Master object distribution

Should I expected master objects to be evenly distributed amongst all nodes in a cluster? Or is it normal to have an uneven distribution? If it’s expected, and that isn’t the case for me, what can I do ?


If you are curious, here is what ChatGPT said :slight_smile:

With Aerospike 4.9, it is expected that the master objects are evenly distributed among all nodes in a cluster. However, in certain situations, an uneven distribution can occur.

The uneven distribution of master objects could occur due to several reasons such as node failure, network partitioning, data distribution skew, etc. It is important to identify the root cause of uneven data distribution before taking any corrective actions.

To address uneven data distribution in Aerospike 4.9, you can consider the following actions:

Run the “asinfo -v ‘tip:host=<IP_address>;port=’” command on all nodes to check for any connectivity issues between nodes. Check for any hardware or network issues that could cause performance degradation or data skew. Run the “asadm” tool to analyze the data distribution and check for any skewness. The tool provides various statistics and insights into the Aerospike cluster, which can help in identifying the issue. Use the “rebalance” command to redistribute data evenly across nodes in the cluster. The rebalance command can be executed through the Aerospike Management Console (AMC) or through the asadm tool. If the above steps do not resolve the issue, consider contacting Aerospike support for further assistance. They can provide additional guidance and help to diagnose and resolve the issue.

If you are curious, here is what ChatGPT said :slight_smile:

Doesn’t seem like good advice to me :slight_smile:

If you are using Enterprise Edition, the distribution of masters should be even by default because prefer-uniform-balance is true by default. If using Community Edition, or prefer-uniform-balance is false then the master distribution is not expected to be evenly distributed. The distribution is determined by the node-ids in the cluster, which you can adjust to try for a better distribution - before prefer-uniform-balance, some users with particularly unfortunate distributions were able to improve their distribution this way.