IV95032: LOGASSERTFAILED: (!"KERNEL REQUESTED TO DIE")

APAR status

• Closed as program error.

Error description

• GPFS assert and reboot:
  Mon Feb 27 02:06:11.902 2017: [X] logAssertFailed:
  (!"kernel requested
  to die")
  Mon Feb 27 02:06:11.980 2017: [X] return code 0, reason
  code 0, log
  record tag 0
  The assert subroutine failed: (!"kernel requested to
  die")
  threadId 107806777 Failure at line 2404 in file
  fetch-vfs-kx.C rc 0
  reason 0 data (nPrefetchedBuffers > 0)
  
  , file ../../../../../../../src/avs/fs/mmfs/ts/fs/svfs.C,
  line 382
  Mon Feb 27 02:06:12.062 2017: [E] Signal 6 at location
  0x90000000079EF14
  in process 7471106, link reg 0xFFFFFFFFFFFFFFFF.
  Mon Feb 27 02:06:12.072 2017: [I]   r0:
  0xFFFFFFFFFFFFFFFF    r1:
  0x0000000111911E20
  Mon Feb 27 02:06:12.080 2017: [I]   r2:
  0xFFFFFFFFFFFFFFFF    r3:
  0x0000000000000000
  Mon Feb 27 02:06:12.088 2017: [I]   r4:
  0xFFFFFFFFFFFFFFFF    r5:
  0xFFFFFFFFFFFFFFFF
  Mon Feb 27 02:06:12.096 2017: [I]   r6:
  0xFFFFFFFFFFFFFFFF    r7:
  0xFFFFFFFFFFFFFFFF
  Mon Feb 27 02:06:12.104 2017: [I]   r8:
  0xFFFFFFFFFFFFFFFF    r9:
  0xFFFFFFFFFFFFFFFF
  Mon Feb 27 02:06:12.112 2017: [I]  r10:
  0xFFFFFFFFFFFFFFFF   r11:
  0xFFFFFFFFFFFFFFFF
  Mon Feb 27 02:06:12.122 2017: [I]  r12:
  0xFFFFFFFFFFFFFFFF   r13:
  0x000000011191B800
  Mon Feb 27 02:06:12.132 2017: [I]  r14:
  0x000000000AFFFFFF   r15:
  0x00000001101D9760
  Mon Feb 27 02:06:12.140 2017: [I]  r16:
  0x0000000000000930   r17:
  0xF100000284989DE0
  Mon Feb 27 02:06:12.150 2017: [I]  r18:
  0x0000000000000001   r19:
  0x00000001101D8C04
  Mon Feb 27 02:06:12.160 2017: [I]  r20:
  0x00000000000000B2   r21:
  0x0000000111913A78
  Mon Feb 27 02:06:12.168 2017: [I]  r22:
  0x0000000000000007   r23:
  0x000000011012BCE8
  Mon Feb 27 02:06:12.176 2017: [I]  r24:
  0x0000000000000000   r25:
  0x000000011191214A
  Mon Feb 27 02:06:12.194 2017: [I]  r26:
  0x0000000111912140   r27:
  0x0000000000000002
  Mon Feb 27 02:06:12.202 2017: [I]  r28:
  0x0000000000000006   r29:
  0x0000000000000006
  Mon Feb 27 02:06:12.210 2017: [I]  r30:
  0x0000000000000000   r31:
  0x0000000000000405
  Mon Feb 27 02:06:12.222 2017: [I]  iar:
  0x090000000079EF14   msr:
  0xA00000000000D032
  Mon Feb 27 02:06:12.230 2017: [I]   cr:
  0x0000000000004000  link:
  0xFFFFFFFFFFFFFFFF
  Mon Feb 27 02:06:12.238 2017: [I]  ctr:
  0xFFFFFFFF0048E120   xer:
  0x00000000FFFFFFFF
  Mon Feb 27 02:06:12.260 2017: [I] exad:
  0x0000000000000000
  Mon Feb 27 02:06:12.270 2017: [I] 0x90000000079EF14
  pthread_kill() +
  0xD4
  Mon Feb 27 02:06:12.280 2017: [I] 0x90000000079E764
  _p_raise() + 0x44
  Mon Feb 27 02:06:12.352 2017: [I] 0x900000000039D68
  raise() + 0x48
  Mon Feb 27 02:06:12.360 2017: [I] 0x9000000000563E4
  abort() + 0xC4
  Mon Feb 27 02:06:12.368 2017: [I] 0x9000000000DDB40
  __assert_c99() +
  0x2C0
  Mon Feb 27 02:06:12.380 2017: [I] 0x100005A0C
  logAssertFailed() + 0x22C
  Mon Feb 27 02:06:12.392 2017: [I] 0x10058529C
  HandleMBDaemonToDie(MBDaemonToDieParms*) + 0x5C
  Mon Feb 27 02:06:12.400 2017: [I] 0x10057EA90
  Mailbox::msgHandlerBody(void*) + 0x3B0
  Mon Feb 27 02:06:12.408 2017: [I] 0x100161D24
  Thread::callBody(Thread*)
  + 0x104
  Mon Feb 27 02:06:12.420 2017: [I] 0x10000218C
  Thread::callBodyWrapper(Thread*) + 0xAC
  Mon Feb 27 02:06:12.432 2017: [I] 0x900000000786E10
  _pthread_body() +
  0xF0
  Mon Feb 27 02:06:12.440 2017: [I] 0xFFFFFFFFFFFFFFFC
  

Local fix

Problem summary

• Kernel assert going off: (nPrefetchedBuffers > 0) in
  fetch-vfs.C.   This can happen when count
  in nPrefetchedBuffers goes over 32767(MAX_INT16) which can
  happen when reading very large file.
  

Problem conclusion

• Change declaration for nPrefetchBuffers from Int16 to UInt16 so
  it would become negative when count go over MAX_INT16.   We have
   code in place trying to keep number of prefetch buffer from
  going over
  32K but it is possible for count to go over it under some
  conditions.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

  IV93813

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  SPECTRUM SCALE

• Fixed component ID

  5725Q01AP

Applicable component levels

• R411 PSY U884675

     19/04/30 I 1000