Installation - HP HPE MR Gen11 User Manual

Table of Contents

transaction. This mode may result in slower performance as it does not use the controller cache.

Write Back - In this mode, the controller sends a data transfer completion signal to the host when the controller cache has received all the data in a

transaction. This option provides a good balance between data protection and performance as the controller switches between write back and write through

depending on the controller status.

Always Write Back - In this mode, the controller sends a data transfer completion signal to the host when the controller cache receives all the data in a

transaction. If you select the Always Write Back policy and the energy pack is absent, the firmware is forced to use the Write Back policy.

This controller now supports two power options to enable the controller cache features. Collectively, these two options are known as the energy pack and include

the following,

HPE Smart Storage Battery

HPE Smart Storage Hybrid Capacitor

Since the interface does not display the cache policy used during the creation of the logical drive, you can identify the policy by:

Viewing the energy pack charge status in iLO and the write cache status.

Viewing cache policy information in the event log.

Temporarily disabling the cache, which displays the original cache setting.

I/O policy

The I/O policy applies to reads on a specific logical drive. It does not affect the read ahead cache. Direct IO option is available for logical drive.

In direct I/O mode, reads are not buffered in cache memory. Data is transferred to the cache and the host concurrently. If the same data block is read again, it

comes from cache memory. This option is the default setting.

Drive caching

The drive caching options for the logical drive are:

Unchanged - Leave the current drive cache policy as is.

Enabled - Enable the drive cache.

Disabled - Disable the drive cache.

Strip size selection

When a logical drive is created, the unit of data that is stored on each drive is defined as a "strip" (ranging in size from 64 KiB to 1 MiB). These strips are

distributed across the physical drives in the array.

Best performance is obtained by aligning and sizing the strip size to the application I/O request size. The smaller (<= 64 KiB) the strip size, the longer the

background parity scans and rebuilds take and the more impact to the host I/O during these operations. However, multiple stripes can be checked or rebuilt

concurrently.