[
https://issues.apache.org/jira/browse/LUCENE-7801?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15981356#comment-15981356
]
Robert Muir commented on LUCENE-7801:
-------------------------------------
Maybe add a simple toString impl too for the components involved here? In
general it makes it easier to understand where the RAM is going, at least if
you use Accountables.toString() to look at it.
Instead of adding .values() as the child elements, you may want to also use
Accountables namedAccountables method:
https://github.com/apache/lucene-solr/blob/master/lucene/core/src/java/org/apache/lucene/util/Accountables.java#L100
So something like namedAccountables("dim", cachedOrdMaps). This way you can see
which dim uses which amount of memory? Look at e.g. docvalues or other places
doing per-field stuff for some examples.
> SortedSetDocValuesReaderState should implement Accountable
> ----------------------------------------------------------
>
> Key: LUCENE-7801
> URL: https://issues.apache.org/jira/browse/LUCENE-7801
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Michael McCandless
> Assignee: Michael McCandless
> Fix For: master (7.0), 6.6
>
> Attachments: LUCENE-7801.patch
>
>
> This class is used by sorted set facets, and wraps per-dimension
> {{MultiDocValues.OrdinalMap}} which already implement accountable, and it's
> helpful to know how much heap this structure is taking for computing Lucene's
> SSDV facets.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
