Block size

Start of change Choose the file system block size based on the projected workload of the file system and the type of storage that it uses. End of change

A block is the largest contiguous amount of disk space that can be allocated to a file and is therefore the largest amount of data that can be accessed in a single I/O operation. A subblock is the smallest unit of contiguous disk space that can be allocated. The size of a subblock and the number of subblocks in a block are determined by the block size. IBM Spectrum Scale™ supports the following file system block sizes: 64 KiB, 128 KiB, 256 KiB, 512 KiB, 1 MiB, 2 MiB, 4 MiB, 8 MiB and 16 MiB (for IBM Spectrum Scale RAID only). You can choose the file system block size based on the workload and the type of storage. For more information, see mmcrfs command.

Note:

A fragment is a contiguous group of one or more subblocks. Files smaller than one block size are stored in fragments. Large files are stored in a number of full blocks plus zero or more subblocks to hold the data at the end of the file. End of change

See the following tables for the suggested file system block sizes.

Table 1. Suggested file system block size per I/O type with examples of applications
I/O Type	Examples of applications	Block size
Large sequential I/O	Scientific computing, Digital media, File-based analytics	1 -16 MiB
Relational database	DB2®, Oracle, Small files on ESS	512 KiB
Small sequential I/O	General File Service, Email, Web applications	256 KiB

Table 2. Suggested file system block size per workload and configuration type.
Workload	Configuration type	Block size
SAP HANA	ESS GL	1 MiB for metadata 16 MiB for data
SAP HANA	FPO	1 MiB for single pool 256 KiB for metadata only 2 MiB for data only
Hadoop	ESS GL	1 MiB for metadata only 8 MiB for data only
Hadoop	FPO	256 KiB for metadata only 2 MiB for data only
Spark	FPO	256 KiB for metadata only 2 MiB for data only
SAP Sybase IQ	ESS GL	256 KiB - 1 MiB for metadata only 16 MiB for data only
Healthcare (Medical Imaging)	ESS GL	256 KiB for metadata only 1 MiB for data only
Healthcare (Medical Imaging)	Other Storage - IBM Spectrum Scale deployed with Block Storage at the backend. See the tip below.	256 KiB for metadata and data
Archive	Other Storage - IBM Spectrum Scale deployed with Block Storage at the backend. See the tip below.	Depends on the filesize.
ECM	Other Storage - IBM Spectrum Scale deployed with Block Storage at the backend. See the tip below.	256 KiB Unless the content is very large files (For example - Videos).
Oracle	Other Storage - IBM Spectrum Scale deployed with Block Storage at the backend. See the tip below.	256 KiB for metadata and data
Technical Computing	ESS GL	1 MiB metadata 4 - 16 MiB data depending on importance of peak sequential performance.
SAS	ESS GL	1 MiB metadata 8 - 16 MiB depending on the SASBUF size (128 KiB or 256 KiB)
Enterprise File (Misc Projects, data sharing)	Other Storage - IBM Spectrum Scale deployed with Block Storage at the backend. See the tip below.	256 KiB metadata and data

Note: To determine the ideal file system block size for your workload, it is best to run performance tests on IBM Spectrum Scale using your workload, or a representative file system benchmark with different file system block sizes

For more information about IBM Spectrum Scale RAID, see IBM Spectrum Scale RAID: Administration in the Elastic Storage Server documentation on IBM® Knowledge Center.

The --metadata-block-size option on the mmcrfs command allows a different block size to be specified for the system storage pool, provided its usage is set to metadataOnly. This can be beneficial if the file system block size is larger than 1 MB. Valid values are the same as those listed for the -B option.

Tip: If you plan to use RAID 5 or 6 devices in your file system, a larger block size might be more effective by reducing the penalties involved in small block-write operations. For example, in a RAID configuration using 4-8 data disks and 2 parity disk (a 84+2P - RAID 6 configuration), which uses a 64 KB stripe size, the optimal file system block size would be an integral multiple of 512 KB (8 data disks × 64 KB stripe size = 512 KB). A block size of an integral multiple of 256 KB results in a single data write that encompasses the 8 data disks and a parity-write to the parity disk. If a block size smaller than 512 KB, such as 256 KB, is used with the same RAID configuration, write performance might be degraded by the read-modify-write behavior depending on the Storage hardware.

The choice of block size affects the performance of certain metadata operations, in particular, block allocation performance. The IBM Spectrum Scale block allocation map is stored in blocks, similar to regular files. When the block size is small:

More blocks are required to store a given amount of data resulting in additional work to allocate those blocks
One block of allocation map data contains less information

Note: The choice of block size is particularly important for large file systems. For file systems larger than 100 TB, you should use a block size of at least 256 KB.