[
https://issues.apache.org/jira/browse/SOLR-9177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16049695#comment-16049695
]
Simon Tower commented on SOLR-9177:
-----------------------------------
I accidentally hosed our entire cluster when I queried for too many rows.
Having this option on by default when running in the foreground would have
saved me a lot of trouble today. One of the shard indexes became corrupt,
forcing us to delete all of the collections, recreate them, and reindex.
> Support oom hook when running Solr in foreground mode
> -----------------------------------------------------
>
> Key: SOLR-9177
> URL: https://issues.apache.org/jira/browse/SOLR-9177
> Project: Solr
> Issue Type: New Feature
> Reporter: Anshum Gupta
>
> After reading through the comments on SOLR-8145 and from my own experience,
> seems like a reasonable number of people run Solr in foreground mode in
> production.
> To give some more context, I've seen Solr hit OOM, which leads to IW being
> closed by Lucene. The Solr process hangs in there and without the oom killer,
> while all queries continue to work, all update requests start failing.
> I think it makes sense to add support to the bin/solr script to add the oom
> hook when running in fg mode.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
