Curator & Knowing when documents have been unIndexed


#1

Is there any way of being notified about this or any way to find out?


(Aaron Mildenstein) #2

What do you mean by "unIndexed?" This term is not one I'm familiar with.


#3

Assume Curator is configured to delete, on a daily basis, documents that are over 30 days old. They get marked as such and once segments are merged, the documents are gone. No longer indexed. Unindexed (I will coin this term).


(Aaron Mildenstein) #4

Curator can only delete entire indices. If you need to delete documents from within an index, you should look into delete_by_query


#5

Yes I understand that, but surely the associated documents also get deleted right?


(Aaron Mildenstein) #6

Think of it like SQL. A document is a row, and an index a table. Drop the table, and all the rows go with it. They don't get "tagged" for deletion. There's not even a need for a segment merge. An index deletion in Elasticsearch is quick and permanent. The segment files are purged from the filesystem immediately, unless you have a very slow cluster/node, in which case it will do it as soon as it can.


#7

Ok I got you. Thanks


(system) #8

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.