Could anyone please tell me if they would expect the presence of a parent /
child query warmer to slow replication down greatly? (From ~5 minutes to
~30 minutes for 12gb of data.)
It is recommended to disable the warmer before bulk indexing, and I wonder
if this applies to replication as well?
I believe this is what we are seeing, and if so, it causes us a problem
around introducing nodes into a cluster in a live environment, as we must
choose between:
Warmers don't slow replication down, replication will take the same time. If you still have the refresh interval to 1 second, then when you index, the warmers will be executed every time the index gets refreshed, which will add more load on the nodes, which will slow down indexing because of the extra load. Thats the main reason why we allow to disable warmers, specifically when doing initial bulk indexing.
Could anyone please tell me if they would expect the presence of a parent / child query warmer to slow replication down greatly? (From ~5 minutes to ~30 minutes for 12gb of data.)
It is recommended to disable the warmer before bulk indexing, and I wonder if this applies to replication as well?
I believe this is what we are seeing, and if so, it causes us a problem around introducing nodes into a cluster in a live environment, as we must choose between:
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.