2882304 - IQ server and host OS start crashing frequently with a range of Memory Corruption type errors – SAP IQ

• Linux host OS (Operating System) running SAP  IQ suddenly starts having frequent  memory related errors that crash the IQ server, or the OS forcing a host reboot,  or both.
  
  
• A range of differing abort messages and stacks can be encountered.  The majority of which will include the error “IQ Internal Error Detected:  Memory Corruption” in the iqmsg file.
  
  
• The more  common aborts seen are in hos_mem.cxx with a rethrown from hos_throw.cxx  with an exception type hos_memexception

  Stacks at ….

  hos_mem::CheckFencesAndThrow()
  hos_memmgr::MarkPerm
  s_bmROCursor::s_bmROCursor
  s_mergeRHSReader::Init
  s_mergeFINALWriter::Init
  hs_dpInsert::Prepare
  dfiio_FP::PrepareFPs()
  dfiio_FP::PrepareOpus()
   
   
  hos_mem::CheckFencesAndThrow()
  hos_memmgr::MarkPerm
  s_bArray::s_bArray
  hs_dp::hs_dp
  db_IndexDef::OpenSIndexWithLock
  db_IndexDef::OpenSIndex
  db_CatalogSIndexPtr::Open
   
  hos_memmgr::FreeHosMem
  db_IndexDef::~db_IndexDef()
  db_IndexDef::~db_IndexDef()
  db_TableDef::~db_TableDef()
  db_TableDef::~db_TableDef()
  db_TableDef::Release
  db_CatalogChain::Flush
  
  
  

• The same server may also see signal 11 aborts at st_server.cxx  with stacks similar to the following…
  
  

  __do_global_dtors_aux_fini_array_entry
  hos_mem::PrintListElement
  hos_mem::Display
  hos_mem::CheckFencesAndThrow()
  hos_memmgr::MarkPerm
  RWCStringRef::getRep
  RWCString::replace
  dfe_Constant::Display
  
  

• Other host OS error characteristics seen are that the messages in the /var/log/messages file become disjointed and chronologically out of sequence.  Also look for "Bad page map" warnings in the OS message file similar to the following…

[ 3602.897107] BUG: unable to handle kernel NULL pointe [ 3603.032967] CPU: 14 PID: 57031 Comm: dbbackup Kdump:

[16064.292371] BUG: Bad page map in process khugepaged [16064.311233] CPU: 14 PID: 353 Comm: khugepaged Kdump:

• SAP IQ 16.0
  
  
• SAP IQ 16.1

sybase, crash, abort, failure, re-boot, "v2s_bmcontext* s_bmROCursor::_context", "Next Element In List", "s_bArrayLocater::_bCursor", RAM, processors, "hos_mem.cxx:249", corruption, RedHat, unix, "hard ware", "h/w" , KBA , BC-SYB-IQ , Sybase IQ , Problem