IV87037: GC CRASH WHEN CONCURRENT GC WORK PACKET OVERFLOW

APAR status

• Closed as program error.

Error description

• Error Message: <OSB>ERR<CSB> Unhandled exception
  <OSB>ERR<CSB> Type=Segmentation error vmState=0x00000000
  <OSB>ERR<CSB> J9Generic_Signal_Number=00000004
  Signal_Number=0000000b Error_Value=24cf9248 Signal_Code=00000001
  ...
  .
  Stack Trace: #12 <signal handler called>
  #13 0x7bebda54 in
  MM_CollectorLanguageInterfaceImpl::private_markingScheme_getRefe
  renceStatus(MM_EnvironmentStandard*, J9Object*, bool*, bool*) ()
     from .../libj9gc29.so
  Cannot access memory at address 0x7b6be1a8
  .
  

Local fix

• This issue can be worked around using -Xgcworkpackets option to
  increase size of work packet so that the buffer overflow does
  not happen.
  The default size of the work packet is proportional to the heap
  size.  For a 6 GB heap, the default size is 15725. In one
  situation, the overflow was avoided using -Xgcworkpackets50000.
  Note this will consume approximately 150 MB extra native memory.
  

Problem summary

• The problem is caused when one of gc state has not been set for
  java thread during concurrent work packet overflow.
  

Problem conclusion

• The JVM has been updated to prevent the crash from checking if
  gc state has been set before using it.
  .
  This APAR will be fixed in the following Java Releases:
     7 R1 SR3 FP60  (7.1.3.60)
     8    SR3 FP11  (8.0.3.11)
  .
  Contact your IBM Product's Service Team for these Service
  Refreshes and Fix Packs.
  For those running stand-alone, information about the available
  Service Refreshes and Fix Packs can be found at:
             https://www.ibm.com/developerworks/java/jdk/
  

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

  J9 COMMON CODE

• Fixed component ID

  620700127

Applicable component levels

• R270 PSY

     UP