2701674 - Database hang during TSM backup with PD_TYPE_SQLO_NLCK lock

Database hang after starting a backup with TSM.
The following error is found in the db2diag.log

PID     : <PID>              TID : <TID          PROC : db2sysc
INSTANCE: db2<sid>               NODE : 000           DB   : <SID>
HOSTNAME: <hostname>
EDUID   : <EDUID>                EDUNAME: db2agent (<SID>) 0
FUNCTION: DB2 UDB, oper system services, sqloNLCKGetInstanceWideNamedLock, probe:10
MESSAGE : Semaphore wait timed out and recovery semaphore not detected yet.
DATA #1 : Codepath, 8 bytes
...
DATA #2 : unsigned integer, 8 bytes
...
DATA #3 : unsigned integer, 8 bytes
...
DATA #4 : unsigned integer, 8 bytes
...
DATA #5 : OSS Named Lock, PD_TYPE_SQLO_NLCK, 40 bytes
   Lock Scope        : Instance-wide
   Semaphore Handle  : 780000000D6A8D8
   Resource Slot     : 780000000D6A880
   File Handle       : N/A
   Lock State        : 0x3, (Initialized)
   Used by Engine    : Yes
   Auto Create File  : No
CALLSTCK: (Static functions may not be resolved correctly, as they are resolved to the nearest symbol)
  [0] 0x090000000D91CD94 sqloNLCKGetInstanceWideNamedLock__FCP16SQLO_NLCK_HANDLECUlCi + 0x554
  [1] 0x090000000D91C640 sqloNLCKLock + 0x2C0
  [2] 0x090000000D915AB4 sqluhLock__FP12SQLUH_HANDLE + 0x1B4
  [3] 0x090000000D917A34 sqluhAllocateHandleVerifyHistoryFilesAndOptToRecover__FPP12SQLUH_HANDLEPCcCbCUlT3PUi + 0xB4
  [4] 0x090000000D91845C sqluhOpen__FPP12SQLUH_HANDLEPCcUlPUib + 0x7C
  [5] 0x090000000D92C450 adminListHist + 0x250
  [6] 0x090000000D92DD14 adminListHistTrusted + 0x414
  [7] 0x090000001885EE8C sqloInvokeFnArgs + 0x592C

Db2 Linux Unix and Windows

IBM, tape, history, prune , KBA , BC-DB-DB6 , DB2 Universal Database for Unix / NT , Problem