1.\" SPDX-License-Identifier: CDDL-1.0 2.\" 3.\" CDDL HEADER START 4.\" 5.\" The contents of this file are subject to the terms of the 6.\" Common Development and Distribution License (the "License"). 7.\" You may not use this file except in compliance with the License. 8.\" 9.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 10.\" or https://opensource.org/licenses/CDDL-1.0. 11.\" See the License for the specific language governing permissions 12.\" and limitations under the License. 13.\" 14.\" When distributing Covered Code, include this CDDL HEADER in each 15.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE. 16.\" If applicable, add the following below this CDDL HEADER, with the 17.\" fields enclosed by brackets "[]" replaced with your own identifying 18.\" information: Portions Copyright [yyyy] [name of copyright owner] 19.\" 20.\" CDDL HEADER END 21.\" 22.\" Copyright (c) 2009 Sun Microsystems, Inc. All Rights Reserved. 23.\" Copyright 2011 Joshua M. Clulow <josh@sysmgr.org> 24.\" Copyright (c) 2011, 2019 by Delphix. All rights reserved. 25.\" Copyright (c) 2011, Pawel Jakub Dawidek <pjd@FreeBSD.org> 26.\" Copyright (c) 2012, Glen Barber <gjb@FreeBSD.org> 27.\" Copyright (c) 2012, Bryan Drewery <bdrewery@FreeBSD.org> 28.\" Copyright (c) 2013, Steven Hartland <smh@FreeBSD.org> 29.\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved. 30.\" Copyright (c) 2014, Joyent, Inc. All rights reserved. 31.\" Copyright (c) 2014 by Adam Stevko. All rights reserved. 32.\" Copyright (c) 2014 Integros [integros.com] 33.\" Copyright (c) 2016 Nexenta Systems, Inc. All Rights Reserved. 34.\" Copyright (c) 2014, Xin LI <delphij@FreeBSD.org> 35.\" Copyright (c) 2014-2015, The FreeBSD Foundation, All Rights Reserved. 36.\" Copyright 2019 Richard Laager. All rights reserved. 37.\" Copyright 2018 Nexenta Systems, Inc. 38.\" Copyright 2019 Joyent, Inc. 39.\" Copyright (c) 2019, Kjeld Schouten-Lebbing 40.\" Copyright (c) 2022 Hewlett Packard Enterprise Development LP. 41.\" 42.Dd June 29, 2024 43.Dt ZFSPROPS 7 44.Os 45. 46.Sh NAME 47.Nm zfsprops 48.Nd native and user-defined properties of ZFS datasets 49. 50.Sh DESCRIPTION 51Properties are divided into two types, native properties and user-defined 52.Po or 53.Qq user 54.Pc 55properties. 56Native properties either export internal statistics or control ZFS behavior. 57In addition, native properties are either editable or read-only. 58User properties have no effect on ZFS behavior, but you can use them to annotate 59datasets in a way that is meaningful in your environment. 60For more information about user properties, see the 61.Sx User Properties 62section, below. 63. 64.Ss Native Properties 65Every dataset has a set of properties that export statistics about the dataset 66as well as control various behaviors. 67Properties are inherited from the parent unless overridden by the child. 68Some properties apply only to certain types of datasets 69.Pq file systems, volumes, or snapshots . 70.Pp 71The values of numeric properties can be specified using human-readable suffixes 72.Po for example, 73.Sy k , 74.Sy KB , 75.Sy M , 76.Sy Gb , 77and so forth, up to 78.Sy Z 79for zettabyte 80.Pc . 81The following are all valid 82.Pq and equal 83specifications: 84.Li 1536M , 85.Li 1.5g , 86.Li 1.50GB . 87.Pp 88The values of non-numeric properties are case sensitive and must be lowercase, 89except for 90.Sy mountpoint , 91.Sy sharenfs , 92and 93.Sy sharesmb . 94.Pp 95The following native properties consist of read-only statistics about the 96dataset. 97These properties can be neither set, nor inherited. 98Native properties apply to all dataset types unless otherwise noted. 99.Bl -tag -width "usedbyrefreservation" 100.It Sy available 101The amount of space available to the dataset and all its children, assuming that 102there is no other activity in the pool. 103Because space is shared within a pool, availability can be limited by any number 104of factors, including physical pool size, quotas, reservations, or other 105datasets within the pool. 106.Pp 107This property can also be referred to by its shortened column name, 108.Sy avail . 109.It Sy compressratio 110For non-snapshots, the compression ratio achieved for the 111.Sy used 112space of this dataset, expressed as a multiplier. 113The 114.Sy used 115property includes descendant datasets, and, for clones, does not include the 116space shared with the origin snapshot. 117For snapshots, the 118.Sy compressratio 119is the same as the 120.Sy refcompressratio 121property. 122Compression can be turned on by running: 123.Nm zfs Cm set Sy compression Ns = Ns Sy on Ar dataset . 124The default value is 125.Sy off . 126.It Sy createtxg 127The transaction group (txg) in which the dataset was created. 128Bookmarks have the same 129.Sy createtxg 130as the snapshot they are initially tied to. 131This property is suitable for ordering a list of snapshots, 132e.g. for incremental send and receive. 133.It Sy creation 134The time this dataset was created. 135.It Sy clones 136For snapshots, this property is a comma-separated list of filesystems or volumes 137which are clones of this snapshot. 138The clones' 139.Sy origin 140property is this snapshot. 141If the 142.Sy clones 143property is not empty, then this snapshot can not be destroyed 144.Po even with the 145.Fl r 146or 147.Fl f 148options 149.Pc . 150The roles of origin and clone can be swapped by promoting the clone with the 151.Nm zfs Cm promote 152command. 153.It Sy defer_destroy 154This property is 155.Sy on 156if the snapshot has been marked for deferred destroy by using the 157.Nm zfs Cm destroy Fl d 158command. 159Otherwise, the property is 160.Sy off . 161.It Sy encryptionroot 162For encrypted datasets, indicates where the dataset is currently inheriting its 163encryption key from. 164Loading or unloading a key for the 165.Sy encryptionroot 166will implicitly load / unload the key for any inheriting datasets (see 167.Nm zfs Cm load-key 168and 169.Nm zfs Cm unload-key 170for details). 171Clones will always share an 172encryption key with their origin. 173See the 174.Sx Encryption 175section of 176.Xr zfs-load-key 8 177for details. 178.It Sy filesystem_count 179The total number of filesystems and volumes that exist under this location in 180the dataset tree. 181This value is only available when a 182.Sy filesystem_limit 183has been set somewhere in the tree under which the dataset resides. 184.It Sy keystatus 185Indicates if an encryption key is currently loaded into ZFS. 186The possible values are 187.Sy none , 188.Sy available , 189and 190.Sy unavailable . 191See 192.Nm zfs Cm load-key 193and 194.Nm zfs Cm unload-key . 195.It Sy guid 196The 64 bit GUID of this dataset or bookmark which does not change over its 197entire lifetime. 198When a snapshot is sent to another pool, the received snapshot has the same 199GUID. 200Thus, the 201.Sy guid 202is suitable to identify a snapshot across pools. 203.It Sy logicalreferenced 204The amount of space that is 205.Qq logically 206accessible by this dataset. 207See the 208.Sy referenced 209property. 210The logical space ignores the effect of the 211.Sy compression 212and 213.Sy copies 214properties, giving a quantity closer to the amount of data that applications 215see. 216However, it does include space consumed by metadata. 217.Pp 218This property can also be referred to by its shortened column name, 219.Sy lrefer . 220.It Sy logicalused 221The amount of space that is 222.Qq logically 223consumed by this dataset and all its descendants. 224See the 225.Sy used 226property. 227The logical space ignores the effect of the 228.Sy compression 229and 230.Sy copies 231properties, giving a quantity closer to the amount of data that applications 232see. 233However, it does include space consumed by metadata. 234.Pp 235This property can also be referred to by its shortened column name, 236.Sy lused . 237.It Sy mounted 238For file systems, indicates whether the file system is currently mounted. 239This property can be either 240.Sy yes 241or 242.Sy no . 243.It Sy objsetid 244A unique identifier for this dataset within the pool. 245Unlike the dataset's 246.Sy guid , No the Sy objsetid 247of a dataset is not transferred to other pools when the snapshot is copied 248with a send/receive operation. 249The 250.Sy objsetid 251can be reused (for a new dataset) after the dataset is deleted. 252.It Sy origin 253For cloned file systems or volumes, the snapshot from which the clone was 254created. 255See also the 256.Sy clones 257property. 258.It Sy receive_resume_token 259For filesystems or volumes which have saved partially-completed state from 260.Nm zfs Cm receive Fl s , 261this opaque token can be provided to 262.Nm zfs Cm send Fl t 263to resume and complete the 264.Nm zfs Cm receive . 265.It Sy redact_snaps 266For bookmarks, this is the list of snapshot GUIDs the bookmark contains a 267redaction 268list for. 269For snapshots, this is the list of snapshot GUIDs the snapshot is redacted with 270respect to. 271.It Sy referenced 272The amount of data that is accessible by this dataset, which may or may not be 273shared with other datasets in the pool. 274When a snapshot or clone is created, it initially references the same amount of 275space as the file system or snapshot it was created from, since its contents are 276identical. 277.Pp 278This property can also be referred to by its shortened column name, 279.Sy refer . 280.It Sy refcompressratio 281The compression ratio achieved for the 282.Sy referenced 283space of this dataset, expressed as a multiplier. 284See also the 285.Sy compressratio 286property. 287.It Sy snapshot_count 288The total number of snapshots that exist under this location in the dataset 289tree. 290This value is only available when a 291.Sy snapshot_limit 292has been set somewhere in the tree under which the dataset resides. 293.It Sy type 294The type of dataset: 295.Sy filesystem , 296.Sy volume , 297.Sy snapshot , 298or 299.Sy bookmark . 300.It Sy used 301The amount of space consumed by this dataset and all its descendants. 302This is the value that is checked against this dataset's quota and reservation. 303The space used does not include this dataset's reservation, but does take into 304account the reservations of any descendent datasets. 305The amount of space that a dataset consumes from its parent, as well as the 306amount of space that is freed if this dataset is recursively destroyed, is the 307greater of its space used and its reservation. 308.Pp 309The used space of a snapshot 310.Po see the 311.Sx Snapshots 312section of 313.Xr zfsconcepts 7 314.Pc 315is space that is referenced exclusively by this snapshot. 316If this snapshot is destroyed, the amount of 317.Sy used 318space will be freed. 319Space that is shared by multiple snapshots isn't accounted for in this metric. 320When a snapshot is destroyed, space that was previously shared with this 321snapshot can become unique to snapshots adjacent to it, thus changing the used 322space of those snapshots. 323The used space of the latest snapshot can also be affected by changes in the 324file system. 325Note that the 326.Sy used 327space of a snapshot is a subset of the 328.Sy written 329space of the snapshot. 330.Pp 331The amount of space used, available, or referenced does not take into account 332pending changes. 333Pending changes are generally accounted for within a few seconds. 334Committing a change to a disk using 335.Xr fsync 2 336or 337.Sy O_SYNC 338does not necessarily guarantee that the space usage information is updated 339immediately. 340.It Sy usedby* 341The 342.Sy usedby* 343properties decompose the 344.Sy used 345properties into the various reasons that space is used. 346Specifically, 347.Sy used No = 348.Sy usedbychildren No + 349.Sy usedbydataset No + 350.Sy usedbyrefreservation No + 351.Sy usedbysnapshots . 352These properties are only available for datasets created on 353.Nm zpool 354.Qo version 13 Qc 355pools. 356.It Sy usedbychildren 357The amount of space used by children of this dataset, which would be freed if 358all the dataset's children were destroyed. 359.It Sy usedbydataset 360The amount of space used by this dataset itself, which would be freed if the 361dataset were destroyed 362.Po after first removing any 363.Sy refreservation 364and destroying any necessary snapshots or descendants 365.Pc . 366.It Sy usedbyrefreservation 367The amount of space used by a 368.Sy refreservation 369set on this dataset, which would be freed if the 370.Sy refreservation 371was removed. 372.It Sy usedbysnapshots 373The amount of space consumed by snapshots of this dataset. 374In particular, it is the amount of space that would be freed if all of this 375dataset's snapshots were destroyed. 376Note that this is not simply the sum of the snapshots' 377.Sy used 378properties because space can be shared by multiple snapshots. 379.It Sy userused Ns @ Ns Ar user 380The amount of space consumed by the specified user in this dataset. 381Space is charged to the owner of each file, as displayed by 382.Nm ls Fl l . 383The amount of space charged is displayed by 384.Nm du No and Nm ls Fl s . 385See the 386.Nm zfs Cm userspace 387command for more information. 388.Pp 389Unprivileged users can access only their own space usage. 390The root user, or a user who has been granted the 391.Sy userused 392privilege with 393.Nm zfs Cm allow , 394can access everyone's usage. 395.Pp 396The 397.Sy userused Ns @ Ns Ar … 398properties are not displayed by 399.Nm zfs Cm get Sy all . 400The user's name must be appended after the 401.Sy @ 402symbol, using one of the following forms: 403.Bl -bullet -compact -offset 4n 404.It 405POSIX name 406.Pq Qq joe 407.It 408POSIX numeric ID 409.Pq Qq 789 410.It 411SID name 412.Pq Qq joe.smith@mydomain 413.It 414SID numeric ID 415.Pq Qq S-1-123-456-789 416.El 417.Pp 418Files created on Linux always have POSIX owners. 419.It Sy userobjused Ns @ Ns Ar user 420The 421.Sy userobjused 422property is similar to 423.Sy userused 424but instead it counts the number of objects consumed by a user. 425This property counts all objects allocated on behalf of the user, 426it may differ from the results of system tools such as 427.Nm df Fl i . 428.Pp 429When the property 430.Sy xattr Ns = Ns Sy on 431is set on a file system additional objects will be created per-file to store 432extended attributes. 433These additional objects are reflected in the 434.Sy userobjused 435value and are counted against the user's 436.Sy userobjquota . 437When a file system is configured to use 438.Sy xattr Ns = Ns Sy sa 439no additional internal objects are normally required. 440.It Sy userrefs 441This property is set to the number of user holds on this snapshot. 442User holds are set by using the 443.Nm zfs Cm hold 444command. 445.It Sy groupused Ns @ Ns Ar group 446The amount of space consumed by the specified group in this dataset. 447Space is charged to the group of each file, as displayed by 448.Nm ls Fl l . 449See the 450.Sy userused Ns @ Ns Ar user 451property for more information. 452.Pp 453Unprivileged users can only access their own groups' space usage. 454The root user, or a user who has been granted the 455.Sy groupused 456privilege with 457.Nm zfs Cm allow , 458can access all groups' usage. 459.It Sy groupobjused Ns @ Ns Ar group 460The number of objects consumed by the specified group in this dataset. 461Multiple objects may be charged to the group for each file when extended 462attributes are in use. 463See the 464.Sy userobjused Ns @ Ns Ar user 465property for more information. 466.Pp 467Unprivileged users can only access their own groups' space usage. 468The root user, or a user who has been granted the 469.Sy groupobjused 470privilege with 471.Nm zfs Cm allow , 472can access all groups' usage. 473.It Sy projectused Ns @ Ns Ar project 474The amount of space consumed by the specified project in this dataset. 475Project is identified via the project identifier (ID) that is object-based 476numeral attribute. 477An object can inherit the project ID from its parent object (if the 478parent has the flag of inherit project ID that can be set and changed via 479.Nm chattr Fl /+P 480or 481.Nm zfs project Fl s ) 482when being created. 483The privileged user can set and change object's project 484ID via 485.Nm chattr Fl p 486or 487.Nm zfs project Fl s 488anytime. 489Space is charged to the project of each file, as displayed by 490.Nm lsattr Fl p 491or 492.Nm zfs project . 493See the 494.Sy userused Ns @ Ns Ar user 495property for more information. 496.Pp 497The root user, or a user who has been granted the 498.Sy projectused 499privilege with 500.Nm zfs allow , 501can access all projects' usage. 502.It Sy projectobjused Ns @ Ns Ar project 503The 504.Sy projectobjused 505is similar to 506.Sy projectused 507but instead it counts the number of objects consumed by project. 508When the property 509.Sy xattr Ns = Ns Sy on 510is set on a fileset, ZFS will create additional objects per-file to store 511extended attributes. 512These additional objects are reflected in the 513.Sy projectobjused 514value and are counted against the project's 515.Sy projectobjquota . 516When a filesystem is configured to use 517.Sy xattr Ns = Ns Sy sa 518no additional internal objects are required. 519See the 520.Sy userobjused Ns @ Ns Ar user 521property for more information. 522.Pp 523The root user, or a user who has been granted the 524.Sy projectobjused 525privilege with 526.Nm zfs allow , 527can access all projects' objects usage. 528.It Sy snapshots_changed 529Provides a mechanism to quickly determine whether snapshot list has 530changed without having to mount a dataset or iterate the snapshot list. 531Specifies the time at which a snapshot for a dataset was last 532created or deleted. 533.Pp 534This allows us to be more efficient how often we query snapshots. 535The property is persistent across mount and unmount operations only if the 536.Sy extensible_dataset 537feature is enabled. 538.It Sy volblocksize 539For volumes, specifies the block size of the volume. 540The 541.Sy blocksize 542cannot be changed once the volume has been written, so it should be set at 543volume creation time. 544The default 545.Sy blocksize 546for volumes is 16 KiB. 547Any power of 2 from 512 bytes to 128 KiB is valid. 548.Pp 549This property can also be referred to by its shortened column name, 550.Sy volblock . 551.It Sy written 552The amount of space 553.Sy referenced 554by this dataset, that was written since the previous snapshot 555.Pq i.e. that is not referenced by the previous snapshot . 556.It Sy written Ns @ Ns Ar snapshot 557The amount of 558.Sy referenced 559space written to this dataset since the specified snapshot. 560This is the space that is referenced by this dataset but was not referenced by 561the specified snapshot. 562.Pp 563The 564.Ar snapshot 565may be specified as a short snapshot name 566.Pq just the part after the Sy @ , 567in which case it will be interpreted as a snapshot in the same filesystem as 568this dataset. 569The 570.Ar snapshot 571may be a full snapshot name 572.Pq Ar filesystem Ns @ Ns Ar snapshot , 573which for clones may be a snapshot in the origin's filesystem 574.Pq or the origin of the origin's filesystem, etc. 575.El 576.Pp 577The following native properties can be used to change the behavior of a ZFS 578dataset. 579.Bl -tag -width "" 580.It Xo 581.Sy aclinherit Ns = Ns Sy discard Ns | Ns Sy noallow Ns | Ns 582.Sy restricted Ns | Ns Sy passthrough Ns | Ns Sy passthrough-x 583.Xc 584Controls how ACEs are inherited when files and directories are created. 585.Bl -tag -compact -offset 4n -width "passthrough-x" 586.It Sy discard 587does not inherit any ACEs. 588.It Sy noallow 589only inherits inheritable ACEs that specify 590.Qq deny 591permissions. 592.It Sy restricted 593default, removes the 594.Sy write_acl 595and 596.Sy write_owner 597permissions when the ACE is inherited. 598.It Sy passthrough 599inherits all inheritable ACEs without any modifications. 600.It Sy passthrough-x 601same meaning as 602.Sy passthrough , 603except that the 604.Sy owner@ , group@ , No and Sy everyone@ 605ACEs inherit the execute permission only if the file creation mode also requests 606the execute bit. 607.El 608.Pp 609When the property value is set to 610.Sy passthrough , 611files are created with a mode determined by the inheritable ACEs. 612If no inheritable ACEs exist that affect the mode, then the mode is set in 613accordance to the requested mode from the application. 614.Pp 615The 616.Sy aclinherit 617property does not apply to POSIX ACLs. 618.It Xo 619.Sy aclmode Ns = Ns Sy discard Ns | Ns Sy groupmask Ns | Ns 620.Sy passthrough Ns | Ns Sy restricted Ns 621.Xc 622Controls how an ACL is modified during chmod(2) and how inherited ACEs 623are modified by the file creation mode: 624.Bl -tag -compact -offset 4n -width "passthrough" 625.It Sy discard 626default, deletes all 627.Sy ACEs 628except for those representing 629the mode of the file or directory requested by 630.Xr chmod 2 . 631.It Sy groupmask 632reduces permissions granted in all 633.Sy ALLOW 634entries found in the 635.Sy ACL 636such that they are no greater than the group permissions specified by 637.Xr chmod 2 . 638.It Sy passthrough 639indicates that no changes are made to the ACL other than creating or updating 640the necessary ACL entries to represent the new mode of the file or directory. 641.It Sy restricted 642will cause the 643.Xr chmod 2 644operation to return an error when used on any file or directory which has 645a non-trivial ACL whose entries can not be represented by a mode. 646.Xr chmod 2 647is required to change the set user ID, set group ID, or sticky bits on a file 648or directory, as they do not have equivalent ACL entries. 649In order to use 650.Xr chmod 2 651on a file or directory with a non-trivial ACL when 652.Sy aclmode 653is set to 654.Sy restricted , 655you must first remove all ACL entries which do not represent the current mode. 656.El 657.It Sy acltype Ns = Ns Sy off Ns | Ns Sy nfsv4 Ns | Ns Sy posix 658Controls whether ACLs are enabled and if so what type of ACL to use. 659When this property is set to a type of ACL not supported by the current 660platform, the behavior is the same as if it were set to 661.Sy off . 662.Bl -tag -compact -offset 4n -width "posixacl" 663.It Sy off 664default on Linux, when a file system has the 665.Sy acltype 666property set to off then ACLs are disabled. 667.It Sy noacl 668an alias for 669.Sy off 670.It Sy nfsv4 671default on 672.Fx , 673indicates that NFSv4-style ZFS ACLs should be used. 674These ACLs can be managed with the 675.Xr getfacl 1 676and 677.Xr setfacl 1 . 678The 679.Sy nfsv4 680ZFS ACL type is not yet supported on Linux. 681.It Sy posix 682indicates POSIX ACLs should be used. 683POSIX ACLs are specific to Linux and are not functional on other platforms. 684POSIX ACLs are stored as an extended 685attribute and therefore will not overwrite any existing NFSv4 ACLs which 686may be set. 687.It Sy posixacl 688an alias for 689.Sy posix 690.El 691.Pp 692To obtain the best performance when setting 693.Sy posix 694users are strongly encouraged to set the 695.Sy xattr Ns = Ns Sy sa 696property. 697This will result in the POSIX ACL being stored more efficiently on disk. 698But as a consequence, all new extended attributes will only be 699accessible from OpenZFS implementations which support the 700.Sy xattr Ns = Ns Sy sa 701property. 702See the 703.Sy xattr 704property for more details. 705.It Sy atime Ns = Ns Sy on Ns | Ns Sy off 706Controls whether the access time for files is updated when they are read. 707Turning this property off avoids producing write traffic when reading files and 708can result in significant performance gains, though it might confuse mailers 709and other similar utilities. 710The values 711.Sy on 712and 713.Sy off 714are equivalent to the 715.Sy atime 716and 717.Sy noatime 718mount options. 719The default value is 720.Sy on . 721See also 722.Sy relatime 723below. 724.It Sy canmount Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy noauto 725If this property is set to 726.Sy off , 727the file system cannot be mounted, and is ignored by 728.Nm zfs Cm mount Fl a . 729Setting this property to 730.Sy off 731is similar to setting the 732.Sy mountpoint 733property to 734.Sy none , 735except that the dataset still has a normal 736.Sy mountpoint 737property, which can be inherited. 738Setting this property to 739.Sy off 740allows datasets to be used solely as a mechanism to inherit properties. 741One example of setting 742.Sy canmount Ns = Ns Sy off 743is to have two datasets with the same 744.Sy mountpoint , 745so that the children of both datasets appear in the same directory, but might 746have different inherited characteristics. 747.Pp 748When set to 749.Sy noauto , 750a dataset can only be mounted and unmounted explicitly. 751The dataset is not mounted automatically when the dataset is created or 752imported, nor is it mounted by the 753.Nm zfs Cm mount Fl a 754command or unmounted by the 755.Nm zfs Cm unmount Fl a 756command. 757.Pp 758This property is not inherited. 759.It Xo 760.Sy checksum Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy fletcher2 Ns | Ns 761.Sy fletcher4 Ns | Ns Sy sha256 Ns | Ns Sy noparity Ns | Ns 762.Sy sha512 Ns | Ns Sy skein Ns | Ns Sy edonr Ns | Ns Sy blake3 763.Xc 764Controls the checksum used to verify data integrity. 765The default value is 766.Sy on , 767which automatically selects an appropriate algorithm 768.Po currently, 769.Sy fletcher4 , 770but this may change in future releases 771.Pc . 772The value 773.Sy off 774disables integrity checking on user data. 775The value 776.Sy noparity 777not only disables integrity but also disables maintaining parity for user data. 778This setting is used internally by a dump device residing on a RAID-Z pool and 779should not be used by any other dataset. 780Disabling checksums is 781.Em NOT 782a recommended practice. 783.Pp 784The 785.Sy sha512 , 786.Sy skein , 787.Sy edonr , 788and 789.Sy blake3 790checksum algorithms require enabling the appropriate features on the pool. 791.Pp 792Please see 793.Xr zpool-features 7 794for more information on these algorithms. 795.Pp 796Changing this property affects only newly-written data. 797.It Xo 798.Sy compression Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy gzip Ns | Ns 799.Sy gzip- Ns Ar N Ns | Ns Sy lz4 Ns | Ns Sy lzjb Ns | Ns Sy zle Ns | Ns Sy zstd Ns | Ns 800.Sy zstd- Ns Ar N Ns | Ns Sy zstd-fast Ns | Ns Sy zstd-fast- Ns Ar N 801.Xc 802Controls the compression algorithm used for this dataset. 803.Pp 804When set to 805.Sy on 806(the default), indicates that the current default compression algorithm should 807be used. 808The default balances compression and decompression speed, with compression ratio 809and is expected to work well on a wide variety of workloads. 810Unlike all other settings for this property, 811.Sy on 812does not select a fixed compression type. 813As new compression algorithms are added to ZFS and enabled on a pool, the 814default compression algorithm may change. 815The current default compression algorithm is either 816.Sy lzjb 817or, if the 818.Sy lz4_compress 819feature is enabled, 820.Sy lz4 . 821.Pp 822The 823.Sy lz4 824compression algorithm is a high-performance replacement for the 825.Sy lzjb 826algorithm. 827It features significantly faster compression and decompression, as well as a 828moderately higher compression ratio than 829.Sy lzjb , 830but can only be used on pools with the 831.Sy lz4_compress 832feature set to 833.Sy enabled . 834See 835.Xr zpool-features 7 836for details on ZFS feature flags and the 837.Sy lz4_compress 838feature. 839.Pp 840The 841.Sy lzjb 842compression algorithm is optimized for performance while providing decent data 843compression. 844.Pp 845The 846.Sy gzip 847compression algorithm uses the same compression as the 848.Xr gzip 1 849command. 850You can specify the 851.Sy gzip 852level by using the value 853.Sy gzip- Ns Ar N , 854where 855.Ar N 856is an integer from 1 857.Pq fastest 858to 9 859.Pq best compression ratio . 860Currently, 861.Sy gzip 862is equivalent to 863.Sy gzip-6 864.Po which is also the default for 865.Xr gzip 1 866.Pc . 867.Pp 868The 869.Sy zstd 870compression algorithm provides both high compression ratios and good 871performance. 872You can specify the 873.Sy zstd 874level by using the value 875.Sy zstd- Ns Ar N , 876where 877.Ar N 878is an integer from 1 879.Pq fastest 880to 19 881.Pq best compression ratio . 882.Sy zstd 883is equivalent to 884.Sy zstd-3 . 885.Pp 886Faster speeds at the cost of the compression ratio can be requested by 887setting a negative 888.Sy zstd 889level. 890This is done using 891.Sy zstd-fast- Ns Ar N , 892where 893.Ar N 894is an integer in 895.Bq Sy 1 Ns - Ns Sy 10 , 20 , 30 , No … , Sy 100 , 500 , 1000 896which maps to a negative 897.Sy zstd 898level. 899The lower the level the faster the compression \(em 900.Sy 1000 901provides the fastest compression and lowest compression ratio. 902.Sy zstd-fast 903is equivalent to 904.Sy zstd-fast- Ns Ar 1 . 905.Pp 906The 907.Sy zle 908compression algorithm compresses runs of zeros. 909.Pp 910This property can also be referred to by its shortened column name 911.Sy compress . 912Changing this property affects only newly-written data. 913.Pp 914When any setting except 915.Sy off 916is selected, compression will explicitly check for blocks consisting of only 917zeroes (the NUL byte). 918When a zero-filled block is detected, it is stored as 919a hole and not compressed using the indicated compression algorithm. 920.Pp 921All blocks are allocated as a whole number of sectors 922.Pq chunks of 2^ Ns Sy ashift No bytes , e.g . Sy 512B No or Sy 4KB . 923Compression may result in a non-sector-aligned size, which will be rounded up 924to a whole number of sectors. 925If compression saves less than one whole sector, 926the block will be stored uncompressed. 927Therefore, blocks whose logical size is a small number of sectors will 928experience less compression 929(e.g. for 930.Sy recordsize Ns = Ns Sy 16K 931with 932.Sy 4K 933sectors, which have 4 sectors per block, 934compression needs to save at least 25% to actually save space on disk). 935.Pp 936There is 937.Sy 12.5% 938default compression threshold in addition to sector rounding. 939.It Xo 940.Sy context Ns = Ns Sy none Ns | Ns 941.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level 942.Xc 943This flag sets the SELinux context for all files in the file system under 944a mount point for that file system. 945See 946.Xr selinux 8 947for more information. 948.It Xo 949.Sy fscontext Ns = Ns Sy none Ns | Ns 950.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level 951.Xc 952This flag sets the SELinux context for the file system file system being 953mounted. 954See 955.Xr selinux 8 956for more information. 957.It Xo 958.Sy defcontext Ns = Ns Sy none Ns | Ns 959.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level 960.Xc 961This flag sets the SELinux default context for unlabeled files. 962See 963.Xr selinux 8 964for more information. 965.It Xo 966.Sy rootcontext Ns = Ns Sy none Ns | Ns 967.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level 968.Xc 969This flag sets the SELinux context for the root inode of the file system. 970See 971.Xr selinux 8 972for more information. 973.It Sy copies Ns = Ns Sy 1 Ns | Ns Sy 2 Ns | Ns Sy 3 974Controls the number of copies of data stored for this dataset. 975These copies are in addition to any redundancy provided by the pool, for 976example, mirroring or RAID-Z. 977The copies are stored on different disks, if possible. 978The space used by multiple copies is charged to the associated file and dataset, 979changing the 980.Sy used 981property and counting against quotas and reservations. 982.Pp 983Changing this property only affects newly-written data. 984Therefore, set this property at file system creation time by using the 985.Fl o Sy copies Ns = Ns Ar N 986option. 987.Pp 988Remember that ZFS will not import a pool with a missing top-level vdev. 989Do 990.Em NOT 991create, for example a two-disk striped pool and set 992.Sy copies Ns = Ns Ar 2 993on some datasets thinking you have setup redundancy for them. 994When a disk fails you will not be able to import the pool 995and will have lost all of your data. 996.Pp 997Encrypted datasets may not have 998.Sy copies Ns = Ns Ar 3 999since the implementation stores some encryption metadata where the third copy 1000would normally be. 1001.It Sy devices Ns = Ns Sy on Ns | Ns Sy off 1002Controls whether device nodes can be opened on this file system. 1003The default value is 1004.Sy on . 1005The values 1006.Sy on 1007and 1008.Sy off 1009are equivalent to the 1010.Sy dev 1011and 1012.Sy nodev 1013mount options. 1014.It Xo 1015.Sy dedup Ns = Ns Sy off Ns | Ns Sy on Ns | Ns Sy verify Ns | Ns 1016.Sy sha256 Ns Oo , Ns Sy verify Oc Ns | Ns Sy sha512 Ns Oo , Ns Sy verify Oc Ns | Ns Sy skein Ns Oo , Ns Sy verify Oc Ns | Ns 1017.Sy edonr , Ns Sy verify Ns | Ns Sy blake3 Ns Oo , Ns Sy verify Oc Ns 1018.Xc 1019Configures deduplication for a dataset. 1020The default value is 1021.Sy off . 1022The default deduplication checksum is 1023.Sy sha256 1024(this may change in the future). 1025When 1026.Sy dedup 1027is enabled, the checksum defined here overrides the 1028.Sy checksum 1029property. 1030Setting the value to 1031.Sy verify 1032has the same effect as the setting 1033.Sy sha256 , Ns Sy verify . 1034.Pp 1035If set to 1036.Sy verify , 1037ZFS will do a byte-to-byte comparison in case of two blocks having the same 1038signature to make sure the block contents are identical. 1039Specifying 1040.Sy verify 1041is mandatory for the 1042.Sy edonr 1043algorithm. 1044.Pp 1045Unless necessary, deduplication should 1046.Em not 1047be enabled on a system. 1048See the 1049.Sx Deduplication 1050section of 1051.Xr zfsconcepts 7 . 1052.It Xo 1053.Sy direct Ns = Ns Sy disabled Ns | Ns Sy standard Ns | Ns Sy always 1054.Xc 1055Controls the behavior of Direct I/O requests 1056.Pq e.g. Dv O_DIRECT . 1057The 1058.Sy standard 1059behavior for Direct I/O requests is to bypass the ARC when possible. 1060These requests will not be cached and performance will be limited by the 1061raw speed of the underlying disks 1062.Pq Dv this is the default . 1063.Sy always 1064causes every properly aligned read or write to be treated as a direct request. 1065.Sy disabled 1066causes the O_DIRECT flag to be silently ignored and all direct requests will 1067be handled by the ARC. 1068This is the default behavior for OpenZFS 2.2 and prior releases. 1069.Pp 1070Bypassing the ARC requires that a direct request be correctly aligned. 1071For write requests the starting offset and size of the request must be 1072.Sy recordsize Ns 1073-aligned, if not then the unaligned portion of the request will be silently 1074redirected through the ARC. 1075For read requests there is no 1076.Sy recordsize 1077alignment restriction on either the starting offset or size. 1078All direct requests must use a page-aligned memory buffer and the request 1079size must be a multiple of the page size or an error is returned. 1080.Pp 1081Concurrently mixing buffered and direct requests to overlapping regions of 1082a file can decrease performance. 1083However, the resulting file will always be coherent. 1084For example, a direct read after a buffered write will return the data 1085from the buffered write. 1086Furthermore, if an application uses 1087.Xr mmap 2 1088based file access then in order to maintain coherency all direct requests 1089are converted to buffered requests while the file is mapped. 1090Currently Direct I/O is not supported with zvols. 1091If dedup is enabled on a dataset, Direct I/O writes will not check for 1092deduplication. 1093Deduplication and Direct I/O writes are currently incompatible. 1094.It Xo 1095.Sy dnodesize Ns = Ns Sy legacy Ns | Ns Sy auto Ns | Ns Sy 1k Ns | Ns 1096.Sy 2k Ns | Ns Sy 4k Ns | Ns Sy 8k Ns | Ns Sy 16k 1097.Xc 1098Specifies a compatibility mode or literal value for the size of dnodes in the 1099file system. 1100The default value is 1101.Sy legacy . 1102Setting this property to a value other than 1103.Sy legacy No requires the Sy large_dnode No pool feature to be enabled . 1104.Pp 1105Consider setting 1106.Sy dnodesize 1107to 1108.Sy auto 1109if the dataset uses the 1110.Sy xattr Ns = Ns Sy sa 1111property setting and the workload makes heavy use of extended attributes. 1112This 1113may be applicable to SELinux-enabled systems, Lustre servers, and Samba 1114servers, for example. 1115Literal values are supported for cases where the optimal 1116size is known in advance and for performance testing. 1117.Pp 1118Leave 1119.Sy dnodesize 1120set to 1121.Sy legacy 1122if you need to receive a send stream of this dataset on a pool that doesn't 1123enable the 1124.Sy large_dnode 1125feature, or if you need to import this pool on a system that doesn't support the 1126.Sy large_dnode No feature . 1127.Pp 1128This property can also be referred to by its shortened column name, 1129.Sy dnsize . 1130.It Xo 1131.Sy encryption Ns = Ns Sy off Ns | Ns Sy on Ns | Ns Sy aes-128-ccm Ns | Ns 1132.Sy aes-192-ccm Ns | Ns Sy aes-256-ccm Ns | Ns Sy aes-128-gcm Ns | Ns 1133.Sy aes-192-gcm Ns | Ns Sy aes-256-gcm 1134.Xc 1135Controls the encryption cipher suite (block cipher, key length, and mode) used 1136for this dataset. 1137Requires the 1138.Sy encryption 1139feature to be enabled on the pool. 1140Requires a 1141.Sy keyformat 1142to be set at dataset creation time. 1143.Pp 1144Selecting 1145.Sy encryption Ns = Ns Sy on 1146when creating a dataset indicates that the default encryption suite will be 1147selected, which is currently 1148.Sy aes-256-gcm . 1149In order to provide consistent data protection, encryption must be specified at 1150dataset creation time and it cannot be changed afterwards. 1151.Pp 1152For more details and caveats about encryption see the 1153.Sx Encryption 1154section of 1155.Xr zfs-load-key 8 . 1156.It Sy keyformat Ns = Ns Sy raw Ns | Ns Sy hex Ns | Ns Sy passphrase 1157Controls what format the user's encryption key will be provided as. 1158This property is only set when the dataset is encrypted. 1159.Pp 1160Raw keys and hex keys must be 32 bytes long (regardless of the chosen 1161encryption suite) and must be randomly generated. 1162A raw key can be generated with the following command: 1163.Dl # Nm dd Sy if=/dev/urandom bs=32 count=1 Sy of= Ns Pa /path/to/output/key 1164.Pp 1165Passphrases must be between 8 and 512 bytes long and will be processed through 1166PBKDF2 before being used (see the 1167.Sy pbkdf2iters 1168property). 1169Even though the encryption suite cannot be changed after dataset creation, 1170the keyformat can be with 1171.Nm zfs Cm change-key . 1172.It Xo 1173.Sy keylocation Ns = Ns Sy prompt Ns | Ns Sy file:// Ns Ar /absolute/file/path Ns | Ns Sy https:// Ns Ar address Ns | Ns Sy http:// Ns Ar address 1174.Xc 1175Controls where the user's encryption key will be loaded from by default for 1176commands such as 1177.Nm zfs Cm load-key 1178and 1179.Nm zfs Cm mount Fl l . 1180This property is only set for encrypted datasets which are encryption roots. 1181If unspecified, the default is 1182.Sy prompt . 1183.Pp 1184Even though the encryption suite cannot be changed after dataset creation, the 1185keylocation can be with either 1186.Nm zfs Cm set 1187or 1188.Nm zfs Cm change-key . 1189If 1190.Sy prompt 1191is selected ZFS will ask for the key at the command prompt when it is required 1192to access the encrypted data (see 1193.Nm zfs Cm load-key 1194for details). 1195This setting will also allow the key to be passed in via the standard input 1196stream, 1197but users should be careful not to place keys which should be kept secret on 1198the command line. 1199If a file URI is selected, the key will be loaded from the 1200specified absolute file path. 1201If an HTTPS or HTTP URL is selected, it will be GETted using 1202.Xr fetch 3 , 1203libcurl, or nothing, depending on compile-time configuration and run-time 1204availability. 1205The 1206.Sy SSL_CA_CERT_FILE 1207environment variable can be set to set the location 1208of the concatenated certificate store. 1209The 1210.Sy SSL_CA_CERT_PATH 1211environment variable can be set to override the location 1212of the directory containing the certificate authority bundle. 1213The 1214.Sy SSL_CLIENT_CERT_FILE 1215and 1216.Sy SSL_CLIENT_KEY_FILE 1217environment variables can be set to configure the path 1218to the client certificate and its key. 1219.It Sy pbkdf2iters Ns = Ns Ar iterations 1220Controls the number of PBKDF2 iterations that a 1221.Sy passphrase 1222encryption key should be run through when processing it into an encryption key. 1223This property is only defined when encryption is enabled and a keyformat of 1224.Sy passphrase 1225is selected. 1226The goal of PBKDF2 is to significantly increase the 1227computational difficulty needed to brute force a user's passphrase. 1228This is accomplished by forcing the attacker to run each passphrase through a 1229computationally expensive hashing function many times before they arrive at the 1230resulting key. 1231A user who actually knows the passphrase will only have to pay this cost once. 1232As CPUs become better at processing, this number should be 1233raised to ensure that a brute force attack is still not possible. 1234The current default is 1235.Sy 350000 1236and the minimum is 1237.Sy 100000 . 1238This property may be changed with 1239.Nm zfs Cm change-key . 1240.It Sy exec Ns = Ns Sy on Ns | Ns Sy off 1241Controls whether processes can be executed from within this file system. 1242The default value is 1243.Sy on . 1244The values 1245.Sy on 1246and 1247.Sy off 1248are equivalent to the 1249.Sy exec 1250and 1251.Sy noexec 1252mount options. 1253.It Sy volthreading Ns = Ns Sy on Ns | Ns Sy off 1254Controls internal zvol threading. 1255The value 1256.Sy off 1257disables zvol threading, and zvol relies on application threads. 1258The default value is 1259.Sy on , 1260which enables threading within a zvol. 1261Please note that this property will be overridden by 1262.Sy zvol_request_sync 1263module parameter. 1264This property is only applicable to Linux. 1265.It Sy filesystem_limit Ns = Ns Ar count Ns | Ns Sy none 1266Limits the number of filesystems and volumes that can exist under this point in 1267the dataset tree. 1268The limit is not enforced if the user is allowed to change the limit. 1269Setting a 1270.Sy filesystem_limit 1271to 1272.Sy on 1273a descendant of a filesystem that already has a 1274.Sy filesystem_limit 1275does not override the ancestor's 1276.Sy filesystem_limit , 1277but rather imposes an additional limit. 1278This feature must be enabled to be used 1279.Po see 1280.Xr zpool-features 7 1281.Pc . 1282.It Sy special_small_blocks Ns = Ns Ar size 1283This value represents the threshold block size for including small file 1284blocks into the special allocation class. 1285Blocks smaller than or equal to this 1286value will be assigned to the special allocation class while greater blocks 1287will be assigned to the regular class. 1288Valid values are zero or a power of two from 512 up to 1048576 (1 MiB). 1289The default size is 0 which means no small file blocks 1290will be allocated in the special class. 1291.Pp 1292Before setting this property, a special class vdev must be added to the 1293pool. 1294See 1295.Xr zpoolconcepts 7 1296for more details on the special allocation class. 1297.It Sy mountpoint Ns = Ns Pa path Ns | Ns Sy none Ns | Ns Sy legacy 1298Controls the mount point used for this file system. 1299See the 1300.Sx Mount Points 1301section of 1302.Xr zfsconcepts 7 1303for more information on how this property is used. 1304.Pp 1305When the 1306.Sy mountpoint 1307property is changed for a file system, the file system and any children that 1308inherit the mount point are unmounted. 1309If the new value is 1310.Sy legacy , 1311then they remain unmounted. 1312Otherwise, they are automatically remounted in the new location if the property 1313was previously 1314.Sy legacy 1315or 1316.Sy none . 1317In addition, any shared file systems are unshared and shared in the new 1318location. 1319.Pp 1320When the 1321.Sy mountpoint 1322property is set with 1323.Nm zfs Cm set Fl u 1324, the 1325.Sy mountpoint 1326property is updated but dataset is not mounted or unmounted and remains 1327as it was before. 1328.It Sy nbmand Ns = Ns Sy on Ns | Ns Sy off 1329Controls whether the file system should be mounted with 1330.Sy nbmand 1331.Pq Non-blocking mandatory locks . 1332Changes to this property only take effect when the file system is unmounted and 1333remounted. 1334This was only supported by Linux prior to 5.15, and was buggy there, 1335and is not supported by 1336.Fx . 1337On Solaris it's used for SMB clients. 1338.It Sy overlay Ns = Ns Sy on Ns | Ns Sy off 1339Allow mounting on a busy directory or a directory which already contains 1340files or directories. 1341This is the default mount behavior for Linux and 1342.Fx 1343file systems. 1344On these platforms the property is 1345.Sy on 1346by default. 1347Set to 1348.Sy off 1349to disable overlay mounts for consistency with OpenZFS on other platforms. 1350.It Sy primarycache Ns = Ns Sy all Ns | Ns Sy none Ns | Ns Sy metadata 1351Controls what is cached in the primary cache 1352.Pq ARC . 1353If this property is set to 1354.Sy all , 1355then both user data and metadata is cached. 1356If this property is set to 1357.Sy none , 1358then neither user data nor metadata is cached. 1359If this property is set to 1360.Sy metadata , 1361then only metadata is cached. 1362The default value is 1363.Sy all . 1364.It Sy quota Ns = Ns Ar size Ns | Ns Sy none 1365Limits the amount of space a dataset and its descendants can consume. 1366This property enforces a hard limit on the amount of space used. 1367This includes all space consumed by descendants, including file systems and 1368snapshots. 1369Setting a quota on a descendant of a dataset that already has a quota does not 1370override the ancestor's quota, but rather imposes an additional limit. 1371.Pp 1372Quotas cannot be set on volumes, as the 1373.Sy volsize 1374property acts as an implicit quota. 1375.It Sy snapshot_limit Ns = Ns Ar count Ns | Ns Sy none 1376Limits the number of snapshots that can be created on a dataset and its 1377descendants. 1378Setting a 1379.Sy snapshot_limit 1380on a descendant of a dataset that already has a 1381.Sy snapshot_limit 1382does not override the ancestor's 1383.Sy snapshot_limit , 1384but rather imposes an additional limit. 1385The limit is not enforced if the user is allowed to change the limit. 1386For example, this means that recursive snapshots taken from the global zone are 1387counted against each delegated dataset within a zone. 1388This feature must be enabled to be used 1389.Po see 1390.Xr zpool-features 7 1391.Pc . 1392.It Sy userquota@ Ns Ar user Ns = Ns Ar size Ns | Ns Sy none 1393Limits the amount of space consumed by the specified user. 1394User space consumption is identified by the 1395.Sy userspace@ Ns Ar user 1396property. 1397.Pp 1398Enforcement of user quotas may be delayed by several seconds. 1399This delay means that a user might exceed their quota before the system notices 1400that they are over quota and begins to refuse additional writes with the 1401.Er EDQUOT 1402error message. 1403See the 1404.Nm zfs Cm userspace 1405command for more information. 1406.Pp 1407Unprivileged users can only access their own groups' space usage. 1408The root user, or a user who has been granted the 1409.Sy userquota 1410privilege with 1411.Nm zfs Cm allow , 1412can get and set everyone's quota. 1413.Pp 1414This property is not available on volumes, on file systems before version 4, or 1415on pools before version 15. 1416The 1417.Sy userquota@ Ns Ar … 1418properties are not displayed by 1419.Nm zfs Cm get Sy all . 1420The user's name must be appended after the 1421.Sy @ 1422symbol, using one of the following forms: 1423.Bl -bullet -compact -offset 4n 1424.It 1425POSIX name 1426.Pq Qq joe 1427.It 1428POSIX numeric ID 1429.Pq Qq 789 1430.It 1431SID name 1432.Pq Qq joe.smith@mydomain 1433.It 1434SID numeric ID 1435.Pq Qq S-1-123-456-789 1436.El 1437.Pp 1438Files created on Linux always have POSIX owners. 1439.It Sy userobjquota@ Ns Ar user Ns = Ns Ar size Ns | Ns Sy none 1440The 1441.Sy userobjquota 1442is similar to 1443.Sy userquota 1444but it limits the number of objects a user can create. 1445Please refer to 1446.Sy userobjused 1447for more information about how objects are counted. 1448.It Sy groupquota@ Ns Ar group Ns = Ns Ar size Ns | Ns Sy none 1449Limits the amount of space consumed by the specified group. 1450Group space consumption is identified by the 1451.Sy groupused@ Ns Ar group 1452property. 1453.Pp 1454Unprivileged users can access only their own groups' space usage. 1455The root user, or a user who has been granted the 1456.Sy groupquota 1457privilege with 1458.Nm zfs Cm allow , 1459can get and set all groups' quotas. 1460.It Sy groupobjquota@ Ns Ar group Ns = Ns Ar size Ns | Ns Sy none 1461The 1462.Sy groupobjquota 1463is similar to 1464.Sy groupquota 1465but it limits number of objects a group can consume. 1466Please refer to 1467.Sy userobjused 1468for more information about how objects are counted. 1469.It Sy projectquota@ Ns Ar project Ns = Ns Ar size Ns | Ns Sy none 1470Limits the amount of space consumed by the specified project. 1471Project space consumption is identified by the 1472.Sy projectused@ Ns Ar project 1473property. 1474Please refer to 1475.Sy projectused 1476for more information about how project is identified and set/changed. 1477.Pp 1478The root user, or a user who has been granted the 1479.Sy projectquota 1480privilege with 1481.Nm zfs allow , 1482can access all projects' quota. 1483.It Sy projectobjquota@ Ns Ar project Ns = Ns Ar size Ns | Ns Sy none 1484The 1485.Sy projectobjquota 1486is similar to 1487.Sy projectquota 1488but it limits number of objects a project can consume. 1489Please refer to 1490.Sy userobjused 1491for more information about how objects are counted. 1492.It Sy readonly Ns = Ns Sy on Ns | Ns Sy off 1493Controls whether this dataset can be modified. 1494The default value is 1495.Sy off . 1496The values 1497.Sy on 1498and 1499.Sy off 1500are equivalent to the 1501.Sy ro 1502and 1503.Sy rw 1504mount options. 1505.Pp 1506This property can also be referred to by its shortened column name, 1507.Sy rdonly . 1508.It Sy recordsize Ns = Ns Ar size 1509Specifies a suggested block size for files in the file system. 1510This property is designed solely for use with database workloads that access 1511files in fixed-size records. 1512ZFS automatically tunes block sizes according to internal algorithms optimized 1513for typical access patterns. 1514.Pp 1515For databases that create very large files but access them in small random 1516chunks, these algorithms may be suboptimal. 1517Specifying a 1518.Sy recordsize 1519greater than or equal to the record size of the database can result in 1520significant performance gains. 1521Use of this property for general purpose file systems is strongly discouraged, 1522and may adversely affect performance. 1523.Pp 1524The size specified must be a power of two greater than or equal to 1525.Ar 512 B 1526and less than or equal to 1527.Ar 128 KiB . 1528If the 1529.Sy large_blocks 1530feature is enabled on the pool, the size may be up to 1531.Ar 16 MiB . 1532See 1533.Xr zpool-features 7 1534for details on ZFS feature flags. 1535.Pp 1536However, blocks larger than 1537.Ar 1 MiB 1538can have an impact on i/o latency (e.g. tying up a spinning disk for 1539~300ms), and also potentially on the memory allocator. 1540.Pp 1541Note that maximum size is still limited by default to 1542.Ar 1 MiB 1543on x86_32, see 1544.Sy zfs_max_recordsize 1545module parameter. 1546.Pp 1547Changing the file system's 1548.Sy recordsize 1549affects only files created afterward; existing files are unaffected. 1550.Pp 1551This property can also be referred to by its shortened column name, 1552.Sy recsize . 1553.It Sy redundant_metadata Ns = Ns Sy all Ns | Ns Sy most Ns | Ns Sy some Ns | Ns Sy none 1554Controls what types of metadata are stored redundantly. 1555ZFS stores an extra copy of metadata, so that if a single block is corrupted, 1556the amount of user data lost is limited. 1557This extra copy is in addition to any redundancy provided at the pool level 1558.Pq e.g. by mirroring or RAID-Z , 1559and is in addition to an extra copy specified by the 1560.Sy copies 1561property 1562.Pq up to a total of 3 copies . 1563For example if the pool is mirrored, 1564.Sy copies Ns = Ns 2 , 1565and 1566.Sy redundant_metadata Ns = Ns Sy most , 1567then ZFS stores 6 copies of most metadata, and 4 copies of data and some 1568metadata. 1569.Pp 1570When set to 1571.Sy all , 1572ZFS stores an extra copy of all metadata. 1573If a single on-disk block is corrupt, at worst a single block of user data 1574.Po which is 1575.Sy recordsize 1576bytes long 1577.Pc 1578can be lost. 1579.Pp 1580When set to 1581.Sy most , 1582ZFS stores an extra copy of most types of metadata. 1583This can improve performance of random writes, because less metadata must be 1584written. 1585In practice, at worst about 1000 blocks 1586.Po of 1587.Sy recordsize 1588bytes each 1589.Pc 1590of user data can be lost if a single on-disk block is corrupt. 1591The exact behavior of which metadata blocks are stored redundantly may change in 1592future releases. 1593.Pp 1594When set to 1595.Sy some , 1596ZFS stores an extra copy of only critical metadata. 1597This can improve file create performance since less metadata 1598needs to be written. 1599If a single on-disk block is corrupt, at worst a single user file can be lost. 1600.Pp 1601When set to 1602.Sy none , 1603ZFS does not store any copies of metadata redundantly. 1604If a single on-disk block is corrupt, an entire dataset can be lost. 1605.Pp 1606The default value is 1607.Sy all . 1608.It Sy refquota Ns = Ns Ar size Ns | Ns Sy none 1609Limits the amount of space a dataset can consume. 1610This property enforces a hard limit on the amount of space used. 1611This hard limit does not include space used by descendants, including file 1612systems and snapshots. 1613.It Sy refreservation Ns = Ns Ar size Ns | Ns Sy none Ns | Ns Sy auto 1614The minimum amount of space guaranteed to a dataset, not including its 1615descendants. 1616When the amount of space used is below this value, the dataset is treated as if 1617it were taking up the amount of space specified by 1618.Sy refreservation . 1619The 1620.Sy refreservation 1621reservation is accounted for in the parent datasets' space used, and counts 1622against the parent datasets' quotas and reservations. 1623.Pp 1624If 1625.Sy refreservation 1626is set, a snapshot is only allowed if there is enough free pool space outside of 1627this reservation to accommodate the current number of 1628.Qq referenced 1629bytes in the dataset. 1630.Pp 1631If 1632.Sy refreservation 1633is set to 1634.Sy auto , 1635a volume is thick provisioned 1636.Po or 1637.Qq not sparse 1638.Pc . 1639.Sy refreservation Ns = Ns Sy auto 1640is only supported on volumes. 1641See 1642.Sy volsize 1643in the 1644.Sx Native Properties 1645section for more information about sparse volumes. 1646.Pp 1647This property can also be referred to by its shortened column name, 1648.Sy refreserv . 1649.It Sy relatime Ns = Ns Sy on Ns | Ns Sy off 1650Controls the manner in which the access time is updated when 1651.Sy atime Ns = Ns Sy on 1652is set. 1653Turning this property on causes the access time to be updated relative 1654to the modify or change time. 1655Access time is only updated if the previous 1656access time was earlier than the current modify or change time or if the 1657existing access time hasn't been updated within the past 24 hours. 1658The default value is 1659.Sy on . 1660The values 1661.Sy on 1662and 1663.Sy off 1664are equivalent to the 1665.Sy relatime 1666and 1667.Sy norelatime 1668mount options. 1669.It Sy reservation Ns = Ns Ar size Ns | Ns Sy none 1670The minimum amount of space guaranteed to a dataset and its descendants. 1671When the amount of space used is below this value, the dataset is treated as if 1672it were taking up the amount of space specified by its reservation. 1673Reservations are accounted for in the parent datasets' space used, and count 1674against the parent datasets' quotas and reservations. 1675.Pp 1676This property can also be referred to by its shortened column name, 1677.Sy reserv . 1678.It Sy secondarycache Ns = Ns Sy all Ns | Ns Sy none Ns | Ns Sy metadata 1679Controls what is cached in the secondary cache 1680.Pq L2ARC . 1681If this property is set to 1682.Sy all , 1683then both user data and metadata is cached. 1684If this property is set to 1685.Sy none , 1686then neither user data nor metadata is cached. 1687If this property is set to 1688.Sy metadata , 1689then only metadata is cached. 1690The default value is 1691.Sy all . 1692.It Sy prefetch Ns = Ns Sy all Ns | Ns Sy none Ns | Ns Sy metadata 1693Controls what speculative prefetch does. 1694If this property is set to 1695.Sy all , 1696then both user data and metadata are prefetched. 1697If this property is set to 1698.Sy none , 1699then neither user data nor metadata are prefetched. 1700If this property is set to 1701.Sy metadata , 1702then only metadata are prefetched. 1703The default value is 1704.Sy all . 1705.Pp 1706Please note that the module parameter zfs_prefetch_disable=1 can 1707be used to totally disable speculative prefetch, bypassing anything 1708this property does. 1709.It Sy setuid Ns = Ns Sy on Ns | Ns Sy off 1710Controls whether the setuid bit is respected for the file system. 1711The default value is 1712.Sy on . 1713The values 1714.Sy on 1715and 1716.Sy off 1717are equivalent to the 1718.Sy suid 1719and 1720.Sy nosuid 1721mount options. 1722.It Sy sharesmb Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Ar opts 1723Controls whether the file system is shared by using 1724.Sy Samba USERSHARES 1725and what options are to be used. 1726Otherwise, the file system is automatically shared and unshared with the 1727.Nm zfs Cm share 1728and 1729.Nm zfs Cm unshare 1730commands. 1731If the property is set to on, the 1732.Xr net 8 1733command is invoked to create a 1734.Sy USERSHARE . 1735.Pp 1736Because SMB shares requires a resource name, a unique resource name is 1737constructed from the dataset name. 1738The constructed name is a copy of the 1739dataset name except that the characters in the dataset name, which would be 1740invalid in the resource name, are replaced with underscore (_) characters. 1741Linux does not currently support additional options which might be available 1742on Solaris. 1743.Pp 1744If the 1745.Sy sharesmb 1746property is set to 1747.Sy off , 1748the file systems are unshared. 1749.Pp 1750The share is created with the ACL (Access Control List) "Everyone:F" ("F" 1751stands for "full permissions", i.e. read and write permissions) and no guest 1752access (which means Samba must be able to authenticate a real user \(em 1753.Xr passwd 5 Ns / Ns Xr shadow 5 Ns - , 1754LDAP- or 1755.Xr smbpasswd 5 Ns -based ) 1756by default. 1757This means that any additional access control 1758(disallow specific user specific access etc) must be done on the underlying file 1759system. 1760.Pp 1761When the 1762.Sy sharesmb 1763property is updated with 1764.Nm zfs Cm set Fl u 1765, the property is set to desired value, but the operation to share, reshare 1766or unshare the the dataset is not performed. 1767.It Sy sharenfs Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Ar opts 1768Controls whether the file system is shared via NFS, and what options are to be 1769used. 1770A file system with a 1771.Sy sharenfs 1772property of 1773.Sy off 1774is managed with the 1775.Xr exportfs 8 1776command and entries in the 1777.Pa /etc/exports 1778file. 1779Otherwise, the file system is automatically shared and unshared with the 1780.Nm zfs Cm share 1781and 1782.Nm zfs Cm unshare 1783commands. 1784If the property is set to 1785.Sy on , 1786the dataset is shared using the default options: 1787.Dl sec=sys,rw,crossmnt,no_subtree_check 1788.Pp 1789Please note that the options are comma-separated, unlike those found in 1790.Xr exports 5 . 1791This is done to negate the need for quoting, as well as to make parsing 1792with scripts easier. 1793.Pp 1794For 1795.Fx , 1796there may be multiple sets of options separated by semicolon(s). 1797Each set of options must apply to different hosts or networks and each 1798set of options will create a separate line for 1799.Xr exports 5 . 1800Any semicolon separated option set that consists entirely of whitespace 1801will be ignored. 1802This use of semicolons is only for 1803.Fx 1804at this time. 1805.Pp 1806See 1807.Xr exports 5 1808for the meaning of the default options. 1809Otherwise, the 1810.Xr exportfs 8 1811command is invoked with options equivalent to the contents of this property. 1812.Pp 1813When the 1814.Sy sharenfs 1815property is changed for a dataset, the dataset and any children inheriting the 1816property are re-shared with the new options, only if the property was previously 1817.Sy off , 1818or if they were shared before the property was changed. 1819If the new property is 1820.Sy off , 1821the file systems are unshared. 1822.Pp 1823When the 1824.Sy sharenfs 1825property is updated with 1826.Nm zfs Cm set Fl u 1827, the property is set to desired value, but the operation to share, reshare 1828or unshare the the dataset is not performed. 1829.It Sy logbias Ns = Ns Sy latency Ns | Ns Sy throughput 1830Provide a hint to ZFS about handling of synchronous requests in this dataset. 1831If 1832.Sy logbias 1833is set to 1834.Sy latency 1835.Pq the default , 1836ZFS will use pool log devices 1837.Pq if configured 1838to handle the requests at low latency. 1839If 1840.Sy logbias 1841is set to 1842.Sy throughput , 1843ZFS will not use configured pool log devices. 1844ZFS will instead optimize synchronous operations for global pool throughput and 1845efficient use of resources. 1846.It Sy snapdev Ns = Ns Sy hidden Ns | Ns Sy visible 1847Controls whether the volume snapshot devices under 1848.Pa /dev/zvol/ Ns Aq Ar pool 1849are hidden or visible. 1850The default value is 1851.Sy hidden . 1852.It Sy snapdir Ns = Ns Sy disabled Ns | Ns Sy hidden Ns | Ns Sy visible 1853Controls whether the 1854.Pa .zfs 1855directory is disabled, hidden or visible in the root of the file system as 1856discussed in the 1857.Sx Snapshots 1858section of 1859.Xr zfsconcepts 7 . 1860The default value is 1861.Sy hidden . 1862.It Sy sync Ns = Ns Sy standard Ns | Ns Sy always Ns | Ns Sy disabled 1863Controls the behavior of synchronous requests 1864.Pq e.g. fsync, O_DSYNC . 1865.Sy standard 1866is the POSIX-specified behavior of ensuring all synchronous requests 1867are written to stable storage and all devices are flushed to ensure 1868data is not cached by device controllers 1869.Pq this is the default . 1870.Sy always 1871causes every file system transaction to be written and flushed before its 1872system call returns. 1873This has a large performance penalty. 1874.Sy disabled 1875disables synchronous requests. 1876File system transactions are only committed to stable storage periodically. 1877This option will give the highest performance. 1878However, it is very dangerous as ZFS would be ignoring the synchronous 1879transaction demands of applications such as databases or NFS. 1880Administrators should only use this option when the risks are understood. 1881.It Sy version Ns = Ns Ar N Ns | Ns Sy current 1882The on-disk version of this file system, which is independent of the pool 1883version. 1884This property can only be set to later supported versions. 1885See the 1886.Nm zfs Cm upgrade 1887command. 1888.It Sy volsize Ns = Ns Ar size 1889For volumes, specifies the logical size of the volume. 1890By default, creating a volume establishes a reservation of equal size. 1891For storage pools with a version number of 9 or higher, a 1892.Sy refreservation 1893is set instead. 1894Any changes to 1895.Sy volsize 1896are reflected in an equivalent change to the reservation 1897.Pq or Sy refreservation . 1898The 1899.Sy volsize 1900can only be set to a multiple of 1901.Sy volblocksize , 1902and cannot be zero. 1903.Pp 1904The reservation is kept equal to the volume's logical size to prevent unexpected 1905behavior for consumers. 1906Without the reservation, the volume could run out of space, resulting in 1907undefined behavior or data corruption, depending on how the volume is used. 1908These effects can also occur when the volume size is changed while it is in use 1909.Pq particularly when shrinking the size . 1910Extreme care should be used when adjusting the volume size. 1911.Pp 1912Though not recommended, a 1913.Qq sparse volume 1914.Po also known as 1915.Qq thin provisioned 1916.Pc 1917can be created by specifying the 1918.Fl s 1919option to the 1920.Nm zfs Cm create Fl V 1921command, or by changing the value of the 1922.Sy refreservation 1923property 1924.Po or 1925.Sy reservation 1926property on pool version 8 or earlier 1927.Pc 1928after the volume has been created. 1929A 1930.Qq sparse volume 1931is a volume where the value of 1932.Sy refreservation 1933is less than the size of the volume plus the space required to store its 1934metadata. 1935Consequently, writes to a sparse volume can fail with 1936.Er ENOSPC 1937when the pool is low on space. 1938For a sparse volume, changes to 1939.Sy volsize 1940are not reflected in the 1941.Sy refreservation . 1942A volume that is not sparse is said to be 1943.Qq thick provisioned . 1944A sparse volume can become thick provisioned by setting 1945.Sy refreservation 1946to 1947.Sy auto . 1948.It Sy volmode Ns = Ns Sy default Ns | Ns Sy full Ns | Ns Sy geom Ns | Ns Sy dev Ns | Ns Sy none 1949This property specifies how volumes should be exposed to the OS. 1950Setting it to 1951.Sy full 1952exposes volumes as fully fledged block devices, providing maximal 1953functionality. 1954The value 1955.Sy geom 1956is just an alias for 1957.Sy full 1958and is kept for compatibility. 1959Setting it to 1960.Sy dev 1961hides its partitions. 1962Volumes with property set to 1963.Sy none 1964are not exposed outside ZFS, but can be snapshotted, cloned, replicated, etc, 1965that can be suitable for backup purposes. 1966Value 1967.Sy default 1968means that volumes exposition is controlled by system-wide tunable 1969.Sy zvol_volmode , 1970where 1971.Sy full , 1972.Sy dev 1973and 1974.Sy none 1975are encoded as 1, 2 and 3 respectively. 1976The default value is 1977.Sy full . 1978.It Sy vscan Ns = Ns Sy on Ns | Ns Sy off 1979Controls whether regular files should be scanned for viruses when a file is 1980opened and closed. 1981In addition to enabling this property, the virus scan service must also be 1982enabled for virus scanning to occur. 1983The default value is 1984.Sy off . 1985This property is not used by OpenZFS. 1986.It Sy xattr Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy dir Ns | Ns Sy sa 1987Controls whether extended attributes are enabled for this file system. 1988Two styles of extended attributes are supported: either directory-based 1989or system-attribute-based. 1990.Pp 1991Directory-based extended attributes can be enabled by setting the value to 1992.Sy dir . 1993This style of extended attribute imposes no practical limit 1994on either the size or number of attributes which can be set on a file. 1995Although under Linux the 1996.Xr getxattr 2 1997and 1998.Xr setxattr 2 1999system calls limit the maximum size to 2000.Sy 64K . 2001This is the most compatible 2002style of extended attribute and is supported by all ZFS implementations. 2003.Pp 2004System-attribute-based xattrs can be enabled by setting the value to 2005.Sy sa 2006(default and equal to 2007.Sy on 2008) . 2009The key advantage of this type of xattr is improved performance. 2010Storing extended attributes as system attributes 2011significantly decreases the amount of disk I/O required. 2012Up to 2013.Sy 64K 2014of data may be stored per-file in the space reserved for system attributes. 2015If there is not enough space available for an extended attribute 2016then it will be automatically written as a directory-based xattr. 2017System-attribute-based extended attributes are not accessible 2018on platforms which do not support the 2019.Sy xattr Ns = Ns Sy sa 2020feature. 2021OpenZFS supports 2022.Sy xattr Ns = Ns Sy sa 2023on both 2024.Fx 2025and Linux. 2026.Pp 2027The use of system-attribute-based xattrs is strongly encouraged for users of 2028SELinux or POSIX ACLs. 2029Both of these features heavily rely on extended 2030attributes and benefit significantly from the reduced access time. 2031.Pp 2032The values 2033.Sy on 2034and 2035.Sy off 2036are equivalent to the 2037.Sy xattr 2038and 2039.Sy noxattr 2040mount options. 2041.It Sy jailed Ns = Ns Sy off Ns | Ns Sy on 2042Controls whether the dataset is managed from a jail. 2043See 2044.Xr zfs-jail 8 2045for more information. 2046Jails are a 2047.Fx 2048feature and this property is not available on other platforms. 2049.It Sy zoned Ns = Ns Sy off Ns | Ns Sy on 2050Controls whether the dataset is managed from a non-global zone or namespace. 2051See 2052.Xr zfs-zone 8 2053for more information. 2054Zoning is a 2055Linux 2056feature and this property is not available on other platforms. 2057.El 2058.Pp 2059The following three properties cannot be changed after the file system is 2060created, and therefore, should be set when the file system is created. 2061If the properties are not set with the 2062.Nm zfs Cm create 2063or 2064.Nm zpool Cm create 2065commands, these properties are inherited from the parent dataset. 2066If the parent dataset lacks these properties due to having been created prior to 2067these features being supported, the new file system will have the default values 2068for these properties. 2069.Bl -tag -width "" 2070.It Xo 2071.Sy casesensitivity Ns = Ns Sy sensitive Ns | Ns 2072.Sy insensitive Ns | Ns Sy mixed 2073.Xc 2074Indicates whether the file name matching algorithm used by the file system 2075should be case-sensitive, case-insensitive, or allow a combination of both 2076styles of matching. 2077The default value for the 2078.Sy casesensitivity 2079property is 2080.Sy sensitive . 2081Traditionally, 2082.Ux 2083and POSIX file systems have case-sensitive file names. 2084.Pp 2085The 2086.Sy mixed 2087value for the 2088.Sy casesensitivity 2089property indicates that the file system can support requests for both 2090case-sensitive and case-insensitive matching behavior. 2091Currently, case-insensitive matching behavior on a file system that supports 2092mixed behavior is limited to the SMB server product. 2093For more information about the 2094.Sy mixed 2095value behavior, see the "ZFS Administration Guide". 2096.It Xo 2097.Sy normalization Ns = Ns Sy none Ns | Ns Sy formC Ns | Ns 2098.Sy formD Ns | Ns Sy formKC Ns | Ns Sy formKD 2099.Xc 2100Indicates whether the file system should perform a 2101.Sy Unicode 2102normalization of file names whenever two file names are compared, and which 2103normalization algorithm should be used. 2104File names are always stored unmodified, names are normalized as part of any 2105comparison process. 2106If this property is set to a legal value other than 2107.Sy none , 2108and the 2109.Sy utf8only 2110property was left unspecified, the 2111.Sy utf8only 2112property is automatically set to 2113.Sy on . 2114The default value of the 2115.Sy normalization 2116property is 2117.Sy none . 2118This property cannot be changed after the file system is created. 2119.It Sy utf8only Ns = Ns Sy on Ns | Ns Sy off 2120Indicates whether the file system should reject file names that include 2121characters that are not present in the 2122.Sy UTF-8 2123character code set. 2124If this property is explicitly set to 2125.Sy off , 2126the normalization property must either not be explicitly set or be set to 2127.Sy none . 2128The default value for the 2129.Sy utf8only 2130property is 2131.Sy off . 2132This property cannot be changed after the file system is created. 2133.El 2134.Pp 2135The 2136.Sy casesensitivity , 2137.Sy normalization , 2138and 2139.Sy utf8only 2140properties are also new permissions that can be assigned to non-privileged users 2141by using the ZFS delegated administration feature. 2142. 2143.Ss Temporary Mount Point Properties 2144When a file system is mounted, either through 2145.Xr mount 8 2146for legacy mounts or the 2147.Nm zfs Cm mount 2148command for normal file systems, its mount options are set according to its 2149properties. 2150The correlation between properties and mount options is as follows: 2151.Bl -tag -compact -offset Ds -width "rootcontext=" 2152.It Sy atime 2153atime/noatime 2154.It Sy canmount 2155auto/noauto 2156.It Sy devices 2157dev/nodev 2158.It Sy exec 2159exec/noexec 2160.It Sy readonly 2161ro/rw 2162.It Sy relatime 2163relatime/norelatime 2164.It Sy setuid 2165suid/nosuid 2166.It Sy xattr 2167xattr/noxattr 2168.It Sy nbmand 2169mand/nomand 2170.It Sy context Ns = 2171context= 2172.It Sy fscontext Ns = 2173fscontext= 2174.It Sy defcontext Ns = 2175defcontext= 2176.It Sy rootcontext Ns = 2177rootcontext= 2178.El 2179.Pp 2180In addition, these options can be set on a per-mount basis using the 2181.Fl o 2182option, without affecting the property that is stored on disk. 2183The values specified on the command line override the values stored in the 2184dataset. 2185The 2186.Sy nosuid 2187option is an alias for 2188.Sy nodevices , Ns Sy nosetuid . 2189These properties are reported as 2190.Qq temporary 2191by the 2192.Nm zfs Cm get 2193command. 2194If the properties are changed while the dataset is mounted, the new setting 2195overrides any temporary settings. 2196. 2197.Ss User Properties 2198In addition to the standard native properties, ZFS supports arbitrary user 2199properties. 2200User properties have no effect on ZFS behavior, but applications or 2201administrators can use them to annotate datasets 2202.Pq file systems, volumes, and snapshots . 2203.Pp 2204User property names must contain a colon 2205.Pq Qq Sy \&: 2206character to distinguish them from native properties. 2207They may contain lowercase letters, numbers, and the following punctuation 2208characters: colon 2209.Pq Qq Sy \&: , 2210dash 2211.Pq Qq Sy - , 2212period 2213.Pq Qq Sy \&. , 2214and underscore 2215.Pq Qq Sy _ . 2216The expected convention is that the property name is divided into two portions 2217such as 2218.Ar module : Ns Ar property , 2219but this namespace is not enforced by ZFS. 2220User property names can be at most 256 characters, and cannot begin with a dash 2221.Pq Qq Sy - . 2222.Pp 2223When making programmatic use of user properties, it is strongly suggested to use 2224a reversed DNS domain name for the 2225.Ar module 2226component of property names to reduce the chance that two 2227independently-developed packages use the same property name for different 2228purposes. 2229.Pp 2230The values of user properties are arbitrary strings, are always inherited, and 2231are never validated. 2232All of the commands that operate on properties 2233.Po Nm zfs Cm list , 2234.Nm zfs Cm get , 2235.Nm zfs Cm set , 2236and so forth 2237.Pc 2238can be used to manipulate both native properties and user properties. 2239Use the 2240.Nm zfs Cm inherit 2241command to clear a user property. 2242If the property is not defined in any parent dataset, it is removed entirely. 2243Property values are limited to 8192 bytes. 2244