IT37723: Mishandling of return codes from OS calls might result in Q filecorruption after file descriptor limit is hit

APAR status

• Closed as program error.

Error description

• If the OS limit on file descriptors (RLIMIT_NOFILE) is hit, then
  OS calls such as open() and pipe() fail with errno=EMFILE.  The
  remedy is for the administrator to increase the limits (eg. by
  setting the ulimit 'nofile' setting).
  
  While reviewing the MQ code for possible defects, it was found
  that, in a few places, there was inadequate checking of the
  return codes from OS calls such as open().  In a scenario in
  which the file descriptor limit has been hit, the MQ code could
  also close a file descriptor used for communication between a
  parent process and a child process.
  
  As a result of these things, there are undefined behaviours in a
  queue manager process after it has suffered such a condition,
  including the writing of bad data into a Queue file.  If this
  has happened, then failure data capture (FDC) records are
  written saying "object damaged".
  

Local fix

• It is strongly recommended that OS file descriptor limits are
  set appropriately based on the expected workload of the queue
  manager and appropriate load testing performed.
  

Problem summary

• ****************************************************************
  USERS AFFECTED:
  All users of MQ on Linux and UNIX platforms.
  
  
  Platforms affected:
  AIX, Linux on Power, Linux on x86-64, Linux on zSeries
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  The IBM MQ code had inadequate checks on the return codes from
  OS calls such as open().  Also the code could close a file
  descriptor that was later needed. Improper handling of file
  handles under this scenario could result in data being written
  to a file handle that had been reassigned to another file - in
  this regard, there is a possibility that arbitrary data could be
  written into a queue file, invalidating that queue file.
  

Problem conclusion

• The IBM MQ code has been corrected.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
  Version    Maintenance Level
  v9.0 LTS   9.0.0.13
  v9.1 LTS   9.1.0.11
  v9.2 LTS   9.2.0.5
  v9.x CD    9.2.5
  
  The latest available maintenance can be obtained from
  'WebSphere MQ Recommended Fixes'
  http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006037
  
  If the maintenance level is not yet available information on
  its planned availability can be found in 'WebSphere MQ
  Planned Maintenance Release Dates'
  http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006309
  ---------------------------------------------------------------
  

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

  MQ BASE V9.2

• Fixed component ID

  5724H7281

Applicable component levels