We are upgrading to ES 6.x from 5.6.7. As the documentation specifies it can be rolling upgrade.
We are using the transport client and my question is regarding the same.
Can we continue writing/reading data when we have a mix of 5.6.7 and 6.x nodes in the cluster using a transport client compatible with 5.6.7.
I assume internal communication between ES nodes happens over the transport interface. How does this work in case of 5.6.7 and 6.x nodes in a cluster at the same time?
So we will need a transport client from 6.x, but ES 5.6.x and ES 6.x nodes can communicate with one another using the either one of the transport interface (5.6.7 or 6.x)?
So if we are using a transport client, the rolling upgrade is not really one for us, as we will have to upgrade all of the nodes before we can start reading/writing?. Please correct me if I am wrong.
Yes. That's what I think. I never tested it myself so I'd say that depending on the exact APIs you are using it can work but I'd not guarantee that at all...
Note that's the same similar concern with the HL Rest Client as it is designed today.
But you should definitely switch to this one anyway because the transport client will be removed in the future.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.