I already know and used Curator to manage indices and was now reading about the ILM. It seems to me that Curator does a lot more that ILM, currently.
However, since ILM is quite new and Curator is some sort of an external tool, I am wondering if ILM will, eventually, replace Curator or at least integrate many of its features.
Is it possible to know the roadmap of ILM? What are the plans regarding this ILM vs Curator?
The answer, like most answers we give, is It depends
ILM replaces most of the basic Curator functionality. If all you are doing is rolling over indices, hot/warm allocation, forceMerge, shrink, and delete, then you will want to use ILM.
If you have a more involved index management pipeline, you can continue to use Curator. It isn’t going away.
Snapshots are not part of ILM, but will be addressed in the future by SLM: Snapshot Lifecycle Management (I’m not providing a target release number or date for that—I don’t have insight into that). Until then, you may want to continue to use Curator for snapshots.
One of the main features I used with Curator was to delete old indices, but always based on their name (when they were time based, for example, indexname-2010.05.05), rather than their creation time.
If I understood correctly, this is currently not supported by ILM, but I am wondering if it will ever be possible to do this with ILM in the future. Is this in the roadmap of ILM?
Who told you you couldn’t use non-rollover indices with ILM? You can! You can configure the different phases, and just set those to occur x days or hours after index creation, rather than after rollover.
That’s part of the trade-off in ILM, though. With Curator, you can use different criteria to determine the age of an index (or the data inside). With ILM, it’s either from the time the index was created, or if using rollover indices, from the time it rolled over. Dig into the UI in Kibana and see what you can do within the “Index Lifecycle Policy” management interface.
To get ILM to operate on new and/or existing indices you do still need to associate the indices with an ILM policy, whether by index template (preferred for new indices), or by updating the settings on older ones.
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.