OA55038: TCP/IP REPLICATE MIGRATION (APPROXIMATELY 490 VOLUMES) SUFFERINGBUFFER SHORTAGES EXHAUSTED TDMF MAXSOC THRESHOLD (544 SOCKETS)

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• An example Migration scenario resulting from a TCP/IP Buffer
  Shortage (other error conditions would cause the same problem
  when they trigger a large number of TCP/IP reconnection events.
  
  The current messages for SMFID xxxx follow.
  GTD1177I The source volume YI3E82 is mounted on
           device C7D2 on this system.
  GTD2840I Attempting to establish a remote
           session with 010.099.156.230 port
           63013.
  GTD2841I Established a TCP/IP session with
           010.099.156.230 port 63013.
  GTD3701I Remote connection established with
           010.099.156.230.
  GTD2406I All storage frames to migrate this
           volume have been successfully page
           fixed.
  GTD2403I This volume successfully selected for
           initialization and is a group volume.
  GTD2281I The Master system is starting the
           initialization process for a volume.
  GTD2283I The Master system is starting the
           migration process for a volume.
  GTD3528I Starting the copy of the source volume.
  GTD3790W TCP/IP transmission error encountered
           (3708/55).
  GTD2838I TCP/IP reconnection successful.
  GTD3790W TCP/IP transmission error encountered
           (3708/55).
  GTD2838I TCP/IP reconnection successful.
  GTD3790W TCP/IP transmission error encountered
           (3708/55).
  GTD2838I TCP/IP reconnection successful.
  GTD3790W TCP/IP transmission error encountered
           (3708/55).
  GTD2838I TCP/IP reconnection successful.
  GTD3790W TCP/IP transmission error encountered
           (3708/55).
  GTD2845E TCP/IP API SOCKET request rejected.
           Return code 1002.
  GTD2293I The Master system is starting the
           termination process for a volume.
  GTD3791E TCP/IP reconnection was unsuccessful.
  GTD2303I The Master system has completed the
           migration process for a volume.
  GTD2410I All storage frames to migrate this
           volume have been successfully page
           freed.
  
  GTD3790W TCP/IP transmission error encountered
  (message/error-code).
  Explanation: The IP data transmission service has
  detected an error condition while writing to the
  REMOTE session or reading an acknowledgment, and
  TDMF will attempt to reconnect to the REMOTE
  session for this volume.
  The message contains the message number and API
  error code relating to the failed read or write request.
  System action: Volume migration will be suspended
  until the TCP/IP connection can be reestablished.
  
  GTD3708E IP WRITE error (err-code).
  Explanation: The IP Connect processor received the
  identified IP error code from its EZASMI TYPE=WRITE
  macro request.
  System action: The volume migration is terminated.
  Programmer response: Determine the reason for the
  IP error code, take appropriate action and retry the
  migration. If unable to determine the cause of the error
  contact IBM Technical Support for assistance.
  Problem determination: Error codes returned by the
  TCP/IP Application Programming Interface (API) for
  socket calls are documented in the IBM manual: z/OS
  Communications Server IP Application Programming
  Interface Guide.
  
  Error number          55
  Message name          ENOBUFS
  Socket API type       All
  Error description     No buffer space is available.
  Programmer's response Check the application for
                        massive storage allocation call.
  
  GTD2845E TCP/IP API SOCKET request rejected.
  Return code err-code.
  Explanation: The EZASMI macro detected an error in
  the TCP/IP environment. Error codes returned by the
  TCP/IP Application Programming Interface (API) for
  socket calls are documented in the IBM manual: z/OS
  Communications Server IP Application Programming
  Interface Guide.
  System action: Migration of this volume is terminated.
  Programmer response: Contact IBM Technical Support
  for assistance.
  
  Error number          1002
  Message name          EIBMSOCKOUTOFRANGE
  Socket API type       Socket, Accept, Takesocket
  Error description     A new socket cannot be created
                        because the MAXSOC value, which
                        is specified on the INITAPI call,
                        has been reached.
  Programmer's response Take either one of the following actions:
                        - Verify whether all open sockets are
                          intended to be in use.
                        - Increase the MAXSOC value to a value
                          that is appropriatefor the current
                          workload. If the default value is
                          currentlybeing used, you might be
                          required to add the INITAPI call.
  

Local fix

• Correct the underlying TCP/IP error conditions, e.g., increase
  buffer pool size.
  
  Restart the TDMF Migration
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: TDMF users performing remote replications    *
  *                 using TCP/IP                                 *
  ****************************************************************
  * PROBLEM DESCRIPTION: TCP/IP buffer shortages can cause       *
  *                      message GTD2854E with a return code     *
  *                      1002 (EIBMSOCKOUTOFRANGE) and thus      *
  *                      preventing TDMF recovery. This leads to *
  *                      many terminated volumes.                *
  ****************************************************************
  * RECOMMENDATION: Install PTF                                  *
  ****************************************************************
  Data connection sockets were not properly closed when given to a
  subtask. When recovery was attempted an additional socket was
  requested which eventually caused an outage of available sockets
  and prevented further recovery.
  

Problem conclusion

• GTDXSL1 was changed to properly close a given socket. In case a
  transmission error occurs this socket will now be available for
  reusage and TDMF no longer runs out of socket.
  

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:

  UA96594 UA96595

Modules/Macros

•    GTDMSGN  GTDMSGV  GTDXSL1  GTDXSR0
  

Fix information

• Fixed component name

  TDMF Z/OS

• Fixed component ID

  TDMFZOS01

Applicable component levels

• R560 PSY UA96594

     UP18/06/19 P F806

• R570 PSY UA96595

     UP18/06/19 P F806

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.