IV68577: APPLICATION BINARY KILLED WHEN A SIGKILL COMMAND IS SENT TO THE PARENT (STARTCOMMAND)

4.1.1-TIV-SAMP-Linux64-FP0001
4.1.1-TIV-SAMP-AIX-FP0001
4.1.0-TIV-SAMP-AIX-FP0002
4.1.0-TIV-SAMP-Linux64-FP0002
4.1.0-TIV-SAMP-Linux-FP0002
3.2.2-TIV-SAMP-Linux-FP0009
3.2.2-TIV-SAMP-AIX-FP0009
3.2.2-TIV-SAMP-Solaris-FP0009
3.2.2-TIV-SAMP-Windows-FP0009
3.2.2-TIV-SAMP-Windows-FP0009
4.1.0-TIV-SAMP-AIX-FP0003
4.1.0-TIV-SAMP-Linux-FP0003
4.1.0-TIV-SAMP-Linux64-FP0003
4.1.0-TIV-SAMP-AIX-FP0006
4.1.0-TIV-SAMP-Linux-FP0006
4.1.0-TIV-SAMP-Linux64-FP0006
4.1.0-TIV-SAMP-AIX-FP0007
4.1.0-TIV-SAMP-Linux64-FP0007
SAMP4107_Verification_Signatures
4.1.1-TIV-SAMP-AIX-FP0000
4.1.1-TIV-SAMP-Linux64-FP0000
3.2.2-TIV-ITSAMP-FP0009, Tivoli System Automation for Multiplatforms
4.1.0-TIV-ITSAMP-FP0003, Tivoli System Automation for Multiplatforms
4.1.0-TIV-ITSAMP-FP0004, Tivoli System Automation for Multiplatforms

APAR status

• Closed as program error.

Error description

• The customer found that after the startcommand exits
  successfully, a SIGKILL is still able to terminate the
  application binary started by it.
  

Local fix

• Introduce a wrapper script that will change the PGID.
  =====/work/setpgrp.pl=====
  #!/usr/bin/perl
  setpgrp 0;
  system (@ARGV);
  =======================
  In the startup command:
  nohup /work/setpgrp.pl  binary.sh >>/tmp/log 2>&1 &
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: Tivoli System Automation for Multiplatforms
  * 4.1.0 users using IBM.Application resources
  ****************************************************************
  * PROBLEM DESCRIPTION:
  * Even a successful ended StartCommand for an IBM.Application
  * resource might receive a SIGKILL, which then is still able to
  * terminate the application started by the StartCommand. The
  * problem is a race condition and might be observed in case the
  * child processes (application processes) started by the
  * StartCommand did not detach fast enough from the StartCommand
  * thread. Using 'nohup' to start the application processes seems
  * to increase the probability to hit the problem.
  ****************************************************************
  * RECOMMENDATION:
  ****************************************************************
  

Problem conclusion

• The internal handling of the GblResRM has been adjusted to
  address the issue. Now there will be no SIGKILL sent to a
  StartCommand that did not run into a time out.
  .
  The official fix for this problem is included in Tivoli System
  Automation for Multiplatforms 4.1.0 fix pack 2
  | 4.1.0-TIV-ITSAMP-FP0002 |
  .
  Additional Search Keywords
  .
  

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

  SA MULTIPLATFOR

• Fixed component ID

  5724M0000

Applicable component levels