Sample Questions of OSS-DB Exam Gold "Performance Tuning"

[Troubleshooting - Recovering damaged cluster]

4.04

The following error message was output when referencing a certain user table.
ERROR: invalid page header in block 0 of relation base/16408/16421

Choose the most appropriate response to this situation.

  1. Perform a VACUUM on the entire database.

  2. Delete the relevant table files and restart PostgreSQL.

  3. Set zero_damaged_pages to on and attempt to access the table again.

  4. Execute CLUSTER on the relevant system table.

  5. Start PostgreSQL in single-user mode and execute REINDEX INDEX on any indexes defined on the relevant table.


[Troubleshooting - Troubleshooting common failures]

4.03

When connecting to PostgreSQL as a superuser, the following message was output.
FATAL: sorry, too many clients already

Choose the one accurate reason for this error message.

  1. The number of simultaneous connections exceeded the value set by the max_connections parameter.

  2. The number of concurrent connections exceeded the value set by the superuser_reserved_connections parameter.

  3. The number of concurrent connections exceeded the value set by the db_connections parameter.

  4. The number of concurrent connections exceeded the value of (max_connections - superuser_reserved_connections).

  5. The number of concurrent connections exceeded the value of (db_connections - superuser_reserved_connections).


[Troubleshooting - Troubleshooting common failures]

4.02

Select the one accurate explanation for the following server log.
LOG: server process (PID 21334) was terminated by signal 11: Segmentation fault
DETAIL: Failed process was running: SELECT user_func();

  1. The query was canceled due to a user request.

  2. The pg_cancel_backend function was executed against the process.

  3. The pg_terminate_backend function was executed against the process.

  4. A SIGKILL signal was sent by the OOM Killer.

  5. A SIGSEGV signal was raised by the user_func user-defined function.


[Troubleshooting - Troubleshooting common failures]

4.01

Which of the following is the correct measure to take when a server goes down due to lack of shared memory?

  1. Set an appropriate memory allocation with maintenance_work_mem.

  2. Set an appropriate memory allocation with shared_buffers.

  3. Set an appropriate memory allocation with autovacuum_work_mem.

  4. Restore the WAL file to a consistent state using pg_resetwal and start PostgreSQL.

  5. Set an appropriate memory allocation with work_mem.