[
https://issues.apache.org/jira/browse/IGNITE-5686?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16080416#comment-16080416
]
Igor Seliverstov commented on IGNITE-5686:
------------------------------------------
{{IgniteQueryIterator}} implements {{scala.Iterator}} and is used inside Spark.
Spark doesn't expect {{AutoClosable}} iterator and isn't responsible for
closing it. Compute operation happens only on terminal action ant it's
guaranteed that {{TaskContext}} will be available, so, we can register a
callback to clean resources after operation finishes.
> Endless partition eviction during node shutdown
> -----------------------------------------------
>
> Key: IGNITE-5686
> URL: https://issues.apache.org/jira/browse/IGNITE-5686
> Project: Ignite
> Issue Type: Bug
> Components: cache
> Affects Versions: 2.1
> Reporter: Alexey Goncharuk
> Assignee: Igor Seliverstov
> Fix For: 2.2
>
> Attachments: dump.txt
>
>
> Observed the following behavior on TC:
> A node cannot stop because system threads are continuously trying to clean
> partitions.
> http://ci.ignite.apache.org/viewLog.html?buildId=702633&buildTypeId=Ignite20Tests_IgniteRdd&tab=buildResultsDiv
> Attached is a thread dump from the test
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
