Hi all, I would like to request review for HDFS-17276:
JIRA: https://issues.apache.org/jira/browse/HDFS-17276 PR: https://github.com/apache/hadoop/pull/8763 In a secure HA+QJM deployment, a valid NameNode SPNEGO request to /getJournal can be rejected with HTTP 403 because authorization receives the short remote user instead of the full Kerberos principal. The patch prefers request.getUserPrincipal().getName() when available, while retaining getRemoteUser() as a fallback. The Yetus run reported +1 overall, including the hadoop-hdfs unit tests. Could an HDFS committer please review the authorization approach, or advise on the appropriate owner/direction? Thanks, kuper
