[HDFS-3915] QJM: Failover fails with auth error in secure cluster

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: QuorumJournalManager (HDFS-3077)
• Fix Version/s: QuorumJournalManager (HDFS-3077)
• Component/s: ha, security
• Hadoop Flags: Reviewed
• Assignee: Todd Lipcon
• Reporter: Todd Lipcon
• Created: 10/Sep/12 22:28
• Updated: 11/Sep/12 04:53
• Resolved: 11/Sep/12 04:53

Description

When testing failover in a secure cluster with QJM, we ran into the following error:

java.io.IOException: Exception trying to open authenticated connection to http://xxxxx:8480/getJournal?jid=journal&segmentTxId=4325&storageInfo=-40%3A1049822920%3A0%3ACID-d7c84ac3-bb09-4d55-baae-0d561bb55e9b
	at org.apache.hadoop.security.SecurityUtil.openSecureHttpConnection(SecurityUtil.java:510)
	at org.apache.hadoop.hdfs.server.namenode.EditLogFileInputStream$URLLog$1.run(EditLogFileInputStream.java:376)
...	at org.apache.hadoop.hdfs.server.namenode.ha.EditLogTailer.doTailEdits(EditLogTailer.java:217)
	at org.apache.hadoop.hdfs.server.namenode.ha.EditLogTailer.catchupDuringFailover(EditLogTailer.java:176)
	at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.startActiveServices(FSNamesystem.java:635)
Caused by: GSSException: No valid credentials provided (Mechanism level: Failed to find any Kerberos tgt)

The issue is that the EditLogFileInputStream uses the "current" user, which in the case of the failover trigger is the admin's remote user, rather than the NN's login user.