Recovery statistics program (DFH$STER)
The recovery statistics program, DFH$STER, replaces the statistics produced by the obsolete recovery utility, DFHRUP. DFH$STER retrieves global statistics from the Recovery Manager, formats and prints the information to the CICS® transient data destination, CSSL. DFH$STER enhances the Recovery Manager Domain recovery messages produced on a CICS emergency restart.
To install DFH$STER, define this program in the second phase PLTPI for your CICS region. For more information, see Writing initialization and shutdown programs.
Example of a report
DFHRM0201 10/29/99 11:10:40 IYK2Z1V1 0 backout-failed and 1 commit-failed UOWs were reconstructed
DFH$STER : CICS Restart Recovery Activity report
No. of syncpoints (forward) : 14
No. of syncpoints (backward) : 0
No. of resynchronizations : 0
No. of UOW indoubt failure shunts completed : 0
No. of UOW com/back failure shunts completed : 1
No. of UOW indoubt failure shunts outstanding : 0
No. of UOW com/back failure shunts outstanding : 0
DFH$STER : End of CICS Recovery activity report
Recovery Manager domain recovery messages
During an emergency restart, the Recovery Manager produces several messages detailing the state of inflight and shunted units of work (UOWs). For further information about the shunting of UOWs, see Units of work.
The messages are produced by the Recovery Manager after processing the system log, and they accurately represent the state of the CICS system workload at the time the cancellation, abend, or termination occurred. When the messages are produced, the Recovery Manager attempts to resolve the UOWs by committing, backing-out, or temporarily suspending completion (shunting) them.
- DFHRM200
<applid> xx Indoubt UOWs were reconstructed
After reading the system log, the CICS Recovery Manager found that there was the specified number of UOWs that were shunted indoubt, or were waiting indoubt at the time the system was terminated. Both types have suffered an indoubt failure with their recovery coordinator, for example, another CICS, or DBCTL. They must wait to be reconnected or resynchronized with the coordinator before they can commit or backout the changes made within a particular shunted UOW, that is to complete a syncpoint. Recovery coordinators will be across LU6.2, LU6.1, MRO (XM/IRC/XCF) links, or a product interface through the Resource Manager Interface (RMI) such as DBCTL, IBM® MQ or Db2®.
- DFHRM201
<applid> xx backout-failed and commit-failed UOWs were reconstructed
After reading the system log, the CICS Recovery Manager finds a specified number of UOWs shunted while awaiting the availability of a LOCAL recoverable resource that was updated by the task in question. The local resources to CICS are files, RLS files, TD queues, TS Queues, RDO objects and user. There are many reasons why a recoverable resource cannot be committed or backed out, for example, RLS server not available, I/O errors on data sets, coupling facility errors, backout exit failures. As with the message DFHRM200, the UOWs could not complete their syncpoint processing. However, they are not shunted while awaiting resolution of an indoubt failure.
- DFHRM202
<applid> xx inflight UOWs were reconstructed
After reading the system log, the CICS Recovery Manager finds a specified number of UOWs in progress. That is, the number of tasks that were executing in the CICS region before the system was terminated abnormally. These tasks are in between syncpoint requests and therefore have to be backed-out (rollbacked) or shunted awaiting resynchronization with an external resource manager such as DBCTL.