P100050: ELIMS KEEP BEING KILLED IN WINDOWS SERVER

APAR status

• Closed as program error.

Error description

• Run several custom elims on a windows server, some elims will be
  killed constantly for hanging too long without successfully
  reporting the resource values.
  
  Turn on the debug for melim log, will see:
  
  
  Oct 24 14:29:23 2012 42676:2888 7 1.2.6 terminateSubElim: SELIM
  <name=C:\LSF_8.3\8.3\etc\elim.systeminfo.bat, pid=42620,
  group=0> is terminated.
  Oct 24 14:29:23 2012 42676:2888 4 1.2.6 checkSubElim: the SELIM
  <C:\LSF_8.3\8.3\etc\elim.systeminfo.bat> is killed for hanging
  longer than LSF_ELIM_BLOCKTIME<35>
  

Local fix

• N/A
  

Problem summary

• Root cause:
  For Windows, all the elim's output are read in together. Need
  to clear the FD_SET after finishing reading the load string of
  each elim. Unlike Linux, which reads the complete load string
  in one function call, Windows use a big loop to read the load
  string and call the same function many times.  So FD_CLR is
  correct for Linux but not Windows.
  

Problem conclusion

• Patch provided.
  
  The elim output string must end with one and only one '\n'
  without extra spaces!
  

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:

Fix information

• Fixed component name

  LSF STAND EDITI

• Fixed component ID

  5725G8201

Applicable component levels

• R830 PSY

     UP