center in the second data center for disaster recovery.We dont want to go with single cluster across zones.what will be ideal way to over come this as we dont have 3 data centers to maintain the quoram factor greater than half if one data center is down for disaster management

Hey @satya

I would suggest configuring a change data capture to go from datacenter 1 to datacenter 2 to record the changes could be your best bet. We discuss the general concept and configuration here.

Let me know if you have any other questions.

Cheers,
Ricardo