What will you recommend to alleviate the issue?

Examine the initialization parameters set for a database instance:

The database supports an OLTP workload. Applications connect to the instance using
shared server connections and perform small, random I/Os. All the data files are on the same
disk. You notice free buffer wait events for sessions in the database instance. To solve the
problem, you increase the size of the buffer cache. But after some time, you notice sessions
waiting again on free buffer waits. What will you recommend to alleviate the issue?
(Choose the best answer.)

Examine the initialization parameters set for a database instance:

The database supports an OLTP workload. Applications connect to the instance using
shared server connections and perform small, random I/Os. All the data files are on the same
disk. You notice free buffer wait events for sessions in the database instance. To solve the
problem, you increase the size of the buffer cache. But after some time, you notice sessions
waiting again on free buffer waits. What will you recommend to alleviate the issue?
(Choose the best answer.)

A.
Run the I/O calibration tool.

B.
Configure the database instance to make asynchronous I/O available to DBWR.

C.
Spread the data files over multiple disks, controllers, and I/O buses to ensure that there
are no hotspots in the I/O subsystem.

D.
Configure dedicated server connections for the applications.



Leave a Reply 4

Your email address will not be published. Required fields are marked *


diablo

diablo

Maybe C.

Aynch is enabled and filesystemio_processes is SO dependent.Btw, formally ASYNCH is enabled just now. Could be useful to increase DB_Writer…but we are using 1 disk and we haven’t details on CPU availables.

vasya

vasya

C looks good, but I suspect the question is about setting the value ASYNC for FILESYSTEMIO_OPTIONS

Oracool

Oracool

B – This is hinted in Oracle training doc – Multiple Writers and I/O Slaves section.

C is close but the question stated that small and random I/O. Single disk rarely an issue for small I/O.

So, along with the exhibit, I think the answer is B and as Vasya suggested, the question is geared towards async.