A Hitachi Unified Storage system is connected to a Solaris server. If you want to optimize performance for the host I/Os, how do you determine the appropriate queue depth value?

Question

Options

• ASet queue depth to 1.
• BUse the operating system default value.
• CThe value is set according to the results of the initial SCSI inquiry.
• DUsing the Host installation guide for Fibre Channel.

• A
  3% (1)
• C
  7% (2)
• D
  90% (27)

Explanation

Using the Host Installation Guide for Fibre Channel (D) is correct because Hitachi provides vendor-specific, tested queue depth recommendations for each supported host OS and HBA combination - these values account for the storage array's port limits, host HBA capabilities, and workload characteristics, ensuring optimal throughput without overwhelming the I/O queues.

Why the distractors fail:

• A is wrong - a queue depth of 1 serializes I/O, eliminating concurrency and crippling performance on any modern storage array.
• B is wrong - OS defaults are generic and conservative; they aren't tuned for enterprise SAN storage like Hitachi Unified Storage, often leaving significant performance on the table.
• C is wrong - the SCSI inquiry process identifies device type and capabilities (vendor ID, LUN parameters), but it does not determine or recommend an optimal queue depth value.

Memory tip: Think of it as "when in doubt, RTFM - but specifically the vendor's FM." Hitachi's Host Installation Guide is the authoritative source because queue depth is array-specific, not a universal OS or protocol constant. If you see a question about tuning for a specific vendor's storage system, the vendor's own documentation beats OS defaults or generic protocol behavior every time.

No community discussion yet for this question.