Just saw that V6 is removing the _type. This is a major hiccup. I have a client or two, where I am using ES as a tad more than just logging. But even those are an issue.
By using a _type, I can accumulate similar type data in one place. Most of this is small.. one in particular has a couple of Bn docs in it.
So, the best way now will be to move these into separate indices.....or using custom' types, which involves changing client facing queries to cope.
By expanding out the indices, I run the risk of having way too many indices for ES to be able to be good with, so may have to consider different clusters also.
To be frank. its a PIA. Thats two features Elastico have removed, the other being _ttl, making t harder to use for its own good.
Having read the docs as to 'why' I still dont see why v6 cant have the switch to allow multiple types in one index, if required. Its a real Pain point for a long term user.. albeit inventively used
The document you're referring to (i.e. this one) goes into the nitty gritty details of why this change was needed. Agreed, this change was due to a technical limitation, but it will bring more good than bad in the future, or so I hope.
I'm in the same situation myself on several customer projects, but if you're using aliases wisely, the required changes will be pretty transparent for your client apps.
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.