IJ26510: LOGASSERTFAILED: !"OLDDISKADDRFOUND.COMPADDR(*OLDDISKADDRP)"

APAR status

• Closed as program error.

Error description

• Spectrum Scale daemon crashes, filesystem gets unmounted
  on the node. The daemon crashes because of daemon code
  hit an assert which is saying the disk address is not
  expected.
  
  Reported in:
  Spectrum Scale 5.0.4.x
  
  Error message:
   [X] logAssertFailed:
  !"oldDiskAddrFound.compAddr(*oldDiskAddrP)"
   [X] return code 0, reason code 0, log record tag 0
   [X] *** Assert
  exp(!"oldDiskAddrFound.compAddr(*oldDiskAddrP)") in line
  11122 of file
  /project/spreltac504/build/rtac504s001a/src/avs/fs/mmfs/t
  s/fs/metadata.C
   [E] *** Traceback:
   [E]         2:0x55776C34CD78 logAssertFailed + 0x418 at
  ??:0
   [E]         3:0x55776C0AB24D
  FileMetadata::updateDataBlockDiskAddr(long long,
  fsDiskAddr const*, fsDiskAddr, unsigned int, int,
  indBlockDesc*, unsigned int*, fsDiskAddr*) + 0xF8D at
  ??:0
   [E]         4:0x55776BF1AA35
  BufferDesc::commitAssigned(int, unsigned int) + 0x1D5 at
  ??:0
   [E]         5:0x55776C12E844
  OpenFile::handleBufferFlush(int, BufferDesc*, long long,
  long long, ByteRange const&, int*, int*, ByteRange*, long
  long*, int*, int*, long long*, cxiUioAio_t**) + 0x1904 at
  ??:0
   [E]         6:0x55776C12F415
  OpenFile::flushBufferAddrs() + 0xE5 at ??:0
   [E]         7:0x55776C12FC3D
  FileMetadata::flushFileMetadata(int, ByteRange const&,
  ByteRange*, unsigned int*) + 0x6CD at ??:0
   [E]         8:0x55776C132143
  FileMetadata::flushFile(int, ByteRange const&, int*,
  int*, ByteRange*) + 0x9B3 at ??:0
   [E]         9:0x55776C132CBB SFSSyncFile(StripeGroup*,
  long long, unsigned int, int, ByteRange const&,
  OpenFile*) + 0x37B at ??:0
   [E]         10:0x55776C11DA01
  HandleMBFSyncFile(MBFSyncFileParms*) + 0xD1 at ??:0
   [E]         11:0x55776BE53F83
  Mailbox::msgHandlerBody(void*) + 0x363 at ??:0
   [E]         12:0x55776BE37703 Thread::callBody(Thread*)
  + 0x63 at ??:0
   [E]         13:0x55776BE24692
  Thread::callBodyWrapper(Thread*) + 0xA2 at ??:0
   [E]         14:0x7FC7F88F9DD5 start_thread + 0xC5 at
  ??:0
   [E]         15:0x7FC7F79FDEAD __clone + 0x6D at ??:0
  

Local fix

Problem summary

• GPFS daemon crashes and the filesystem
  gets unmounted.
  The GPFS daemon crashes
  because the daemon code hit an
  assert which is saying the disk
  address is not expected.
  

Problem conclusion

• Benefits of the solution: 
  No crash, and the file system is keeped active
  Work around:
  None
  Problem trigger:
  Failed writes and seeks within a file
  Symptom:
  Scale daemon crashes with logAssertFailed:
  !"oldDiskAddrFound.compAddr(*oldDiskAddrP)",
  and the file system gets unmounted.
  Platforms affected:
  All
  Functional Area affected:
  Core
  Customer Impact:
  Critical
  

Temporary fix

Comments

APAR Information

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

  IJ25547

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

  IJ26956

Fix information

• Fixed component name

  SPEC SCALE STD

• Fixed component ID

  5737F33AP

Applicable component levels