Linux filesystem tuning

ext2-3-4 tuning

XFS is currently the recommended filesystem. It can achieve up to three times the performance of a tuned ext2/ext3 solution. At this time, there is no known additional tuning for running XFS in a single- or multi-IO Accelerator configuration.

Setting stride size and stripe width for ext2/3 (extN) when using

RAID

The extN filesystem family has create-time options of stride and stripe width. Stride helps the filesystem ensure that critical metadata structures are spread across the disks in the RAID evenly, to keep any given disk from becoming a hotspot. Stripe width allows for filesystem optimization.

To calculate the correct values for stride size, take the chunk size of the RAID array and divide it by the block size of the filesystem. For example:

stride = (chunk size / filesystem block size)

Stripe width calculation requires the stride size and the number of data-bearing disks. The following table shows the calculation for number of data bearing disks per raid type.

dbd = # of data bearing disks total_disks = # of active disks

mirrored_sets = The number of raid 1 mirrored sets that are used to form the higher-level group

RAID level		Data bearing disks (dbd)

0	(Striping)	total_disks
1	(Mirroring)	1
5		total_disks - 1
6		total_disks - 2
10		mirrored_sets
50		mirrored_sets - 1

To calculate the stripe width, take the number of data bearing disks and multiply it by the stride.

stripe_width = dbd * stride

Sample Configuration

NOTE: The example below is not a recommended configuration, but it is a good demonstration of how to use the above equations.

Create a RAID50 with 10 IO Accelerators combined into five mirrored sets, (two mirrored IO Accelerators per set) and 256KB chunk size.

Create an ext3 filesystem with an 8K block size:

Linux filesystem tuning 21