The CICSplex Unit-of-work DSN Failure
Details attribute group enables you to inquire on the reasons why
UOWs were shunted because of a failure during syncpoint associated
with a specified data set. Failures during syncpoint processing
result in locks held by the UOW against the data set (or data sets)
which suffered the failure being retained. Thus, when a failure is
reported by this command, it also indicates the presence of retained
locks.
Cause Indicates
which failed component has caused the UOW to have retained locks for
this data set. These are the values:
- CACHE
- A VSAM RLS cache structure, or connection to it, has failed.
- CONNECTION
- An intersystem connection error has caused the UOW to fail while
indoubt. The name of the system to which connectivity was lost is
returned on the SYSID parameter and its netname is returned on the
NETNAME parameter. CICS returns additional information in the REASON
parameter about the connection failure.
- DATASET
- The backout of a UOW has failed for this data set. The reason
for the data set failure is returned in the REASON parameter.
- RLSSERVER
- The SMSVSAM server has failed. The reason for the data set failure
is returned in the REASON parameter.
- UNDEFINED
- The UOW is probably being retried. This occurs following a SET
DSN RETRY command, or automatically when the failed resource returns.
It also occurs following an emergency restart.
CICS Region Name Indicates
the job name or modify ID of the CICS region being monitored. Each
CICS region in a z/OS operating system has a unique name. This name
is an alphanumeric string, with a maximum of eight characters, and
case-sensitive. CICS region names are always in uppercase characters.
Dataset name Is
the name of the data set that has suffered a backout failure in the
current UOW. It is a 44 character value.
Connected System ID Is
the ID of the connected CICS region.
Network name Is
the network name of the remote system to which connectivity has been
lost. It is an eight character netname.
Origin Node Is
the combination of z/OS System ID (SMFID) and CICS region name. The
value format is an alphanumeric string, maximum 32 bytes, and case-sensitive.
When a situation is true, the system compares the origin node name
in the attribute to the names of managed systems assigned to the item.
If the origin node name matches the name of a managed system, the
status of the item changes accordingly. If the names of the origin
node and the managed systems do not match, the status of the item
remains unchanged.
Reason Is the
specific failure reason indicator. These are the values:
- BACKUPNONBWO
- Backout of the updates made to the data set by the UOW failed
because a non-BWO backup of the data set was in progress while the
UOW was being backed out. When the backup completes, CICS automatically
retries the UOW.
- COMMITFAIL
- An error occurred at some point when RLS locks were in the process
of being released. This is an error that is typically be resolved
by recycling the SMSVSAM server (which should happen automatically).
The locks were acquired as a result of recoverable requests having
been issued against the data set.
- DATASETFULL
- No space is available on the direct access device for adding records
to a data set. You need to reallocate the data set with more space
and then retry the backout using the SET DSNAME RETRY command.
- DEADLOCK
- (For non-RLS data sets only.) A deadlock was detected during backout.
This is a transient condition that goes away if the backout is retried.
- DELEXITERROR
- Backout of a write to an ESDS failed because a logical delete
global user exit program was not enabled, or a logical delete global
user exit program decided not to execute the logical delete.
- FAILEDBKOUT
- This occurs as a result of a severe error being identified during
backout, and is possibly an error in either CICS or VSAM. The problem
might go away if the backout is retried. Note that CICS performs some
first-failure data capture (FFDC) at the point where the error is
first detected.
- INDEXRECFULL
- A larger alternate index record size needs to be defined for the
data set. This error occurs when a unique alternate index key, for
a non-RLS data set, has been reused and CICS is now backing out the
request which had removed that key value.
- INDOUBT
- The unit of work had issued recoverable requests against the data
set, and has now failed indoubt. The connection to the coordinating
system needs to be reestablished.
- IOERROR
- A hard I/O error occurred during backout. To correct this error,
restore a full backup copy of the data set and perform forward recovery.
If you use CICSVR as your forward recovery utility, the backout is
automatically retried for an RLS data set. For a non-RLS data set,
use the SET DSNAME (...) RETRY command to drive the backout retry.
- LCKSTRUCFULL
- An attempt to acquire a lock during backout of an update to this
data set failed because the RLS lock structure was full. You must
allocate a larger lock structure in an available coupling facility
and rebuild the existing lock structure into it, then use the SET
DSNAME (...) RETRY command to drive the backout retry.
- NOTAPPLIC
- The CVDA for CAUSE is not CONNECTION, RLSSERVER, or DATASET.
- OPENERROR
- Error on opening the file for backout. A console message notifies
you of the reason for the open error. One likely reason could be that
the data set was quiesced.
- RLSGONE
- An error occurred when backing out of the UOW because the SMSVSAM
RLS server was inactive. This might also be the reason why the UOW
went into backout originally. This is an error is resolved by recycling
the server (which should happen automatically). Generally, when the
server recovers, the units of work are retried automatically. In very
exceptional circumstances, it might be necessary to issue a SET DSNAME(...)
RETRY command to retry the units of work that were not retried when
the server returned.
- RRCOMMITFAIL
- An error occurred while RLS locks for the unit of work were being
released. For this data set, the locks being released were all repeatable
read locks, so if the failure was because of the RLS server being
unavailable, the locks are released. If the failure was because of
some other error from the SMSVSAM server, the locks might still be
held.
- RRINDOUBT
- The unit of work had issued repeatable read requests against the
data set, and has now failed with an indoubt condition. The locks
are released, so this failure does not prevent you from running a
batch job against the data set. However, if you want to open the data
set in non-RLS mode from CICS, you need to resolve the indoubt failure
before you define the file as having RLSACCESS(NO). If the unit of
work has updated any other data sets, or any other resources, try
to resolve the indoubt failure correctly. If the unit of work has
only performed repeatable reads against VSAM data sets and has made
no updates to other resources, it is safe to force the unit of work
using the SET DSNAME or SET UOW commands.
RLS Access Is
the RLS indicator for the connection. These are the values:
- NOTRLS
- The last open in this CICS region was in non-RLS mode.
- RLS
- The last open in this CICS region was in RLS mode.
System ID Indicates
the four character name that uniquely identifies an active z/OS operating
system within a given CICSplex. The value format is an alphanumeric
string, maximum four characters, and case-sensitive. z/OS System IDs
are always in uppercase characters.
Unit-of-work ID Is
the identifier token for the current unit-of-work.