PI81590: AMBARI-17539 AMBARI AGENT MEMORY LEAK FIX

APAR status

• Closed as fixed if next.

Error description

• IBM BigInsights 4.2 is based off Ambari 2.2.0
  When using Ambari 2.2.0, the ambar-agent process memory
  utilisation may increase over time from around 500 MB to huge
  values 10 GB and higher having been reported.
  
  AMBARI-17539 addresses this in IBM BigInsights 4.3 which is
  based off 2.4.2
  

Local fix

• In a test environment where you see this behaviour:
  a. Make a backup of the file
  /usr/lib/python2.6/site-packages/ambari_agent/main.py
  b. Stop the ambari agent.
  c. Amend /usr/lib/python2.6/site-packages/ambari_agent/main.py
  exactly as follows:
  def fix_subprocess_racecondition():
    """
    !!! PLEASE NOTE THIS SHOULD BE CALLED BEFORE ANY OTHER
  INITIALIZATION was done to avoid already created links to
  subprocess or subprocess.gc or gc
    """
    # monkey patching subprocess
    import subprocess
    subprocess.gc.isenabled = lambda: True
  
    # re-importing gc to have correct isenabled for
  non-subprocess contexts
    import sys
    del sys.modules['gc']
    import gc
  
  d. Start the Ambari agent - confirm that this starts
  successfully!!
  e. Monitor memory usage over a period of time.
  f. If this does seem to address the memory growth, then apply
  this change to all nodes running ambari-agents.
  

Problem summary

• Fixed in BigInsights 4.3 / Ambari 2.4.2
  Specific code from:
  https://bugs.python.org/issue27448;#msg269804
  

Problem conclusion

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Modules/Macros

• ambari
  

Fix information

Applicable component levels

• R430 PSY

     UP