You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@kylin.apache.org by "Shaofeng SHI (JIRA)" <ji...@apache.org> on 2019/01/24 00:52:00 UTC

[jira] [Commented] (KYLIN-3783) The mapreduce.map.java.opts config in kylin_job_conf_inmem.xml overrides the krb5.conf config in Cluster

    [ https://issues.apache.org/jira/browse/KYLIN-3783?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16750589#comment-16750589 ] 

Shaofeng SHI commented on KYLIN-3783:
-------------------------------------

The default configuration file may not consider all cases. You can modify the kylin_job_conf_inmem.xml manually to match your environment need, the file in conf/ folder is for that case. 

> The mapreduce.map.java.opts config in kylin_job_conf_inmem.xml overrides the krb5.conf config in Cluster
> --------------------------------------------------------------------------------------------------------
>
>                 Key: KYLIN-3783
>                 URL: https://issues.apache.org/jira/browse/KYLIN-3783
>             Project: Kylin
>          Issue Type: Bug
>    Affects Versions: v2.5.2
>         Environment: hadoop 2.7
>            Reporter: yuzhang qiu
>            Priority: Minor
>
> In our cluster, we use kerberos for authorization, and config `-Djava.security.krb5.conf` in `mapreduce.map.java.opts`. But the default configuration in kylin_job_conf_inmem.xml is `-Xmx2700m -XX:OnOutOfMemoryError='kill -9 %p'`, which will overrides the krb5 config when the cubing algorithm is **inmem**. Then, we got `Caused by: KrbException: Cannot locate default realm
> `



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)