2524578 - Standstill scenario caused by lack of EG (extended global memory)

• The affected instance suffers from an apparent standstill situation. Most actions will just hang and new users cannot logon to the application server.
   
• You also notice that a few processes seem to be waiting for semaphore 39, in the snapshot generated in dispatcher trace file (dev_disp)
   
• One of the following warning entries is noticed at work process trace files:
   *** WARNING => MtxLock xxYYxxYYxxYY EMADM owner=<wp number> deadlock ? [mtxxx.c 2574]
   
  or
   
  *** WARNING => MtxLock xxYYxxYYxxYY EGPHxx owner=<wp number> deadlock ? [mtxxx.c 642]
   
• By extracting the C-STACK from the work process number displayed as the mutex holder (owner=<wp number>), the following (or similar) is seen:

(SemRq+0x596)[0x1565506]
(EgIGlobalLock+0x18)[0x1fc63b8]
(EgGetPtr+0x2d)[0x1fc65dd]
(EmHyperContextCreate+0x90)[0x1fbad20]
(ThEmInit+0xac)[0x15ab14c]
(ThSessionInit+0x2f1)[0x1573661]
(TskhLoop+0x1362)[0x1589af2]
(ThStart+0x2bc)[0x4ab317]
(DpMain+0x3b3)[0x1530a03]
 
or
 
<signal handler called> [0x7f94b12a8c00,0x11c00]
libpthread.so.0[S](__GI___nanosleep+0x2d)[0x7f94b12a84cd,0x114cd]
dw.sapP10_D30[S](MtxMicrosleep+0x36)[0x559924748306,0x26d8306]
dw.sapP10_D30[S](MtxLock_SPIN+0x2e9)[0x559924160239,0x20f0239]
dw.sapP10_D30[S](EgPhSetMmDiagmode+0x6d)[0x559924ac769d,0x2a5769d]
dw.sapP10_D30[S](EgSetMmDiagmode+0x8e)[0x559923cf5e38,0x1c85e38]

SAP NetWeaver release independent

semaphore 36, semaphore 39, mutex, freeze, hanging situation, extended global, SAP memory , KBA , BC-CST-DP , Dispatcher, Task Handler , BC-CST-MM , Memory Management , Problem