The site I am supporting (a local history club, non-profit) is currently using "Search Lucene". But since the index is growing (german and english, 16 MB disk space for directory luceneapi_node/ ) the system is going slower. So I try to move the site to SolR. To check the feasibility for migration, the set up for the local test environment holds the following settings

  1. TAB "multilingual":
    German and English
    No Mapping
    Limit search to user's language by default
  2. TAB "Searchfields":
    All "language neutral" fields are set to "omit", as described in the README (see attached image 1)
  3. TAB Multilingual:
    Configuration files were downloaded and installed

The effects are

  1. TAB "search index" (admin/settings/apachesolr/index):
    There are fields like "body" and "spell" (see image-2) which look to mirror e.g. the "body_de" field (see "copy" commands in schema.xml)
  2. therefor the resulting index is doubling its size.
  3. after indexing 3000 nodes (about 15% of the site) SolR's data folder has a size of about 23 MB. This estimates to about 160 MB for the complete index.

These 160 MB are about 10 time more that the former one built by "Search Lucene".
So my questions are:

  • how can I get rid of the duplicated index?
  • which parts (fields) are candidates for compression?
CommentFileSizeAuthor
solr-2.png37.87 KBschildi
solr-1.png55.31 KBschildi

Comments

mkalkbrenner’s picture

Status: Active » Closed (won't fix)

6.x-1.x is not supported anymore.