Reducing the size of the SOLR index

We have several massive SOLR indexes for a large project and its consumption is over 50 GB of space.

We've looked at several ways to reduce the size associated with changes in content in the indices, but I'm curious if there might be any changes we can make to the SOLR index that will reduce its size by 2 orders of magnitude or more that are directly related to any ( 1) support commands that we can run, or (2) simple configuration options that might not be set correctly.

Another topical question: (3) Is there a way to trade the size index for performance inside SOLR, and if so, how does it work?

Any thoughts on this would be appreciated ... Thanks!

+3


source to share


1 answer


There are several things you could do to trade the performance for the size of the index. For example, an integer (int) field uses less space than an integer trie (tint), but range queries will be slower when using int.

To shorten your index significantly, you will almost certainly have to take a closer look at the fields you use.



  • Are you using a lot of stored fields? If so, try removing the stored fields from the index and querying your database for the data you need as soon as you get the results back from Solr.
  • Add omitNorms = "true" to text fields that don't need length normalization.
  • Add omitPositions = "true" to text fields that don't require a phrase match
  • Special fields like NGrams can take up a lot of space.
  • Are you removing stop words from text boxes?
+8


source







All Articles