The default facet.limit is 10, but it's set to 50 for most of the
facets. I've included the query parameters below. In case it makes any
difference, there are quite a lot of facet fields with large numbers of
terms, and the queries are being generated by the Sarnia Drupal module.

Thanks,
Andy

---
<lst name="params">
    <str name="f.links_public_suffixes.facet.limit">50</str>
    <str name="f.postcode_district.facet.limit">50</str>
    <str name="facet">true</str>
    <str name="sort">wayback_date asc</str>
    <str name="f.content_type_served.facet.limit">50</str>
    <str name="f.sentiment.facet.limit">50</str>
    <str name="facet.limit">10</str>
    <str name="f.content_encoding.facet.limit">50</str>
    <str name="f.last_modified_year.facet.limit">50</str>
    <str name="f.links_hosts.facet.limit">50</str>
    <str name="facet.method">enum</str>
    <str name="f.author.facet.limit">50</str>
    <str name="fl">*,score</str>
    <str name="f.content_type_full.facet.limit">50</str>
    <str name="f.content_type.facet.limit">50</str>
    <str name="f.content_type_ext.facet.limit">50</str>
    <arr name="facet.field">
        <str>sentiment</str>
        <str>private_suffix</str>
        <str>public_suffix</str>
        <str>postcode_district</str>
        <str>content_type_ext</str>
        <str>content_type_full</str>
        <str>{!ex=content_type_norm}content_type_norm</str>
        <str>content_type_served</str>
        <str>content_type</str>
        <str>content_language</str>
        <str>author</str>
        <str>content_encoding</str>
        <str>content_ffb</str>
        <str>{!ex=crawl_year}crawl_year</str>
        <str>domain</str>
        <str>links_public_suffixes</str>
        <str>links_private_suffixes</str>
        <str>links_hosts</str>
        <str>generator</str>
        <str>last_modified_year</str>
    </arr>
    <str name="qt">standard</str>
    <str name="facet.enum.cache.minDf">500</str>
    <str name="facet.missing">false</str>
    <str name="f.crawl_year.facet.limit">50</str>
    <str name="facet.mincount">1</str>
    <str name="f.content_language.facet.limit">50</str>
    <str name="json.nl">map</str>
    <str name="wt">xml</str>
    <str name="f.private_suffix.facet.limit">50</str>
    <str name="rows">20</str>
    <str name="f.content_ffb.facet.limit">50</str>
    <str name="f.generator.facet.limit">50</str>
    <str name="f.links_private_suffixes.facet.limit">50</str>
    <str name="f.domain.facet.limit">50</str>
    <str name="facet.sort">count</str>
    <str name="start">0</str>
    <str name="q">*:*</str>
    <str name="f.public_suffix.facet.limit">50</str>
    <str name="f.content_type_norm.facet.limit">50</str>
</lst>
---

> -----Original Message-----
> From: Michael Ryan [mailto:mr...@moreover.com]
> Sent: 03 September 2013 13:41
> To: solr-user@lucene.apache.org
> Subject: RE: Memory usage during aggregation - SolrCloud with very
large
> numbers of facet terms.
> 
> > However, the Solr instance we direct our  client query to is
consuming
> significantly more RAM (10GB) and is still failing after a few queries
when it
> runs out of heap space. This is presumably due to the role it plays,
> aggregating the results from each shard.
> 
> That seems quite odd... What facet parameters are you using in the
query? I
> could imagine memory issues if you're using facet.limit=-1, or some
very
> large number.
> 
> -Michael

Reply via email to