JR60633: Parallel engine jobs fail intermittently with "Temporary failure in name resolution" during job startup.

IBM InfoSphere Information Server, Version 11.7.1.1 (Fix Pack 1)
IBM InfoSphere Information Server, Version 11.7.1.1 for AIX
IBM InfoSphere Information Server, Version 11.7.1.1 for Windows
IBM InfoSphere Information Server, Version 11.7.1.1 for Linux

APAR status

• Closed as program error.

Error description

• PX uses the getaddr() system call to set up connections among
  processes in a job. Even in a stable network, there is a chance
  that the getaddrinfo() system call may return the error
  "Temporary failure in name resolution"
  due to network packet  dropping. The fix for this issuem makes
  PX more forgiving in this case by having PX retry the system
  call after  a defined interval.
  

Local fix

Problem summary

• Patches are available which fix the issue.
  

Problem conclusion

• The fix for this issue  creates two new PX environment
  variables
  APT_ADDRINFO_RETRY and APT_ADDRINFO_RETRY_INTERVAL
   that allow users to define the number of retries, and the
  interval
  between each retry of getaddrinfo(). The interval is in
  seconds, e.g. 3 means
  sleep 3 seconds then retry the call.
  APT_ADDRINFO_RETRY
      If set and the value is a positive number, the player will
  try the number
      of getaddrinfo() calls when the call's return code is
  EAI_AGAIN, e.g.
      "Temporary failure in name resolution" error from the
  system. If the error
      persists after the defined number of retries, the job will
  abort and will log
      a new message with the information of how many times of
  retry, the interval,
      and the system error.
  APT_ADDRINFO_RETRY_INTERVAL
      Set it to a positive number if sleep(n) is desired between
  each retry.
      (n means n seconds.) This environment variable only works
  when
      APT_ADDRINFO_RETRY is defined.
  

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

• None
  PX
  

Fix information

• Fixed component name

  WIS DATASTAGE

• Fixed component ID

  5724Q36DS

Applicable component levels

• R912 PSY

     UP