You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Chesnay Schepler (JIRA)" <ji...@apache.org> on 2017/12/12 19:29:00 UTC
[jira] [Closed] (FLINK-8213) Prevent potential metric-related
ClassCastExceptions
[ https://issues.apache.org/jira/browse/FLINK-8213?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Chesnay Schepler closed FLINK-8213.
-----------------------------------
Resolution: Fixed
Fix Version/s: 1.5.0
master: 0bf0fdc26ea86020929fa64d083dce02ba2a2ae2
> Prevent potential metric-related ClassCastExceptions
> ----------------------------------------------------
>
> Key: FLINK-8213
> URL: https://issues.apache.org/jira/browse/FLINK-8213
> Project: Flink
> Issue Type: Bug
> Components: Metrics, Streaming
> Affects Versions: 1.5.0
> Reporter: Chesnay Schepler
> Assignee: Chesnay Schepler
> Priority: Minor
> Fix For: 1.5.0
>
>
> There are a few places, in particular revolving around the {{StreamOperator}}, that may run into ClassCastExceptions.
> For example, in the {{DataSinkTask}} you can find this line:
> {{final Counter numRecordsIn = ((OperatorMetricGroup) ctx.getMetricGroup()).getIOMetricGroup().getNumRecordsInCounter();}}
> The cast here is necessary since the {{AbstractStreamOperator}} exposes the {{OperatorMetricGroup}} as a plain {{MetricGroup}} since it is an internal class. However there is obviously no guarantee that {{StreamOperator#getMetricGroup()}} returns an actual {{OperatorMetricGroup}}. In fact, a fall-back case in {{AbstractStreamOperator#setup}} causes exactly this assumption to be violated.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)