[
https://issues.apache.org/jira/browse/SOLR-17018?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
rajanimaski updated SOLR-17018:
-------------------------------
Description: LTR queries are producing oom and looks like the issue is same
as what is described on this SOLR-5986. I think this happens because LTR
component does not implement `timeAllowed`. During peak traffic hours, when 80%
of resources are in use, one expensive LTR query that either has more terms or
paginated with start>=500 are not ceased and hence leading to full heap usage,
invoking full GC, and reporting an oom. I do see queries with ltr param are
taking more than 15 seconds which seems like is the reason for oom issue, it is
during this time interval the tenured(old generation) heap also gets filled up.
(was: LTR queries are producing oom and looks like the issue is same as
what is described on this SOLR-5986. Because LTR component does not implement
`timeAllowed`, during peak traffic hours, when 80% of resources are in use, one
expensive LTR query that either has 10 or more terms or paginated with
start>=500 are leading to full heap usage, invoking full GC, reporting an oom.
Certain LTR queries are taking more than 15 seconds which seems like is the
issue, it is during this time interval the tenured(old generation) heap is
getting filled up as well. )
> LTR queries producing out of memory issues
> ------------------------------------------
>
> Key: SOLR-17018
> URL: https://issues.apache.org/jira/browse/SOLR-17018
> Project: Solr
> Issue Type: Improvement
> Components: contrib - LTR, ltr
> Affects Versions: 9.1.1
> Reporter: rajanimaski
> Priority: Major
> Labels: ltr
>
> LTR queries are producing oom and looks like the issue is same as what is
> described on this SOLR-5986. I think this happens because LTR component does
> not implement `timeAllowed`. During peak traffic hours, when 80% of resources
> are in use, one expensive LTR query that either has more terms or paginated
> with start>=500 are not ceased and hence leading to full heap usage, invoking
> full GC, and reporting an oom. I do see queries with ltr param are taking
> more than 15 seconds which seems like is the reason for oom issue, it is
> during this time interval the tenured(old generation) heap also gets filled
> up.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
