We are looking to implement Hot-Warm-Cold-Frozen-Delete phases for a client but the idea is to move the indexes to other storage: For example, keep hot on local SSDs but move frozen to offsite HDDs (just a example of course).
Reading the documentation though, Im kind of confused on how to set this up exactly. Can someone explain a bit better the phases? The documentation IMO isnt very clear on this, it just states what the parameters do .
No. The is no way using the native ILM capabilities to have data moved between different local storage locations on the same node for hot, warm, cold rollover. It's designed to move between nodes with a single storage type for each node.
Yes, though normally a node wouldn't be associated with more than one phase because the whole point is having the super high performance hardware on the hot nodes and the less performant hardware on the warm/cold phases. I don't think the delete phase has nodes associated with it, the indices just get deleted from wherever they are when they enter that phase.
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.