Categories of problem

The following categories of problem (in order of ascending impact on the user) may be encountered by the CICS® log manager.

Those problems within z/OS® System Logger that z/OS System Logger resolves for itself
CICS has no involvement in this category and might only experience the problem as an increase in response times.
z/OS System Logger is unable to satisfy the CICS log manager's request immediately
This problem state can be encountered:
  • For a log stream that uses a coupling facility structure, on a STRUCTURE FULL condition, where the coupling facility has reached its capacity before offloading data to DASD. This state may also be encountered during the rebuilding of a coupling facility structure.
  • For a DASD-only log stream, on a STAGING DATA SET FULL condition, where the staging data set has reached its capacity before offloading data to secondary storage.

If either of these conditions occur, CICS issues message DFHLG0771 (for a general log) or DFHLG0777 (for a system log). The CICS log manager retries the request every three seconds until the request is satisfied. Typically, this can take up to a minute.

z/OS System Logger fails
If z/OS System Logger fails, CICS is abended. If the system log has not been damaged, a subsequent emergency restart of CICS should succeed.
The CICS system log has been damaged

If a return code implies that the CICS system log has been damaged, CICS is quiesced, meaning transactions are allowed to run to completion as far as possible, with no further records being written to the system log. To get CICS back into production, you must perform an initial start. However, before doing so you may want to perform a diagnostic run, to gather information for problem diagnosis; see Dealing with a corrupt system log.

If a return code implies damage to a forward recovery log or autojournal, all files using the log stream are quiesced and their transactions run to completion. Message DFHFC4800, DFHFC4801, or DFHFC4802 is issued. User transactions writing journal records to the log stream experience a write error.

For a forward recovery log, before you can continue to use the log stream, you must:
  1. Take an image copy of all data sets referencing the log stream.
  2. Redefine the log stream.
  3. Unquiesce the data sets using the affected logs. You may then explicitly open the files but they open automatically at the first READ or WRITE if they are in a CLOSED ENABLED state after the unquiesce.
For an autojournal, before you can continue to use the log stream, you must:
  1. Try to read and recover data from the damaged autojournal.
  2. Redefine the log stream.