[
https://issues.apache.org/jira/browse/HBASE-3234?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12931989#action_12931989
]
Gary Helmling commented on HBASE-3234:
--------------------------------------
@Stack: For "stack" vs. "unknown user", that's probably not related...
The first is from a Hadoop RPC call (ipc.Client), where the second is from an
HBase RPC call (ipc.HBaseClient). HBase RPC _always_ passes "null" for the
user in the RPC header -- it's hard coded in HBaseClient. We've changed this
in the secure HBase branch, of course, but it's always been this way in the
mainline. Master and RS should be using the process username for all DFS
client interactions.
> hdfs-724 "breaks" TestHBaseTestingUtility multiClusters
> -------------------------------------------------------
>
> Key: HBASE-3234
> URL: https://issues.apache.org/jira/browse/HBASE-3234
> Project: HBase
> Issue Type: Bug
> Reporter: stack
> Priority: Critical
> Fix For: 0.90.0
>
> Attachments:
> org.apache.hadoop.hbase.TestHBaseTestingUtility-output.txt,
> org.apache.hadoop.hbase.TestHBaseTestingUtility-output.txt,
> org.apache.hadoop.hbase.TestHBaseTestingUtility.txt
>
>
> We upgraded our hadoop jar in TRUNK to latest on 0.20-append branch.
> TestHBaseTestingUtility started failing reliably. If I back out hdfs-724,
> the test passes again. This issue is about figuring whats up here.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
