You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Aleksandr Shulman (JIRA)" <ji...@apache.org> on 2013/12/06 04:12:36 UTC

[jira] [Commented] (HBASE-10073) [Hadoop1]: hbase zkcli broken due to slf4j incompatibility

    [ https://issues.apache.org/jira/browse/HBASE-10073?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13840904#comment-13840904 ] 

Aleksandr Shulman commented on HBASE-10073:
-------------------------------------------

Assigned to Andrew Purtell to have a look, since he is the author of the patch. The patch itself is not incorrect, but reveals a larger issue of maintaining compatibility and harmony among the dependencies.

> [Hadoop1]: hbase zkcli broken due to slf4j incompatibility
> ----------------------------------------------------------
>
>                 Key: HBASE-10073
>                 URL: https://issues.apache.org/jira/browse/HBASE-10073
>             Project: HBase
>          Issue Type: Bug
>          Components: Zookeeper
>    Affects Versions: 0.96.1
>         Environment: Centos6, sun-jdk-64bit-1.7.0.25
>            Reporter: Aleksandr Shulman
>            Assignee: Andrew Purtell
>
> Observed behavior:
> In my automation, I have a call to hbase zkcli. That call recently broke with this checkin: https://github.com/apache/hbase/commit/5af0a60efed91ac2084f25f13edb21db0f510e7c
> The error that is reported is:
> {code}++ ./hbase zkcli
> 11:19:58  Warning: $HADOOP_HOME is deprecated.
> 11:19:58  
> 11:20:00  Exception in thread "main" java.lang.IllegalAccessError: tried to access field org.slf4j.impl.StaticLoggerBinder.SINGLETON from class org.slf4j.LoggerFactory
> 11:20:00  	at org.slf4j.LoggerFactory.<clinit>(LoggerFactory.java:60)
> 11:20:00  	at org.apache.zookeeper.ZooKeeperMain.<clinit>(ZooKeeperMain.java:50)
> 11:20:00  	at org.apache.hadoop.hbase.zookeeper.ZooKeeperMainServer.main(ZooKeeperMainServer.java:78)
> 11:20:00  Build step 'Execute shell' marked build as failure{code}
> That said, this checkin is perfectly valid as each component should be allowed to specify its own dependencies.
> The issue is a deeper one of dependency mismatches.
> Note: This issue only affects hadoop1, not hadoop2. It also appears in trunk, where there is a similar checkin, but since trunk is not required to work against hadoop1, this is not an issue for trunk.



--
This message was sent by Atlassian JIRA
(v6.1#6144)