I have a three node cluster set up with no explicitly defined roles for each nodes. I can see by disk usage and index document count that no data was ingested into master node - only non-master nodes have data on them. How do I equally distribute the data across all three nodes.
10.0.2.98 73 99 6 0.70 0.45 0.22 cdfhilmrstw * esnode03
10.0.2.97 3 94 0 0.23 0.29 0.19 cdfhilmrstw - esnode02
10.0.2.96 17 50 6 0.48 0.65 0.47 cdfhilmrstw - esnode01
What is the full output of the cluster stats API?
Here's the full output. Originally and at the time of ingestion esnode01 was the master, but after I bounced it esnode03 was elected master. The original jvm size for esnode01 was 32GB just as 02 and 03. The data was sent to the API of esnode01.
That does not look like the output from the API I linked to. Are you sure you ran the correct API?
Note that when you are indexing data into Elasticsearch data can be sent to any node and the master role does not participate in request processing unles something results in a change to the cluster state, e.g. new index is created or mappings changed. Which node is master at any time does therefore have no imapct at all on which node you index data into nor where the data ends up.
What is the configuration of the nodes? Are they all identical?
They are all identicial, but I suspect I've made a rookie mistake and didn't add enough shards. I'll try to reindex and see how it goes
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.