[AMBARI-12636] Kerberos: fails check during enable Kerb with SLES

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.0.0
• Fix Version/s: 2.3.0, 2.1.1
• Component/s: ambari-server
• Labels: directory-permissions install

Environment

Ambari 2.1.1 Build #107
HDP 2.3 GA
ZK + AMS + Kafka
SLES 11 SP3
MIT KDC, all single node
Register hosts / bootstrap agents via SSH

• Assignee: Robert Levas
• Reporter: Robert Levas
• Created: 04/Aug/15 20:06
• Updated: 10/Aug/15 23:15
• Resolved: 10/Aug/15 19:38

Description

When executing the Kerberos service check, the following error occurs:

stderr:   /var/lib/ambari-agent/data/errors-24.txt

Traceback (most recent call last):
  File "/var/lib/ambari-agent/cache/common-services/KERBEROS/1.10.3-10/package/scripts/service_check.py", line 81, in <module>
    KerberosServiceCheck().execute()
  File "/usr/lib/python2.6/site-packages/resource_management/libraries/script/script.py", line 218, in execute
    method(env)
  File "/var/lib/ambari-agent/cache/common-services/KERBEROS/1.10.3-10/package/scripts/service_check.py", line 64, in service_check
    user=params.smoke_user
  File "/usr/lib/python2.6/site-packages/resource_management/core/base.py", line 157, in __init__
    self.env.run()
  File "/usr/lib/python2.6/site-packages/resource_management/core/environment.py", line 152, in run
    self.run_action(resource, action)
  File "/usr/lib/python2.6/site-packages/resource_management/core/environment.py", line 118, in run_action
    provider_action()
  File "/usr/lib/python2.6/site-packages/resource_management/core/providers/system.py", line 258, in action_run
    tries=self.resource.tries, try_sleep=self.resource.try_sleep)
  File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 70, in inner
    result = function(command, **kwargs)
  File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 92, in checked_call
    tries=tries, try_sleep=try_sleep)
  File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 140, in _call_wrapper
    result = _call(command, **kwargs_copy)
  File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 291, in _call
    raise Fail(err_msg)
resource_management.core.exceptions.Fail: Execution of '/usr/bin/kinit -c /var/lib/ambari-agent/data/tmp/kerberos_service_check_cc_dd529fe1e15538ddfe9ce0347604d64c -kt /etc/security/keytabs/kerberos.service_check.080315.keytab [e-mail]' returned 1. kinit(v5): Credentials cache permissions incorrect when initializing cache /var/lib/ambari-agent/data/tmp/kerberos_service_check_cc_dd529fe1e15538ddfe9ce0347604d64c

stdout:   /var/lib/ambari-agent/data/output-24.txt

Performing kinit using [e-mail]
2015-08-03 19:11:57,085 - Execute['/usr/bin/kinit -c /var/lib/ambari-agent/data/tmp/kerberos_service_check_cc_dd529fe1e15538ddfe9ce0347604d64c -kt /etc/security/keytabs/kerberos.service_check.080315.keytab [e-mail]'] {'user': 'jambari-qa'}
2015-08-03 19:11:57,179 - File['/var/lib/ambari-agent/data/tmp/kerberos_service_check_cc_dd529fe1e15538ddfe9ce0347604d64c'] {'action': ['delete']}

This error happens only on SLES, however the cause exists on all platforms. The other platforms silently ignore the condition; which, however, does not have any bearing on the results of the kinit test.

Cause
The "Credentials cache permissions incorrect when initializing cache" issue is caused by the inability to write the Kerberos ticket cache file to the specified location. In the case it is /var/lib/ambari-agent/data/tmp/kerberos_service_check_cc_dd529fe1e15538ddfe9ce0347604d64c. The reason for the write failure is that /var/lib/ambari-agent/data/tmp is not writable by the user executing the kinit call - which is the Ambari smoke test user (typically ambari-qa). The directory's permissions are

drwxr-xr-x. 4 root root 4096 Aug  3 22:20 /var/lib/ambari-agent/data/tmp/

Solution
In order for the ambari smoke test user to be able to write to the relevant directory (/var/lib/ambari-agent/data/tmp), the permissions must be set at least as follows

drwxrwxr-x. 4 root hadoop 4096 Aug  3 22:20 /var/lib/ambari-agent/data/tmp/

However, at the time this directory is created, it is not known what the name of the hadoop group is, so the next best solution is to set the permissions as

drwxrwxrwx. 4 root root 4096 Aug  3 22:20 /var/lib/ambari-agent/data/tmp/

If the ambari-agent is installed manually via the relevant package manager, the directory is created with the open permissions (777, drwxrwxrwx) via the packages install_helper.sh post install script. However if Ambari installs the agent via SSH, the directory is created with the more restrictive permissions (755, drwxr-xr-x) via the agent bootstrap.py script.

To make these consistent, the following needs to be changed

bootstrap.py:650

   command = "sudo mkdir -p {0} ; sudo chown -R {1} {0} ; sudo chmod 755 {3} ; sudo chmod 755 {2} ; sudo chmod 755 {0}".format(
      self.TEMP_FOLDER, quote_bash_args(params.user), DEFAULT_AGENT_DATA_FOLDER, DEFAULT_AGENT_LIB_FOLDER)

to

bootstrap.py (change)

   command = "sudo mkdir -p {0} ; sudo chown -R {1} {0} ; sudo chmod 755 {3} ; sudo chmod 755 {2} ; sudo chmod 777 {0}".format(
      self.TEMP_FOLDER, quote_bash_args(params.user), DEFAULT_AGENT_DATA_FOLDER, DEFAULT_AGENT_LIB_FOLDER)

Note: self.TEMP_FOLDER contains the path to the Ambari agent temp folder (typically, /var/lib/ambari-agent/data/tmp).

Issue Links

• links to: ReviewBoard