ZFS recordsize Property - Revision history

Snehring: merge into ZFS, pending deletion

2026-06-05T19:22:04Z

merge into ZFS, pending deletion

← Older revision		Revision as of 12:22, 5 June 2026
Line 1:		Line 1:
	The <code>recordsize</code> property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed, but for workloads that create very large files, increasing the value of <code>recordsize</code> can deliver a performance benefit. The chosen size must be a power of 2 with the minimum allowed size being 512 bytes. For ZFS on Linux, the maximum value is 1MiB in version 0.6.5. Prior to this release, the maximum value was 128K (which was also the default setting for Lustre 2.10 and earlier).		see [[ZFS#recordsize]]

	The default setting is not sufficient to sustain the performance of throughput oriented workloads that are typical of the I/O patterns for OSTs, and it is recommended to increase the <code>recordsize</code> to 1MiB (1024K), in order to better match the Lustre IO transaction sizes for block I/O. This is the default for ZFS OSTs with Lustre 2.11 and later.

	[[~~Category:~~ZFS ~~OSD Tuning]]~~
	~~[[Category:Lustre Systems Administration~~]]

Adilger: Note that default recordsize is 1MB with 2.11 and later

2018-05-25T20:28:00Z

Note that default recordsize is 1MB with 2.11 and later

← Older revision		Revision as of 13:28, 25 May 2018
Line 1:		Line 1:
	The <code>recordsize</code> property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed, but for workloads that create very large files, increasing the value of <code>recordsize</code> can deliver a performance benefit. The chosen size must be a power of 2 with the minimum allowed size being 512 bytes. For ZFS on Linux, the maximum value is 1MiB in version 0.6.5. Prior to this release, the maximum value was 128K (which is also the default setting).		The <code>recordsize</code> property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed, but for workloads that create very large files, increasing the value of <code>recordsize</code> can deliver a performance benefit. The chosen size must be a power of 2 with the minimum allowed size being 512 bytes. For ZFS on Linux, the maximum value is 1MiB in version 0.6.5. Prior to this release, the maximum value was 128K (which was also the default setting for Lustre 2.10 and earlier).

	The default setting is not sufficient to sustain the performance of ~~the~~ throughput oriented workloads that are typical of the I/O patterns for OSTs, and it is recommended to increase the <code>recordsize</code> to 1MiB (1024K), in order to better match the Lustre IO transaction sizes for block I/O.		The default setting is not sufficient to sustain the performance of throughput oriented workloads that are typical of the I/O patterns for OSTs, and it is recommended to increase the <code>recordsize</code> to 1MiB (1024K), in order to better match the Lustre IO transaction sizes for block I/O. This is the default for ZFS OSTs with Lustre 2.11 and later.

	[[Category:ZFS OSD Tuning]]		[[Category:ZFS OSD Tuning]]
	[[Category:Lustre Systems Administration]]		[[Category:Lustre Systems Administration]]

Malcolm at 05:53, 31 August 2017

2017-08-31T05:53:14Z

← Older revision		Revision as of 22:53, 30 August 2017
Line 4:		Line 4:

	[[Category:ZFS OSD Tuning]]		[[Category:ZFS OSD Tuning]]
			[[Category:Lustre Systems Administration]]

Malcolm: Created page with "The `recordsize` property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed,..."

2017-08-14T08:27:37Z

Created page with "The <code>recordsize</code> property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed,..."

New page

The <code>recordsize</code> property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed, but for workloads that create very large files, increasing the value of <code>recordsize</code> can deliver a performance benefit. The chosen size must be a power of 2 with the minimum allowed size being 512 bytes. For ZFS on Linux, the maximum value is 1MiB in version 0.6.5. Prior to this release, the maximum value was 128K (which is also the default setting).

The default setting is not sufficient to sustain the performance of the throughput oriented workloads that are typical of the I/O patterns for OSTs, and it is recommended to increase the <code>recordsize</code> to 1MiB (1024K), in order to better match the Lustre IO transaction sizes for block I/O.

[[Category:ZFS OSD Tuning]]

ZFS recordsize Property - Revision history

Snehring: merge into ZFS, pending deletion

Adilger: Note that default recordsize is 1MB with 2.11 and later

Malcolm at 05:53, 31 August 2017

Malcolm: Created page with "The recordsize property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed,..."

Malcolm: Created page with "The `recordsize` property of ZFS datasets is used to specify the maximum block size for files in the file system. Often, this property does not need to be changed,..."