Recovering from a failure during a log RBA read request

A failure might occur during a log RBA read request. For example, because of problems with the BSDS, the requested RBA, which contains the dropped log data set, cannot be read. You can recover from the situation.

Symptoms

Abend code 00D1032A is issued, and message DSNJ113E is displayed:

DSNJ113E RBA log-rba NOT IN ANY ACTIVE OR ARCHIVE
        LOG DATA SET.  CONNECTION-ID=aaaaaaaa, CORRELATION-ID=aaaaaaaa

Causes

The BSDS is wrapping around too frequently when log RBA read requests are submitted; when the last archive log data sets were added to the BSDS, the maximum allowable number of log data sets in the BSDS was exceeded. This caused the earliest data sets in the BSDS to be displaced by the new entry. Subsequently, the requested RBA containing the dropped log data set cannot be read after the wrap occurs.

Resolving the problem

System programmer response:
  1. Stop Db2 with the STOP DB2 command, if it has not already been stopped automatically as a result of the problem.
  2. Check any other messages and reason codes that are displayed, and correct the errors that are indicated. Locate the output from an old execution of the print log map utility, and identify the data set that contains the missing RBA. If the data set has not been reused, run the change log inventory utility to add this data set back into the inventory of log data sets.
  3. Increase the maximum number of archive log volumes that can be recorded in the BSDS. To do this, update the MAXARCH system parameter value as follows:
    1. Start the installation CLIST.
    2. On panel DSNTIPA1, select UPDATE mode.
    3. On panel DSNTIPT, change any data set names that are not correct.
    4. On panel DSNTIPB, select the ARCHIVE LOG DATA SET PARAMETERS option.
    5. On panel DSNTIPA, increase the value of RECORDING MAX.
    6. When the installation CLIST editing completes, rerun job DSNTIJUZ to recompile the system parameters.
  4. Start Db2 with the START DB2 command.