Testing IOPS on DDB Disk and Disk Library

Question

I have a few questions regarding the documentation for testing IOPS for both the DDB disk (https://documentation.commvault.com/2024e/commcell-console/testing_iops_of_deduplication_database_disk_on_windows.html) and the Disk Library (https://documentation.commvault.com/2024e/commcell-console/testing_iops_for_disk_library_mount_path_with_iometer.html)The documentation states to run the IOMeter test “on the complete raw disk”, and “IOMeter allows allows physical disks to be selected only when there are no volumes on the disk”. Does this mean that the disk should not be formatted before running the IOMeter test?	For the Disk Library, I am mounting a LUN from a SAN to the MediaAgent that is about 48 TB. What is the best way to test IOPS for this? I am guessing I will need to set the Maximum Disk Size, otherwise creating that huge iobw.tst file would take ages.	Is the CVDiskPerf tool the preferred method testing Disk Library (like the above)? https://documentation.commvault.com/2024e/commcell-console/disk_performance_tool.html

Erase4ndReuseMedia · Answer

For the DDB Volume, it is dependent on the level of accuracy you’re after.

If I am handed a MediaAgent with already formatted disks, I will generally just run IOMeter on the formatted disks. There may be performance implications in doing so and there can be significant preparation time to write the test file, but it is probably closer to real-world performance levels, which may be more valuable than knowing the outright performance capabilities of the disk.

Typically, I wouldn’t test IOPS for library mount paths unless specifically requested to, but I do recommend testing with CVDiskPerf.

Testing IOPS on DDB Disk and Disk Library

1 reply

Readiverse Academy Certifications

Readiverse Academy Certifications

Sign up

Login to the community