[
https://issues.apache.org/jira/browse/LUCENE-6157?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14263190#comment-14263190
]
Adrien Grand commented on LUCENE-6157:
--------------------------------------
That would be nice as core keys don't cary any meaning but it felt a bit
awkward to implement. For instance, when we evict a filter in order to clear up
some space, we don't have a reference to a LeafReader available. So we would
need the per-leaf cache to keep a reference to the first used LeafReader and
use this instance when performing evictions. But then it means that we might
give a reference to a closed LeafReader (if the original leaf reader was closed
but there is still another LeafReader sharing the same core because of eg.
deletions) in these callbacks? Would it be ok?
> Add the ability to compute fine-grained statistics on the filter cache
> ----------------------------------------------------------------------
>
> Key: LUCENE-6157
> URL: https://issues.apache.org/jira/browse/LUCENE-6157
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Adrien Grand
> Assignee: Adrien Grand
> Priority: Minor
> Attachments: LUCENE-6157.patch
>
>
> The filter cache exposes some useful statistics about its usage, eg. hit
> count, eviction count, etc. In some cases it could be useful to give users
> the ability to compute finer-grained statistics though, for example by
> breaking up statistics by segment, index or by type of filter.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
