Ingres service stuck in "Starting" status

Document ID : KB000053426
Last Modified Date : 14/02/2018
Show Technical Document Details

Description

The Ingres service is not starting or is stuck in starting. An 'E_DM942A_RCP_ DBINCON_FAIL' error such as the following can be seen in the Ingres error log:

E_DM942A_RCP_ DBINCON_FAIL - One or more failures during RCP recovery has prevented all open databases from being properly recovered. Due to the existence of the II_DMFRCP_ STOP_ON_INCONS_DB flag, the RCP will bring down the installation rather than mark any databases inconsistent. Previous messages in the error log and the RCP log file may help diagnose the recovery problems.

If the problem is resource- or environment-related, the problem should be fixed and the system restarted.

Otherwise, II_DMFRCP_STOP_ON_INCONS_DB should be unset (or deassigned) and the RCP allowed to mark the database(s) inconsistent. A rollforward will be required in this case.

Solution

Perform the following steps to recover Ingres:

  1. While ingres is down, issue the command: ingunset II_DMFRCP_STOP_ON_INCONS_DB


    (kill any Ingres processes if necessary to ensure it is down)

  2. Run ingstart to start ingres. If ingres comes up successfully go to step 3.

    If Ingres still complains of problems with the Transaction log or the recovery server:

    shut down ingres, erase the transaction log via "cbf.exe" and restart ingres.

  3. The log shows database mdb is inconsistent. Force it consistent with command: verifydb -mrun -sdbname mdb -oforce_consistent

  4. Run "rollforwarddb mdb -edd-mmm-yyyy:hh:mm:ss"

    where the -e parameter has a date and time just prior to the first occurrence of the problem.

  5. Stop the ingres using 'ingstop -service' and run: ingsetenv II_DMFRCP_STOP_ON_INCONS_DB Yes