IJ28321: ASSERT EXP(ISVALIDSOCKET(SOCK))

APAR status

• Closed as program error.

Error description

• Due to a change in Spectrum Scale 5.0.5.1 OpenDevice
  call was not followed by a CloseDevice, thus
  leaving unused open file descriptors open.
  When these reach a threshold of >64k, the mmfsd will
  assert on opening a socket or a file with :
  [X] logAssertFailed: isValidSocket(sock)
  causing this node to recycle and temporarily lose access
  to the filesystem.
  

Local fix

• one cause of using up FDs is the periodical run
  diskIOcheck every 300 sec (default) by disabling this
  via:
   # mmchconfig diskIOCheckInterval=0 -i
  these check will not anymore add to the leakage.
  

Problem summary

• logAssertFailed:isValidSocket(sock) line 2661
  of file  thread.C
  

Problem conclusion

• Benefits of the solution:
  No more assert
  
  Work Around:
  No stable workaround to avoid this, but
  stopping the unmount of the file system
  and a stable NSD server node with stable
  disks could help.
  
  Problem trigger:
  File system unmount operations and
  disk path rediscovery or NSD disk stats
  changes.
  
  Symptom: mmfsd crash
  
  Platforms affected: All Operating Systems
  
  Functional Area affected: All Scale users
  
  Customer Impact: Critical
  
  Changed Externals:None
  

Temporary fix

Comments

APAR Information

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

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

Fix information

• Fixed component name

  SPEC SCALE STD

• Fixed component ID

  5737F33AP

Applicable component levels