JR58138: PARALLEL JOB FAILS WITH UNEXPECTED PLAYER TERMINATION WHEN THE ENVIRONMENT VARIABLE APT_PM_POLL_SYSCALL IS SET.

APAR status

• Closed as program error.

Error description

• The environment variable APT_PM_POLL_SYSCALL  changes the
  parallel engine
   process manager messaging system to use the poll() system
  call
   instead of the default select() call for polling on socket
  descriptors.
    The poll() call does not use a Unixfile descriptor,
  sosetting APT_PM_POLL_SYSCALL can prevent job failures when a
  busy system or large job runs out
   of file descriptors. These errors look like:
   IIS-DSEE-TUTL-00015 Fatal Error: File descriptor out of
  range in fd_set (requested 1,025, limit 1,023)
   Unfortunately, users discovered that the implementation of
  the environment variable test for
   APT_PM_POLL_SYSCALL had an error which could lead to
  intermittent job failures with messages like:
   IIS-DSEE-TFPM-00338  APT_PMsectionLeader(1, node1), player
  1 - Unexpected exit status 1.
   In once case, this problem occurred when users set the
  environment variable at the project level
   but there was no error when the environment variable was
  set at the job level .
  

Local fix

• Non
  

Problem summary

• PARALLEL JOB FAILS WITH UNEXPECTED PLAYER TERMINATION WHEN THE
  ENVIRONMENT VARIABLE APT_PM_POLL_SYSCALL IS SET
  The environment variable APT_PM_POLL_SYSCALL  changes the
  parallel engine
   process manager messaging system to use the poll() system
  call
   instead of the default select() call for polling on socket
  descriptors.
    The poll() call does not use a Unix
   file descriptor, so setting APT_PM_POLL_SYSCALL can prevent
  job failures when a busy system or large job runs out
   of file descriptors. These errors look like:
   IIS-DSEE-TUTL-00015 Fatal Error: File descriptor out of
  range in fd_set (requested 1,025, limit 1,023)
  
   Unfortunately, users discovered that the implementation of
  the environment variable test for
   APT_PM_POLL_SYSCALL had an error which could lead to
  intermittent job failures with messages like:
   IIS-DSEE-TFPM-00338  APT_PMsectionLeader(1, node1), player
  1 - Unexpected exit status 1.
   In once case, this problem occurred when users set the
  environment variable at the project level
   but there was no error when the environment variable was
  set at the job level .
  

Problem conclusion

• There was an error in the code which tested the environment
  variable. There are patches is which fix the problem.
  

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

• SERVER
  

Fix information

• Fixed component name

  WIS DATASTAGE

• Fixed component ID

  5724Q36DS

Applicable component levels

• RB70 PSY

     UP