Deleting docs is expensive but deleting whole indices is quick/cheap.
This is why we advocate using "time-based indices" e.g. keeping one per month and then moving old indices to backup storage or dropping whole indices after a time. This practice has been formalised in the idea of datastreams
We have Index where we are logging documents with request and response payload for few number of transaction ,not for all the transaction.
So what we are thinking is we can delete the just request -response payload of that particular transaction after 15 days may be, so that storage issue will not be there.
So we have problem for using DBQ for this as it take so much time.
Please suggest.
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.