[
https://issues.apache.org/jira/browse/HDFS-13795?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16574412#comment-16574412
]
Ewan Higgs commented on HDFS-13795:
-----------------------------------
{quote}Shouldn't the approach be to make sure aliasMap always exists by
checking the configuration and refusing to start unless the configuration is
valid?{quote}
I think there should not be a null check inside the AliasMap. By returning
{{Option.empty()}} we give the impression that there is no data in the AliasMap
when there may be - we're not misconfigured.
The issue here is that the NN should detect the misconfinguration and fail in a
nicer way than NPE here.
> Fix potential NPE in InMemoryLevelDBAliasMapServer
> --------------------------------------------------
>
> Key: HDFS-13795
> URL: https://issues.apache.org/jira/browse/HDFS-13795
> Project: Hadoop HDFS
> Issue Type: Bug
> Reporter: Virajith Jalaparti
> Assignee: Virajith Jalaparti
> Priority: Major
> Attachments: HDFS-13795.001.patch, HDFS-13795.002.patch,
> HDFS-13795.003.patch, HDFS-13795.004.patch
>
>
> Namenode fails to stop correctly due to NPE in InMemoryAliasMapServer, when
> it is configured incorrectly.
> {code:java}
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.server.aliasmap.InMemoryLevelDBAliasMapServer.close(InMemoryLevelDBAliasMapServer.java:149)
> at
> org.apache.hadoop.hdfs.server.namenode.NameNode.stop(NameNode.java:1023)
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
