IO Scheduler
Introduction
The Linux kernel, the core of the operating system, is responsible for controlling disk access by using kernel IO scheduling.
This article explains how-to change the IO scheduler without recompiling the kernel and without restart.
The IO scheduler can be set per block device. In the examples below we will use /dev/sda.
Check the currently used IO scheduler
cat /sys/block/sda/queue/scheduler
[none] mq-deadline bfq
For example the scheduler mq-deadline delivers best performance on hardware raid and SAN environments. While none delivers better performance for SSDs.
If you don't see bfq as an option, it means that the kernel module is not loaded. It is still possible to configure the BFQ scheduler however, which will automatically load the kernel module. To load the module manually, use:
modprobe bfq
Switching IO Schedulers on runtime
Set the scheduler for /dev/sda to Multi-Queue Deadline:
echo mq-deadline > /sys/block/sda/queue/scheduler
Set the scheduler for /dev/sda to BFQ:
echo bfq > /sys/block/sda/queue/scheduler
Set IO Schedulers permanently
This can be done via udev rules. For a basic example, the following is a rule to set the BFQ scheduler for all SCSI/SAS/SATA attached disks. Create /etc/udev/rules.d/60-io-scheduler.rules with content:
ACTION=="add|change", KERNEL=="sd[a-z]*", ATTR{queue/scheduler}="bfq"