[
https://issues.apache.org/jira/browse/SOLR-12914?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vadim Miller updated SOLR-12914:
--------------------------------
Description:
TermsComponent class always tries to fetch all terms from all shards for a
further processing. There is {{java.lang.OutOfMemoryError}} __ exception if
the resulting list is too long.
There is a very common use case when the full terms list is not required: a
client needs to see next N terms in alphabetically sorted list starting with a
given value. Usually, this is needed for some autocomplete field on a page.
Example URL:
{{[http://localhost:8983/solr/mycollection/terms?terms.fl=fulltext&terms.sort=index&terms.lower=cat&terms.limit=50]}}
In this example TermsComponent needs to fetch only 50 terms from each shard
starting with a value provided in {{terms.lower}} URL parameter. So, it should
not reset TermsParams.TERMS_LIMIT parameter when generates a shard query in
createSmartShardQuery() method.
The patch is attached.
was:
TermsComponent class always tries to fetch all terms from all shards for a
further processing. There is {{java.lang.OutOfMemoryError}} __ exception if
the resulting list is too long.
There is a very common use case when the full terms list is not required: a
client needs to see next N terms in alphabetically sorted list starting with a
given value. Usually, this is needed for some autocomplete field on a page.
Example URL:
{{http://localhost:8983/solr/mycollection/terms?terms.fl=fulltext&terms.sort=index&terms.lower=cat&terms.limit=50}}
In this example TermsComponent needs to fetch only 50 terms from each shard
starting with a value provided in {{terms.lower}} URL parameter. So, it should
not reset TermsParams.TERMS_LIMIT parameter when generates a shard query in
createSmartShardQuery() method.
The patch is attached.
> Solr crashes in /terms request handler
> --------------------------------------
>
> Key: SOLR-12914
> URL: https://issues.apache.org/jira/browse/SOLR-12914
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Affects Versions: 7.5
> Reporter: Vadim Miller
> Priority: Major
> Labels: terms
> Attachments: terms.patch
>
>
> TermsComponent class always tries to fetch all terms from all shards for a
> further processing. There is {{java.lang.OutOfMemoryError}} __ exception if
> the resulting list is too long.
> There is a very common use case when the full terms list is not required: a
> client needs to see next N terms in alphabetically sorted list starting with
> a given value. Usually, this is needed for some autocomplete field on a page.
> Example URL:
>
> {{[http://localhost:8983/solr/mycollection/terms?terms.fl=fulltext&terms.sort=index&terms.lower=cat&terms.limit=50]}}
>
> In this example TermsComponent needs to fetch only 50 terms from each shard
> starting with a value provided in {{terms.lower}} URL parameter. So, it
> should not reset TermsParams.TERMS_LIMIT parameter when generates a shard
> query in createSmartShardQuery() method.
> The patch is attached.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
