Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
9.6.1
-
None
-
None
Description
Initializing a Hunspell dictionary is a fairly costly process, as it involves reading and parsing fairly large files.
In my case, I have several collections that use the same set of 3 Hunspell dictionaries (they all use the same configset).
Dictionaries are initialized for each collection (and probably for each shard if there were any). This represents a significant overhead when starting up the Solr service.
Would it be possible to "pool" dictionaries across collections that use the same settings? In this way, the number of collections would have no effect on initialization time.
Translated with DeepL.com (free version)