ZFS pools and health monitoring

ZFS pools are typically configured to detect and mitigate the failure modes of traditional storage media. The IO Accelerator device attempts to fail as gracefully as possible, and these new failure mechanisms are compatible with ZFS pools. A device in a zpool fails to receive data at a sufficient rate if the device is in a write-reduced state and it is participating in a write-heavy workload. In this case, the drive is evicted from the zpool. A drive in read-only mode is evicted when write I/Os are returned from the device as failed. Catastrophic failures are detected and handled just as though they are on traditional storage devices.

Using module parameters

The following table describes the module parameters you can set using the esxcfg-modulecommand. The remote option (--server) is only required for the vCLI.

Sample command:

esxcfg-module --server <server-name> iomemory-vsl -s '<parameter>=<value>'

You must reboot the ESX or ESXi system to enforce any parameter changes.

Module parameter	Default (min/max)	Description

auto_attach	1	1	= Always attach the device on startup.
		0	= Do not attach the device on startup.
external_power_override*	No devices selected	Allows selected devices to draw full power from
		the PCIe slot. The <value> for this parameter is
		a comma-separated list of adapter serial
		numbers.

force_minimal_mode	0	1	= Force minimal mode on the device.
		0	= Do not force minimal mode on the device.
parallel_attach	True	1	= Enable parallel attach of multiple drivers.
		0	= Disable parallel attach of multiple drives.
tintr_hw_wait	0 (0, 255)	Interval (microseconds) to wait between
		hardware interrupts. Also known as interrupt
		coalescing.
		0	= Off.

*Use this parameter with care. For more information, see "Enabling PCIe power override (on page 14)."

Except for the external_power_override parameter, module parameters apply to all IO Accelerator devices in the system.

To query the current module parameters, use the following command:

esxcfg-module --server <server-name> iomemory-vsl -g

Monitoring IO Accelerator health 35