ZFS Tunables for Lustre Metadata Servers (MDS): Difference between revisions

From Lustre Wiki
Jump to navigation Jump to search
(Created page with " {| class="wikitable" style="text-align:left;" !Parameter !Notes !Default !Suggested |- |zfs_arc_max |Maximum size of the ARC. |50% RAM |75% RAM |- |zfs_arc_meta_limit |Increa...")
 
(Remove some older options, update.)
 
(3 intermediate revisions by 2 users not shown)
Line 7: Line 7:
|-
|-
|zfs_arc_max
|zfs_arc_max
|Maximum size of the ARC.
|Maximum size of the ARC in bytes.
|50% RAM
|0 (~50% RAM)
|75% RAM
|75% RAM
|-
|-
|zfs_arc_meta_limit
|zfs_arc_meta_balance
|Increasing this value will improve performance if the workload involves operations on a large number of files and directories, or frequent metadata operations, at the cost of less file data fitting in the ARC.
|Increasing this value can improve performance if the workload is metadata-op heavy, at the cost of less file data fitting in the ARC. The default of 500 would prefer metadata 5x over data.
|25% ARC
|500
|75% ARC
|500+
|-
|zfs_vdev_scheduler
|VDEV scheduler
|noop
|Deadline
|-
|-
|metaslab_debug_unload
|metaslab_debug_unload
|This option prevents ZFS from unloading the spacemaps from a metaslab once it is read in.  
|This option prevents ZFS from unloading the spacemaps from a metaslab once it is read in. Unless you've noticed significant overhead from loading in spacemaps, you probably don't need this.  
|0
|0
|1
|1
|-
|-
|spl_kmem_cache_slab_limit
|Objects of spl_kmem_cache_slab_limit or smaller will be allocated using the Linux slab allocator, large objects use the SPL allocator. A cut-off of 16K  was determined to be optimal for architectures using 4K pages.
|16384
|16384
|}
|}
For a comprehensive description of all available ZFS module parameters, refer to the [https://openzfs.github.io/openzfs-docs/Performance%20and%20Tuning/Module%20Parameters.html module parameters] docs.


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

Latest revision as of 12:00, 19 October 2026

Parameter Notes Default Suggested
zfs_arc_max Maximum size of the ARC in bytes. 0 (~50% RAM) 75% RAM
zfs_arc_meta_balance Increasing this value can improve performance if the workload is metadata-op heavy, at the cost of less file data fitting in the ARC. The default of 500 would prefer metadata 5x over data. 500 500+
metaslab_debug_unload This option prevents ZFS from unloading the spacemaps from a metaslab once it is read in. Unless you've noticed significant overhead from loading in spacemaps, you probably don't need this. 0 1


For a comprehensive description of all available ZFS module parameters, refer to the module parameters docs.

Retrieved from "http://wiki.lustre.org/index.php?title=ZFS_Tunables_for_Lustre_Metadata_Servers_(MDS)&oldid=5788"