[ 
https://issues.apache.org/jira/browse/SOLR-16628?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17678576#comment-17678576
 ] 

Jan Høydahl commented on SOLR-16628:
------------------------------------

Wow. This is an argument for not only using beefy machines for testing. We've 
seen similar Jenkins test failures due to resource constraints, timeouts etc, 
but yea, it's hard to reproduce in a consistent manner. Thanks for fixing.

> Occasional resource leak around XmlConfigFile parsing 
> ------------------------------------------------------
>
>                 Key: SOLR-16628
>                 URL: https://issues.apache.org/jira/browse/SOLR-16628
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>    Affects Versions: 8.9
>            Reporter: Michael Gibney
>            Priority: Minor
>          Time Spent: 1h 10m
>  Remaining Estimate: 0h
>
> Currently, xml config file parsing can in exceptional circumstances lead to 
> resource leaks (InputStream not being closed). This has occasionally [shown 
> up in 
> tests|https://issues.apache.org/jira/browse/SOLR-16336?focusedCommentId=17678323#comment-17678323]
>  and presumably could also manifest in the wild.
> I suspect (but am not certain) that these changes may have been introduced in 
> version 8.9 by SOLR-15337 (closing of byte streams in xml parsing is unusual 
> iiuc in that according to spec, closing is handled internal to the parser, so 
> it's easy to rely on that behavior and things will usually be ok. But we 
> should take extra steps to ensure the InputStreams are always closed).



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to