z/OS MVS System Codes
Previous topic | Next topic | Contents | Contact z/OS | Library | PDF


09A

z/OS MVS System Codes
SA38-0665-00

09A

Explanation

Global resource serialization encountered an error.

Note: The X'09A' abend code does not propagate to GQSCAN's caller.

The two low-order bytes of register 15 contain a hexadecimal function and reason code:

Code
Explanation
1208
Issued when the GRS receives notification for a global resource that is not known to GRS.
1210
Issued when the GRS receives notification for a global ENQ or DEQ request that is not known to GRS.
1904
An invalid control block was detected while running in GRS=STAR mode.
2204
Issued by when the system cannot create the GRS address space.
2505
GRS could not create a dependent enclave for the GRS EQDQ monitor to use because of an unexpected return or reason code from the WLM IWMECREA service. The IWMECREA return and reason codes were placed in registers 2 and 3 before the abend is issued.
3106
Issued by the GRS Lock Structure Complete Exit and indicates the resource specified in the exit is not known to GRS.
3705
Issued when GRS STAR Mode global GQSCAN processing could not add the ISGQSCAN dataspace to the DU-AL.
xxD5
Issued when the GRS Message Exit failed to receive an incoming message through the IXCMSGI macro.
46FA
Issued by ISGCDSP as part of "DISPLAY GRS" D GRS processing. X'FA' indicates a non-zero return code was received from MSGSERV. This typically occurs when the user issues a form of the command that generates too much output, "D GRS,ALL" or "D GRS,RES=(*,SYSDSN)", for example.
4814
An error occurred in the Queue Merge processing when this system tried to rejoin the GRS ring.
4818
An error occurred in Queue Merge processing when this system tried to rejoin the GRS ring.
6324
Issued when an attempt is made to initiate I/O on a CTC that has I/O still pending.
68C8
Indicates a non-zero return code from an IXCMOD request. GRS issues IXCMOD to change its XCF monitoring interval.
68CA
An IXCPART partition request returned a return code or reason code that is not expected. GRS issues IXCPART to remove a system from the sysplex.
81xx or 85xx
An error occurred in storage manager allocation processing while trying to allocate a cell.
82xx or 87xx
An error occurred in storage manager deallocation processing while trying to deallocate a cell.
A114
Issued when the GQSCAN RESNAME specification contains a QNAME ADDR and QNAME length of 0-7, but with RNAME ADDR, RNAME length, and specific/generic omitted (for example, RESNAME=(QNAME,,,,7)).
A142
Indicates a GQSCAN with SYSNAME was in progress, but the SYSNAME was unknown. This can occur when normal ENQ/DEQ processing changes control blocks while GQSCAN processing is looking at them. A dump is produced because recovery cannot determine the cause the cause of the changing control blocks (for example, if they changed because of normal ENQ/DEQ activity, or if they changed because of an overlay or other serious event).
A2xx
ISGQSCNR, which is the GQSCAN service's functional recovery routine, converted the ABEND.
A220
An undefined error occurred during execution of the GQSCAN service, which is the GQSCAN service routine. The system saves the original completion code in register 2 and the original reason code in register 3.
A22C
An unexpected error occurred when the ISGQSCAN information routine was activated to obtain requestor information from the service address space. GQSCAN recovery will not percolate to the GQSCAN caller's recovery since retry is requested.
C50C
The GRS STAR sysplex wide global resource contention notification server task, ENF 51 issuer, has determined that the length of its processing queue has reached a critical threshold. This indicates that the task can not keep up with the amount of work asked of it. This could be due to dispatch priority, a large amount of sysplex wide contention, or a problem with the server.
C544
Function code C5 is module ISGGCN the contention notification task. Subfunction 44 indicates GQSCAN. This reason code indicates a failure occurred in GQSCAN during contention notification processing.
Others
Internal error. IBM® might request this information for diagnosis. The contents of registers 3 through 5 may contain diagnostic data.

System action

System processing continues. The system usually writes a logrec data set error record and issues an SVC dump. For reason codes X'4814' and X'4818', the system will attempt to force a dump on the system in the complex which was assisting with the rejoin process.

For C50C: The contention notification task:
  1. Takes an SVC dump to identify the instance and to gather diagnostic data.
  2. Prevents any more elements from being queued.
  3. Informs ENF listeners that the ENF 51 data is temporarily suppressed and that some data could be lost.
  4. Purges all queued elements.
  5. Gives up ownership of the sysplex wide function. Because ownership is relinquished, another contention notification server on a different sysplex system, who may or may not have more resources, could become the sysplex wide notification server. If there is no other system, the same system will continue being the owner. Use the DISPLAY GRS command, found in z/OS MVS System Commands to determine which system is the notification server.
  6. The new owner will inform ENF listeners that the data is valid again.

Operator response

For C50C: Inform the system programmer.

For some 09A abends, you should obtain a dump of the GRS address space from each system in the global resource serialization complex. The IBM Support Center will request the dumps for diagnosis. The following function and reason codes require a dump from each system in the complex:

  • 43yy
  • 44yy
  • 45yy
  • 62yy
  • 65yy
  • 66yy
  • 67yy
  • 6Ayy
  • 6Cyy
  • C1yy
  • E1yy
  • E3yy
You do not need to obtain a dump for the codes above when yy is one of the following:
  • 45
  • 90
  • B2
  • B6
  • E0
  • EC
  • FA
  • FC

System programmer response

If the problem persists, search problem reporting data bases for a fix for the problem. If no fix exists, contact the IBM Support Center. Provide the function and reason codes, the logrec data set error record, and the dump. For the appropriate reason and function codes, provide a dump of the GRS address space dumps for each system in the global resource serialization complex. Provide all dumps taken by GRS recovery for the ABEND.

For C50C, before you contact the IBM support center, use the monitors to determine if an application created an abnormal amount of contention. If so, consider changing the application, or use the following installation exits to suppress notification of specific high frequency contention cases:
  • ISGCNFXITSYSTEM
  • ISGCNFXITSYSPLEX.
Also, check for any possible dispatch priority issues.

Programmer response

Do one of the following:
  • If mmff is 4808, 480C, or 4810, this is probably a user error. Correct the problem and resubmit the job.
  • If mmff is 46FA, check to see if the problem was caused by a shortage of WTO buffer space. If not, search problem reporting databases for a fix for the problem. If no fix exists, contact the IBM Support Center and provide the return code in register 2.
  • If the reason and function codes are A2xx, the problem is with the GQSCAN macro. Respecify the GQSCAN macro.

If the problem persists, search problem reporting data bases for a fix for the problem. If no fix exists, contact the IBM Support Center.

Source

Global resource serialization

Go to the previous page Go to the next page




Copyright IBM Corporation 1990, 2014