nAMgr access violation crashes server

  • I called this intermittent because it doesn’t happen all the time. Today, however, it happened five times. Not good for the production system users. Server is R8.5FP1 on windows server 2003 R2 SP2.

Here’s the “FATAL THREAD” bit of the NSD:

############################################################

FATAL THREAD 8/14 [ nAMgr: 1180: 1584]

FP=0x1accfc54, PC=0x600074a2, SP=0x1accfc38

stkbase=1acd0000, total stksize=32768, used stksize=968

EAX=0x1accfc4c, EBX=0x00000000, ECX=0x00000000, EDX=0x00000004

ESI=0x00000000, EDI=0x13ef3d6e, CS=0x0000001b, SS=0x00000023

DS=0x00000023, ES=0x00000023, FS=0x0000003b, GS=0x00000000 Flags=0x00010246

Exception code: c0000005 (ACCESS_VIOLATION)

############################################################

@[ 1] 0x600074a2 nnotes.OSLocalAllocNZ@8+50 (108,1accfca8)

@[ 2] 0x601866c1 nnotes.OSLocalAlloc@8+17 (108,1accfca8)

@[ 3] 0x600c47d2 nnotes.SECRandomNoSem@12+386 (0,0,1accfcbc)

@[ 4] 0x60034c67 nnotes.GetNSFT@4+71 (0)

@[ 5] 0x608605e5 nnotes.NSFHealthCheck@8+21 (5,4)

@[ 6] 0x60019d02 nnotes.NSFNoteClose@4+98 (35)

@[ 7] 0x61cc5a66 nlsxbe.ANNote::~ANNote+230 (0)

@[ 8] 0x61cc55eb nlsxbe.ANNote::`scalar deleting destructor’+11 (1)

@[ 9] 0x61d49ec4 nlsxbe.Java_lotus_domino_local_NotesBase_RecycleObject@12+292 (197bc200,1991fda0,1991fd9c)

############################################################

PASS 2 : FATAL THREAD with STACK FRAMES 8/14 [ nAMgr: 1180: 1584]

FP=1accfc54, PC=600074a2, SP=1accfc38

stkbase=1acd0000, total stksize=32768, used stksize=968

Exception code: c0000005 (ACCESS_VIOLATION)

############################################################

Disassembly of c. 10 instructions before and after faulting address 600074a2:



    6000747c c745f80c000000   mov     dword ptr [ebp+0xf8],0xc  ss:1ca2a79a=6543206c

    60007483 e858ffffff       call    600073e0

    60007488 8945e8           mov     [ebp+0xe8],eax            ss:1ca2a79a=6543206c

    6000748b c745e400000000   mov     dword ptr [ebp+0xe4],0x0  ss:1ca2a79a=6543206c

    60007492 6a00             push    0x0

    60007494 6a04             push    0x4

    60007496 8d4df4           lea     ecx,[ebp+0xf4]            ss:1ca2a79a=6543206c

    60007499 51               push    ecx

    6000749a e851a1ffff       call    600015f0

    6000749f 8b55e8           mov     edx,[ebp+0xe8]            ss:1ca2a79a=6543206c

FAULT ->600074a2 833a00 cmp dword ptr [edx],0x0 ds:00000004=???

    600074a5 751e             jnz     600074c5

    600074a7 8b45e8           mov     eax,[ebp+0xe8]            ss:1ca2a79a=6543206c

    600074aa 50               push    eax

    600074ab e8b0eb1700       call    60186060

    600074b0 668945ec         mov     [ebp+0xec],ax                 ss:1ca2a79b=4320

    600074b4 0fb74dec         movzx   ecx,word ptr [ebp+0xec]       ss:1ca2a79b=4320

    600074b8 85c9             test    ecx,ecx

    600074ba 7409             jz      600074c5

    600074bc 668b45ec         mov     ax,[ebp+0xec]                 ss:1ca2a79b=4320

    600074c0 e944010000       jmp     60007609

    600074c5 ba7eff0000       mov     edx,0xff7e
  • Naturally the entire thing can be produced on-demand. Sorta like Netflix.

  • The crashing application is a Java-based system-to-system interface written for R5 and migrated along to various newer versions, though we skipped R7. Java code is all Designer-based, there are no external JARs involved. It was pretty darned reliable until we put it on R8.5, then the troubles started.

  • I’ve tossed a ton of JVM and overall memory debugging bits into the server’s notes.ini, hoping to capture something wonky there.

Any thoughts? Thanks for your time…

Subject: Did you ever got any response to this ??

please mail me rahuremote@gmail.com

Thanks a ton