[
https://issues.apache.org/jira/browse/SOLR-11200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16444514#comment-16444514
]
David Smiley commented on SOLR-11200:
-------------------------------------
It would be nice if somehow Solr could be smart enough to know when this
setting is appropriate vs not. Otherwise we have yet another magic setting
that expert users may or may not eventually find. For example if there are no
searches going on then don't throttle. Perhaps Solr could wrap the merge
scheduler so that when a merge is about to happen that it looks at the
SolrIndexSearcher to get some stats. Just a straw-man; I dunno. Perhaps
another similar direction is to enhance SolrIndexSearcher to close lazily if
it's not actually getting used (I've heard of this strategy used in a forked
Solr to reduce memory for massive # cores). And then what we detect in this
Merge Scheduler is quite simply if there is an active SolrIndexSearcher or not.
Today we always have one. Any way, something to consider.
> provide a config to enable disable ConcurrentMergeSchedule.doAutoIOThrottle
> ---------------------------------------------------------------------------
>
> Key: SOLR-11200
> URL: https://issues.apache.org/jira/browse/SOLR-11200
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Nawab Zada Asad iqbal
> Priority: Minor
> Attachments: SOLR-11200.patch, SOLR-11200.patch, SOLR-11200.patch
>
>
> This config can be useful while bulk indexing. Lucene introduced it
> https://issues.apache.org/jira/browse/LUCENE-6119 .
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
