[AMBARI-16765] While restarting node manager through Ambari, the yarn.local.dir permission got change from 755 to 775, and the other way around, causing long restarting time for nodemanager

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.2.1.1
• Fix Version/s: 2.4.0
• Component/s: ambari-server
• Assignee: Papirkovskyy Myroslav
• Reporter: Papirkovskyy Myroslav
• Created: 19/May/16 11:14
• Updated: 19/May/16 11:43
• Resolved: 19/May/16 11:38

Description

Seems that Ambari changes the directory to 775, then change it back to 755:

2016-03-14 13:03:33,638 - Directory['/opt/hd/db/data/01/yarn/local'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,638 - Changing permission for /opt/hd/db/data/01/yarn/local from 755 to 775 
2016-03-14 13:03:33,655 - Directory['/opt/hd/db/data/02/yarn/local'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,656 - Changing permission for /opt/hd/db/data/02/yarn/local from 755 to 775 
2016-03-14 13:03:33,674 - Directory['/opt/hd/db/data/03/yarn/local'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,674 - Changing permission for /opt/hd/db/data/03/yarn/local from 755 to 775 
2016-03-14 13:03:33,689 - Directory['/opt/hd/db/data/04/yarn/local'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,689 - Changing permission for /opt/hd/db/data/04/yarn/local from 755 to 775 
2016-03-14 13:03:33,699 - Directory['/opt/hd/db/data/05/yarn/local'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,699 - Changing permission for /opt/hd/db/data/05/yarn/local from 755 to 775 
2016-03-14 13:03:33,709 - Directory['/opt/hd/db/data/01/yarn/log'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,709 - Directory['/opt/hd/db/data/02/yarn/log'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,710 - Directory['/opt/hd/db/data/03/yarn/log'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,710 - Directory['/opt/hd/db/data/04/yarn/log'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,711 - Directory['/opt/hd/db/data/05/yarn/log'] {'group': 'hadoop', 'recursive': True, 'cd_access': 'a', 'ignore_failures': True, 'mode': 0775, 'owner': 'yarn'} 
2016-03-14 13:03:33,712 - Execute[('chmod', '-R', '755', '/opt/hd/db/data/01/yarn/local', '/opt/hd/db/data/02/yarn/local', '/opt/hd/db/data/03/yarn/local', '/opt/hd/db/data/04/yarn/local'

This seems to cause long restart time for nodemanager.

Issue Links

• links to: ReviewBoard