We have recently setup a federation cluster of tribe nodes with ES 1.7.4. The service that ingests data into the federated cluster(with indices already created) is experiencing master not discovered exception. The ingesting service uses bulk processor to do the same. Has anybody used bulk processor with tribe nodes to ingest data into elastic search before?
Are you using the tribe nodes to index?
Yes we want to use tribe node to index documents into elastic search. The clients create monthly rolling index currently when document comes for the next month's index. But considering the scale , we had to split our Elastic Search cluster into two,and now the client has to talk to two clusters so we are exploring tribe nodes instead of client talking to two clusters on its own.
We use bulk processor with transport client internally for indexing documents.
I have setup a test federation cluster which interacts with two federated clusters and I am trying to insert a document in one of the federated clusters. The index exists in one of the federated clusters, but still the document is not getting indexed and the exception being obtained in bulk response listener is MasterNotFoundException.
I have seen bulk processor code and didn't find any other master level operation being done which tribe node is trying to carry out, or I am missing something here?
Any issues that have been faced while using tribe nodes for write purposes would greatly help.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.