Explanation
The scheduler work area (SWA) manager detected
an uncorrectable error. Register 15 contains a reason code explaining
the error:
- Code
- Explanation
- 04
- The routine that called SWA manager requested an incorrect function.
- 08
- The routine that called SWA manager passed an incorrect SVA (SWA
virtual address). The SVA does not point to the beginning of a SWA
prefix or the SWA prefix was destroyed.
- 0C
- A SWA manager routine tried to read a record not yet written into
SWA.
- 10
- The system tried to read or write a block that is not 176 bytes
or the system tried to assign a block with a specified length of 0
or a negative number.
- 14
- The routine that called SWA manager specified an incorrect count
field. For move mode, an incorrect count is 0 for a READ, WRITE, or
ASSIGN function, and 00 for WRITE/ASSIGN.
- 18
- The routine that called SWA manager by issuing the QMNGRIO macro
specified both or neither of the READ or WRITE options.
- 1C
- The routine that called SWA manager tried to write into a SWA
block for the first time and has either passed a nonexistent ID, or
has failed to pass an ID at all and the block does not have an embedded
ID.
- 20
- The system tried to write a block using an incorrect pointer to
the block.
- 24
- An error occurred while the system was translating an SVA for
an SWA block that is above 16 megabytes.
- 28
- The routine that called the SWA manager specified zero on the
EPA parameter of the SWAREQ macro. The EPA parameter must be the
address of a fullword pointer to the EPA. Check the value specified
on the EPA parameter of the SWAREQ macro to ensure that it is the
address of a pointer to the EPA and not the EPA itself.
- 34
- An attempt has been made to allocate or delete an SWA block in
an unauthorized environment.
- 3C
- While attempting to process an assign request, the SWA manager
encountered a zero value as the pointer to the QMPA. The most likely
cause of this error condition is that a task running under the master
subsystem (SUB=MSTR) attempted a dynamic allocation in its own address
space, but that address space is not yet totally established.
Some
specific causes are as follows:
- Attempting to run a started JOB as SUB=MSTR or under a WLM initiator
(INIT), when the JOB to be started has an instream procedure (proc).
The converter issues a dynamic allocation to create a temporary PDS
to hold the procedure being read in.
- Attempting to run a started task (STC) as SUB=MSTR, but not all
of the data sets needed for that started task are cataloged in the
MASTER CATALOG. Either a JOBCAT or STEPCAT is included in the JCL,
or simply one of the data sets in the JCL is cataloged in a USER catalog
(usercat). Allocation issues a dynamic allocation for the required
catalog.
- The device type in the catalog for one of the data sets that is
started as SUB=MSTR is incorrect (for example, the information specifies
that the data set is on a 3380 device, but the data set is really
on a 3390 device.).
Corrective actions are:
- You cannot run started jobs with PROC statements as SUB=MSTR.
Do
one of the following actions:
- Remove the PROC statement, or
- Remove the JOB statement and run it as a started task
- Run it as SUB=JES2 (or JES3)
- Ensure that all data sets required by jobs or tasks running SUB=MSTR
are cataloged in the master catalog or are directly pointed to by
the specification of UNIT and VOL=SER parameters on the DD statement.
- Ensure that the device types for all required catalogued data
sets and any UNIT parameters on DD statements are correct.
System action
The task ends and the system writes a dump.
Operator response
Notify the system programmer.
System programmer response
If the problem recurs and if the program
is not in error, search problem reporting data bases for a fix for
the problem. If no fix exists, contact the IBM® Support Center. It is the responsibility
of the caller to SWA manager to take the dump, and if none was taken,
then a SLIP would need to be set up to capture the failure. Provide
the dump to the IBM Support
Center.
Programmer response
If the program that ended was a started
task, and the code in register 15 is X‘08’ or X‘0C’,
check the JCL for this task. For all started tasks, the JCL must either:
- refer to existing data sets that are cataloged in the master catalog;
or
- include unit and volume information on the DD statements for existing
data sets.
Modify the JCL to meet one of these requirements, and resubmit
the job.
Source
Scheduler work area (SWA) manager