2 .TH NFS 5 "2 November 2007"
4 nfs \- fstab format and options for the
10 NFS is an Internet Standard protocol
11 created by Sun Microsystems in 1984. NFS was developed
12 to allow file sharing between systems residing
13 on a local area network.
14 The Linux NFS client supports three versions
16 NFS version 2 [RFC1094],
17 NFS version 3 [RFC1813],
18 and NFS version 4 [RFC3530].
22 command attaches a file system to the system's
23 name space hierarchy at a given mount point.
28 should assemble a system's file name hierarchy
29 from various independent file systems
30 (including file systems exported by NFS servers).
33 file describes a single file system, its mount point,
34 and a set of default mount options for that mount point.
36 For NFS file system mounts, a line in the
38 file specifies the server name,
39 the path name of the exported server directory to mount,
40 the local directory that is the mount point,
41 the type of file system that is being mounted,
42 and a list of mount options that control
43 the way the filesystem is mounted and
44 how the NFS client behaves when accessing
45 files on this mount point.
46 The fifth and sixth fields on each line are not used
47 by NFS, thus conventionally each contain the digit zero. For example:
51 .TA 2.5i +0.75i +0.75i +1.0i
52 server:path /mountpoint fstype option,option,... 0 0
55 The server's hostname and export pathname
56 are separated by a colon, while
57 the mount options are separated by commas. The remaining fields
58 are separated by blanks or tabs.
60 The server's hostname can be an unqualified hostname,
61 a fully qualified domain name,
62 a dotted quad IPv4 address, or
63 an IPv6 address enclosed in square brackets.
64 Link-local and site-local IPv6 addresses must be accompanied by an
68 for details on specifying raw IPv6 addresses.
72 field contains "nfs", for whatever version of the protocol.
75 allow several mount options, which are described below.
79 for a description of generic mount options
80 available for all file systems. If you do not need to
81 specify any mount options, use the generic option
86 .SS "Options supported by all versions"
87 These options are valid to use with any NFS version.
90 Determines the recovery behavior of the NFS client
91 after an NFS request times out.
92 If neither option is specified (or if the
94 option is specified), NFS requests are retried indefinitely.
97 option is specified, then the NFS client fails an NFS request
100 retransmissions have been sent,
101 causing the NFS client to return an error
102 to the calling application.
105 A so-called "soft" timeout can cause
106 silent data corruption in certain cases. As such, use the
108 option only when client responsiveness
109 is more important than data integrity.
110 Using NFS over TCP or increasing the value of the
112 option may mitigate some of the risks of using the
117 The time (in tenths of a second) the NFS client waits for a
118 response before it retries an NFS request. If this
119 option is not specified, requests are retried every
120 60 seconds for NFS over TCP.
121 The NFS client does not perform any kind of timeout backoff
124 However, for NFS over UDP, the client uses an adaptive
125 algorithm to estimate an appropriate timeout value for frequently used
126 request types (such as READ and WRITE requests), but uses the
128 setting for infrequently used request types (such as FSINFO requests).
131 option is not specified,
132 infrequently used request types are retried after 1.1 seconds.
133 After each retransmission, the NFS client doubles the timeout for
135 up to a maximum timeout length of 60 seconds.
138 The number of times the NFS client retries a request before
139 it attempts further recovery action. If the
141 option is not specified, the NFS client tries each request
144 The NFS client generates a "server not responding" message
147 retries, then attempts further recovery (depending on whether the
149 mount option is in effect).
152 The maximum number of bytes in each network READ request
153 that the NFS client can receive when reading data from a file
155 The actual data payload size of each NFS READ request is equal to
158 setting. The largest read payload supported by the Linux NFS client
159 is 1,048,576 bytes (one megabyte).
163 value is a positive integral multiple of 1024.
166 values lower than 1024 are replaced with 4096; values larger than
167 1048576 are replaced with 1048576. If a specified value is within the supported
168 range but not a multiple of 1024, it is rounded down to the nearest
173 value is not specified, or if the specified
175 value is larger than the maximum that either client or server can support,
176 the client and server negotiate the largest
178 value that they can both support.
182 mount option as specified on the
184 command line appears in the
186 file. However, the effective
188 value negotiated by the client and server is reported in the
193 The maximum number of bytes per network WRITE request
194 that the NFS client can send when writing data to a file
195 on an NFS server. The actual data payload size of each
196 NFS WRITE request is equal to
199 setting. The largest write payload supported by the Linux NFS client
200 is 1,048,576 bytes (one megabyte).
206 value is a positive integral multiple of 1024.
209 values lower than 1024 are replaced with 4096; values larger than
210 1048576 are replaced with 1048576. If a specified value is within the supported
211 range but not a multiple of 1024, it is rounded down to the nearest
216 value is not specified, or if the specified
218 value is larger than the maximum that either client or server can support,
219 the client and server negotiate the largest
221 value that they can both support.
225 mount option as specified on the
227 command line appears in the
229 file. However, the effective
231 value negotiated by the client and server is reported in the
236 Selects whether the client may cache file attributes. If neither
237 option is specified (or if
239 is specified), the client caches file
242 To improve performance, NFS clients cache file
243 attributes. Every few seconds, an NFS client checks the server's version of each
244 file's attributes for updates. Changes that occur on the server in
245 those small intervals remain undetected until the client checks the
248 option prevents clients from caching file
249 attributes so that applications can more quickly detect file changes
252 In addition to preventing the client from caching file attributes,
255 option forces application writes to become synchronous so
256 that local changes to a file become visible on the server
257 immediately. That way, other clients can quickly detect recent
258 writes when they check the file's attributes.
262 option provides greater cache coherence among NFS clients
263 accessing the same files,
264 but it extracts a significant performance penalty.
265 As such, judicious use of file locking is encouraged instead.
266 The DATA AND METADATA COHERENCE section contains a detailed discussion
270 The minimum time (in seconds) that the NFS client caches
271 attributes of a regular file before it requests
272 fresh attribute information from a server.
273 If this option is not specified, the NFS client uses
277 The maximum time (in seconds) that the NFS client caches
278 attributes of a regular file before it requests
279 fresh attribute information from a server.
280 If this option is not specified, the NFS client uses
284 The minimum time (in seconds) that the NFS client caches
285 attributes of a directory before it requests
286 fresh attribute information from a server.
287 If this option is not specified, the NFS client uses
291 The maximum time (in seconds) that the NFS client caches
292 attributes of a directory before it requests
293 fresh attribute information from a server.
294 If this option is not specified, the NFS client uses
307 If this option is not specified, the NFS client uses
308 the defaults for each of these options listed above.
313 command behaves if an attempt to mount an export fails.
318 to exit with an error status if any part of the mount request
319 times out or fails outright.
320 This is called a "foreground" mount,
321 and is the default behavior if neither the
325 mount option is specified.
329 option is specified, a timeout or failure causes the
331 command to fork a child which continues to attempt
333 The parent immediately returns with a zero exit code.
334 This is known as a "background" mount.
336 If the local mount point directory is missing, the
338 command acts as if the mount request timed out.
339 This permits nested NFS mounts specified in
341 to proceed in any order during system initialization,
342 even if some NFS servers are not yet available.
343 Alternatively these issues can be addressed
344 using an automounter (refer to
349 The number of minutes that the
351 command retries an NFS mount operation
352 in the foreground or background before giving up.
353 If this option is not specified, the default value for foreground mounts
354 is 2 minutes, and the default value for background mounts is 10000 minutes
355 (80 minutes shy of one week).
356 If a value of zero is specified, the
358 command exits immediately after the first failure.
361 The RPCGSS security flavor to use for accessing files on this mount point.
364 option is not specified, or if
366 is specified, the NFS client uses the AUTH_SYS security flavor
367 for all NFS requests on this mount point.
368 Valid security flavors are
381 Refer to the SECURITY CONSIDERATIONS section for details.
383 .BR sharecache " / " nosharecache
384 Determines how the client's data cache and attribute cache are shared
385 when mounting the same export more than once concurrently. Using the
386 same cache reduces memory requirements on the client and presents
387 identical file contents to applications when the same remote file is
388 accessed via different mount points.
390 If neither option is specified, or if the
393 specified, then a single cache is used for all mount points that
394 access the same export. If the
397 then that mount point gets a unique cache. Note that when data and
398 attribute caches are shared, the mount options from the first mount
399 point take effect for subsequent concurrent mounts of the same export.
401 As of kernel 2.6.18, the behavior specified by
403 is legacy caching behavior. This
404 is considered a data risk since multiple cached copies
405 of the same file on the same client can become out of sync
406 following a local update of one of the copies.
408 .BR resvport " / " noresvport
409 Specifies whether the NFS client should use a privileged source port
410 when communicating with an NFS server for this mount point.
411 If this option is not specified, or the
413 option is specified, the NFS client uses a privileged source port.
416 option is specified, the NFS client uses a non-privileged source port.
417 This option is supported in kernels 2.6.28 and later.
419 Using non-privileged source ports helps increase the maximum number of
420 NFS mount points allowed on a client, but NFS servers must be configured
421 to allow clients to connect via non-privileged source ports.
423 Refer to the SECURITY CONSIDERATIONS section for important details.
425 .BI lookupcache= mode
426 Specifies how the kernel manages its cache of directory entries
427 for a given mount point.
435 This option is supported in kernels 2.6.28 and later.
437 The Linux NFS client caches the result of all NFS LOOKUP requests.
438 If the requested directory entry exists on the server,
439 the result is referred to as
441 If the requested directory entry does not exist on the server,
442 the result is referred to as
445 If this option is not specified, or if
447 is specified, the client assumes both types of directory cache entries
448 are valid until their parent directory's cached attributes expire.
451 .BR pos " or " positive
452 is specified, the client assumes positive entries are valid
453 until their parent directory's cached attributes expire, but
454 always revalidates negative entires before an application
460 the client revalidates both types of directory cache entries
461 before an application can use them.
462 This permits quick detection of files that were created or removed
463 by other clients, but can impact application and server performance.
465 The DATA AND METADATA COHERENCE section contains a
466 detailed discussion of these trade-offs.
467 .SS "Options for versions 2 and 3 only"
468 Use these options, along with the options in the above subsection,
469 for NFSv2/v3 only. They will be ignored for newer versions.
472 The transport protocol name and protocol family the NFS client uses
473 to transmit requests to the NFS server for this mount point.
474 If an NFS server has both an IPv4 and an IPv6 address, using a specific
475 netid will force the use of IPv4 or IPv6 networking to communicate
478 If support for TI-RPC is built into the
482 is a valid netid listed in
486 is one of "tcp," "udp," or "rdma," and only IPv4 may be used.
488 Each transport protocol uses different default
493 Refer to the description of these two mount options for details.
495 In addition to controlling how the NFS client transmits requests to
496 the server, this mount option also controls how the
498 command communicates with the server's rpcbind and mountd services.
499 Specifying a netid that uses TCP forces all traffic from the
501 command and the NFS client to use TCP.
502 Specifying a netid that uses UDP forces all traffic types to use UDP.
506 mount option is not specified, the
508 command discovers which protocols the server supports
509 and chooses an appropriate transport for each service.
510 Refer to the TRANSPORT METHODS section for more details.
515 option is an alternative to specifying
517 It is included for compatibility with other operating systems.
522 option is an alternative to specifying
524 It is included for compatibility with other operating systems.
527 The numeric value of the server's NFS service port.
528 If the server's NFS service is not available on the specified port,
529 the mount request fails.
531 If this option is not specified, or if the specified port value is 0,
532 then the NFS client uses the NFS service port number
533 advertised by the server's rpcbind service.
534 The mount request fails if the server's rpcbind service is not available,
535 the server's NFS service is not registered with its rpcbind service,
536 or the server's NFS service is not available on the advertised port.
539 The numeric value of the server's mountd port.
540 If the server's mountd service is not available on the specified port,
541 the mount request fails.
543 If this option is not specified,
544 or if the specified port value is 0, then the
546 command uses the mountd service port number
547 advertised by the server's rpcbind service.
548 The mount request fails if the server's rpcbind service is not available,
549 the server's mountd service is not registered with its rpcbind service,
550 or the server's mountd service is not available on the advertised port.
552 This option can be used when mounting an NFS server
553 through a firewall that blocks the rpcbind protocol.
555 .BI mountproto= netid
556 The transport protocol name and protocol family the NFS client uses
557 to transmit requests to the NFS server's mountd service when performing
558 this mount request, and when later unmounting this mount point.
560 If support for TI-RPC is built into the
564 is a valid netid listed in
568 is one of "tcp" or "udp," and only IPv4 may be used.
570 This option can be used when mounting an NFS server
571 through a firewall that blocks a particular transport.
572 When used in combination with the
574 option, different transports for mountd requests and NFS requests
576 If the server's mountd service is not available via the specified
577 transport, the mount request fails.
579 Refer to the TRANSPORT METHODS section for more on how the
581 mount option interacts with the
586 The hostname of the host running mountd.
587 If this option is not specified, the
589 command assumes that the mountd service runs
590 on the same host as the NFS service.
593 The RPC version number used to contact the server's mountd.
594 If this option is not specified, the client uses a version number
595 appropriate to the requested NFS version.
596 This option is useful when multiple NFS services
597 are running on the same remote server host.
600 The maximum length of a pathname component on this mount.
601 If this option is not specified, the maximum length is negotiated
602 with the server. In most cases, this maximum length is 255 characters.
604 Some early versions of NFS did not support this negotiation.
605 Using this option ensures that
607 reports the proper maximum component length to applications
611 The NFS protocol version number used to contact the server's NFS service.
612 If the server does not support the requested version, the mount request fails.
613 If this option is not specified, the client negociate a suitable version with
614 the server, trying version 4 first, version 3 second, and version 2 last.
617 This option is an alternative to the
620 It is included for compatibility with other operating systems.
622 .BR lock " / " nolock
623 Selects whether to use the NLM sideband protocol to lock files on the server.
624 If neither option is specified (or if
626 is specified), NLM locking is used for this mount point.
629 option, applications can lock files,
630 but such locks provide exclusion only against other applications
631 running on the same client.
632 Remote applications are not affected by these locks.
634 NLM locking must be disabled with the
636 option when using NFS to mount
640 contains files used by the NLM implementation on Linux.
643 option is also required when mounting exports on NFS servers
644 that do not support the NLM protocol.
646 .BR intr " / " nointr
647 Selects whether to allow signals to interrupt file operations
648 on this mount point. If neither option
652 signals do not interrupt NFS file operations. If
654 is specified, system calls return EINTR if an in-progress NFS operation is interrupted by
659 option is preferred to using the
661 option because it is significantly less likely to result in data corruption.
664 .BR intr " / " nointr
665 mount option is deprecated after kernel 2.6.25.
666 Only SIGKILL can interrupt a pending NFS operation on these kernels,
667 and if specified, this mount option is ignored to provide backwards
668 compatibility with older kernels.
671 Selects whether to use close-to-open cache coherence semantics.
672 If neither option is specified (or if
674 is specified), the client uses close-to-open
675 cache coherence semantics. If the
677 option is specified, the client uses a non-standard heuristic to determine when
678 files on the server have changed.
682 option may improve performance for read-only mounts,
683 but should be used only if the data on the server changes only occasionally.
684 The DATA AND METADATA COHERENCE section discusses the behavior
685 of this option in more detail.
688 Selects whether to use the NFSACL sideband protocol on this mount point.
689 The NFSACL sideband protocol is a proprietary protocol
690 implemented in Solaris that manages Access Control Lists. NFSACL was never
691 made a standard part of the NFS protocol specification.
698 the NFS client negotiates with the server
699 to see if the NFSACL protocol is supported,
700 and uses it if the server supports it.
701 Disabling the NFSACL sideband protocol may be necessary
702 if the negotiation causes problems on the client or server.
703 Refer to the SECURITY CONSIDERATIONS section for more details.
705 .BR rdirplus " / " nordirplus
706 Selects whether to use NFS version 3 READDIRPLUS requests.
707 If this option is not specified, the NFS client uses READDIRPLUS requests
708 on NFS version 3 mounts to read small directories.
709 Some applications perform better if the client uses only READDIR requests
711 .SS "Options for version 4 only"
712 Use these options, along with the options in the first subsection above,
713 for NFSv4 only. They will be ignored with older versions.
716 The transport protocol name and protocol family the NFS client uses
717 to transmit requests to the NFS server for this mount point.
718 If an NFS server has both an IPv4 and an IPv6 address, using a specific
719 netid will force the use of IPv4 or IPv6 networking to communicate
722 If support for TI-RPC is built into the
726 is a valid netid listed in
730 is one of "tcp" or "udp," and only IPv4 may be used.
732 All NFS version 4 servers are required to support TCP,
733 so if this mount option is not specified, the NFS version 4 client
734 uses the TCP protocol.
735 Refer to the TRANSPORT METHODS section for more details.
738 The numeric value of the server's NFS service port.
739 If the server's NFS service is not available on the specified port,
740 the mount request fails.
742 If this mount option is not specified,
743 the NFS client uses the standard NFS port number of 2049
744 without first checking the server's rpcbind service.
745 This allows an NFS version 4 client to contact an NFS version 4
746 server through a firewall that may block rpcbind requests.
748 If the specified port value is 0,
749 then the NFS client uses the NFS service port number
750 advertised by the server's rpcbind service.
751 The mount request fails if the server's rpcbind service is not available,
752 the server's NFS service is not registered with its rpcbind service,
753 or the server's NFS service is not available on the advertised port.
755 .BR intr " / " nointr
756 Selects whether to allow signals to interrupt file operations
757 on this mount point. If neither option is specified (or if
759 is specified), system calls return EINTR if an in-progress NFS operation
760 is interrupted by a signal. If
762 is specified, signals do not
763 interrupt NFS operations.
767 option is preferred to using the
769 option because it is significantly less likely to result in data corruption.
772 .BR intr " / " nointr
773 mount option is deprecated after kernel 2.6.25.
774 Only SIGKILL can interrupt a pending NFS operation on these kernels,
775 and if specified, this mount option is ignored to provide backwards
776 compatibility with older kernels.
779 Selects whether to use close-to-open cache coherence semantics
780 for NFS directories on this mount point.
786 the default is to use close-to-open cache coherence
787 semantics for directories.
789 File data caching behavior is not affected by this option.
790 The DATA AND METADATA COHERENCE section discusses
791 the behavior of this option in more detail.
793 .BI clientaddr= n.n.n.n
794 Specifies a single IPv4 address (in dotted-quad form),
795 or a non-link-local IPv6 address,
796 that the NFS client advertises to allow servers
797 to perform NFS version 4 callback requests against
798 files on this mount point. If the server is unable to
799 establish callback connections to clients, performance
800 may degrade, or accesses to files may temporarily hang.
802 If this option is not specified, the
804 command attempts to discover an appropriate callback address automatically.
805 The automatic discovery process is not perfect, however.
806 In the presence of multiple client network interfaces,
807 special routing policies,
808 or atypical network topologies,
809 the exact address to use for callbacks may be nontrivial to determine.
810 .SH nfs4 FILE SYSTEM TYPE
813 file system type is an old syntax for specifying NFSv4 usage. It can still
814 be used with all NFSv4-specific and common options, excepted the
817 .SH MOUNT CONFIGURATION FILE
818 If the mount command is configured to do so, all of the mount options
819 described in the previous section can also be configured in the
820 .I /etc/nfsmount.conf
825 To mount an export using NFS version 2,
828 file system type and specify the
831 To mount using NFS version 3,
834 file system type and specify the
837 To mount using NFS version 4,
840 file system type, with the
846 The following example from an
848 file causes the mount command to negotiate
849 reasonable defaults for NFS behavior.
852 .TA 2.5i +0.7i +0.7i +.7i
853 server:/export /mnt nfs defaults 0 0
856 Here is an example from an /etc/fstab file for an NFS version 2 mount over UDP.
859 .TA 2.5i +0.7i +0.7i +.7i
860 server:/export /mnt nfs nfsvers=2,proto=udp 0 0
863 Try this example to mount using NFS version 4 over TCP
864 with Kerberos 5 mutual authentication.
867 .TA 2.5i +0.7i +0.7i +.7i
868 server:/export /mnt nfs4 sec=krb5 0 0
871 This example can be used to mount /usr over NFS.
874 .TA 2.5i +0.7i +0.7i +.7i
875 server:/export /usr nfs ro,nolock,nocto,actimeo=3600 0 0
878 This example shows how to mount an NFS server
879 using a raw IPv6 link-local address.
882 .TA 2.5i +0.7i +0.7i +.7i
883 [fe80::215:c5ff:fb3e:e2b1%eth0]:/export /mnt nfs defaults 0 0
885 .SH "TRANSPORT METHODS"
886 NFS clients send requests to NFS servers via
887 Remote Procedure Calls, or
889 The RPC client discovers remote service endpoints automatically,
890 handles per-request authentication,
891 adjusts request parameters for different byte endianness on client and server,
892 and retransmits requests that may have been lost by the network or server.
893 RPC requests and replies flow over a network transport.
897 command, NFS client, and NFS server
898 can automatically negotiate proper transport
899 and data transfer size settings for a mount point.
900 In some cases, however, it pays to specify
901 these settings explicitly using mount options.
903 Traditionally, NFS clients used the UDP transport exclusively for
904 transmitting requests to servers. Though its implementation is
905 simple, NFS over UDP has many limitations that prevent smooth
906 operation and good performance in some common deployment
907 environments. Even an insignificant packet loss rate results in the
908 loss of whole NFS requests; as such, retransmit timeouts are usually
909 in the subsecond range to allow clients to recover quickly from
910 dropped requests, but this can result in extraneous network traffic
913 However, UDP can be quite effective in specialized settings where
914 the networks MTU is large relative to NFSs data transfer size (such
915 as network environments that enable jumbo Ethernet frames). In such
916 environments, trimming the
920 settings so that each
921 NFS read or write request fits in just a few network frames (or even
922 in a single frame) is advised. This reduces the probability that
923 the loss of a single MTU-sized network frame results in the loss of
924 an entire large read or write request.
926 TCP is the default transport protocol used for all modern NFS
927 implementations. It performs well in almost every conceivable
928 network environment and provides excellent guarantees against data
929 corruption caused by network unreliability. TCP is often a
930 requirement for mounting a server through a network firewall.
932 Under normal circumstances, networks drop packets much more
933 frequently than NFS servers drop requests. As such, an aggressive
934 retransmit timeout setting for NFS over TCP is unnecessary. Typical
935 timeout settings for NFS over TCP are between one and ten minutes.
936 After the client exhausts its retransmits (the value of the
938 mount option), it assumes a network partition has occurred,
939 and attempts to reconnect to the server on a fresh socket. Since
940 TCP itself makes network data transfer reliable,
944 can safely be allowed to default to the largest values supported by
945 both client and server, independent of the network's MTU size.
946 .SS "Using the mountproto mount option"
947 This section applies only to NFS version 2 and version 3 mounts
948 since NFS version 4 does not use a separate protocol for mount
951 The Linux NFS client can use a different transport for
952 contacting an NFS server's rpcbind service, its mountd service,
953 its Network Lock Manager (NLM) service, and its NFS service.
954 The exact transports employed by the Linux NFS client for
955 each mount point depends on the settings of the transport
956 mount options, which include
959 .BR udp ", and " tcp .
961 The client sends Network Status Manager (NSM) notifications
962 via UDP no matter what transport options are specified, but
963 listens for server NSM notifications on both UDP and TCP.
964 The NFS Access Control List (NFSACL) protocol shares the same
965 transport as the main NFS service.
967 If no transport options are specified, the Linux NFS client
968 uses UDP to contact the server's mountd service, and TCP to
969 contact its NLM and NFS services by default.
971 If the server does not support these transports for these services, the
973 command attempts to discover what the server supports, and then retries
974 the mount request once using the discovered transports.
975 If the server does not advertise any transport supported by the client
976 or is misconfigured, the mount request fails.
979 option is in effect, the mount command backgrounds itself and continues
980 to attempt the specified mount request.
988 option is specified but the
990 option is not, the specified transport is used to contact
991 both the server's mountd service and for the NLM and NFS services.
995 option is specified but none of the
996 .BR proto ", " udp " or " tcp
997 options are specified, then the specified transport is used for the
998 initial mountd request, but the mount command attempts to discover
999 what the server supports for the NFS protocol, preferring TCP if
1000 both transports are supported.
1003 .BR mountproto " and " proto
1005 .BR udp " or " tcp )
1006 options are specified, then the transport specified by the
1008 option is used for the initial mountd request, and the transport
1012 .BR udp " or " tcp " options)"
1013 is used for NFS, no matter what order these options appear.
1014 No automatic service discovery is performed if these options are
1018 .BR proto ", " udp ", " tcp ", "
1021 options are specified more than once on the same mount command line,
1022 then the value of the rightmost instance of each of these options
1024 .SH "DATA AND METADATA COHERENCE"
1025 Some modern cluster file systems provide
1026 perfect cache coherence among their clients.
1027 Perfect cache coherence among disparate NFS clients
1028 is expensive to achieve, especially on wide area networks.
1029 As such, NFS settles for weaker cache coherence that
1030 satisfies the requirements of most file sharing types. Normally,
1031 file sharing is completely sequential:
1032 first client A opens a file, writes something to it, then closes it;
1033 then client B opens the same file, and reads the changes.
1035 .SS "Close-to-open cache consistency"
1036 When an application opens a file stored on an NFS server,
1037 the NFS client checks that it still exists on the server
1038 and is permitted to the opener by sending a GETATTR or ACCESS request.
1039 When the application closes the file,
1040 the NFS client writes back any pending changes
1041 to the file so that the next opener can view the changes.
1042 This also gives the NFS client an opportunity to report
1043 any server write errors to the application
1044 via the return code from
1046 The behavior of checking at open time and flushing at close time
1047 is referred to as close-to-open cache consistency.
1048 .SS "Weak cache consistency"
1049 There are still opportunities for a client's data cache
1050 to contain stale data.
1051 The NFS version 3 protocol introduced "weak cache consistency"
1052 (also known as WCC) which provides a way of efficiently checking
1053 a file's attributes before and after a single request.
1054 This allows a client to help identify changes
1055 that could have been made by other clients.
1057 When a client is using many concurrent operations
1058 that update the same file at the same time
1059 (for example, during asynchronous write behind),
1060 it is still difficult to tell whether it was
1061 that client's updates or some other client's updates
1062 that altered the file.
1063 .SS "Attribute caching"
1066 mount option to achieve attribute cache coherence
1067 among multiple clients.
1068 Almost every file system operation checks
1069 file attribute information.
1070 The client keeps this information cached
1071 for a period of time to reduce network and server load.
1074 is in effect, a client's file attribute cache is disabled,
1075 so each operation that needs to check a file's attributes
1076 is forced to go back to the server.
1077 This permits a client to see changes to a file very quickly,
1078 at the cost of many extra network operations.
1080 Be careful not to confuse the
1082 option with "no data caching."
1085 mount option prevents the client from caching file metadata,
1086 but there are still races that may result in data cache incoherence
1087 between client and server.
1089 The NFS protocol is not designed to support
1090 true cluster file system cache coherence
1091 without some type of application serialization.
1092 If absolute cache coherence among clients is required,
1093 applications should use file locking. Alternatively, applications
1094 can also open their files with the O_DIRECT flag
1095 to disable data caching entirely.
1096 .SS "Directory entry caching"
1097 The Linux NFS client caches the result of all NFS LOOKUP requests.
1098 If the requested directory entry exists on the server,
1099 the result is referred to as a
1100 .IR positive " lookup result.
1101 If the requested directory entry does not exist on the server
1102 (that is, the server returned ENOENT),
1103 the result is referred to as
1104 .IR negative " lookup result.
1106 To detect when directory entries have been added or removed
1108 the Linux NFS client watches a directory's mtime.
1109 If the client detects a change in a directory's mtime,
1110 the client drops all cached LOOKUP results for that directory.
1111 Since the directory's mtime is a cached attribute, it may
1112 take some time before a client notices it has changed.
1113 See the descriptions of the
1114 .BR acdirmin ", " acdirmax ", and " noac
1115 mount options for more information about
1116 how long a directory's mtime is cached.
1118 Caching directory entries improves the performance of applications that
1119 do not share files with applications on other clients.
1120 Using cached information about directories can interfere
1121 with applications that run concurrently on multiple clients and
1122 need to detect the creation or removal of files quickly, however.
1125 mount option allows some tuning of directory entry caching behavior.
1127 Before kernel release 2.6.28,
1128 the Linux NFS client tracked only positive lookup results.
1129 This permitted applications to detect new directory entries
1130 created by other clients quickly while still providing some of the
1131 performance benefits of caching.
1132 If an application depends on the previous lookup caching behavior
1133 of the Linux NFS client, you can use
1134 .BR lookupcache=positive .
1136 If the client ignores its cache and validates every application
1137 lookup request with the server,
1138 that client can immediately detect when a new directory
1139 entry has been either created or removed by another client.
1140 You can specify this behavior using
1141 .BR lookupcache=none .
1142 The extra NFS requests needed if the client does not
1143 cache directory entries can exact a performance penalty.
1144 Disabling lookup caching
1145 should result in less of a performance penalty than using
1147 and has no effect on how the NFS client caches the attributes of files.
1149 .SS "The sync mount option"
1150 The NFS client treats the
1152 mount option differently than some other file systems
1155 for a description of the generic
1164 is specified (or if the
1166 option is specified),
1167 the NFS client delays sending application
1168 writes to the server
1169 until any of these events occur:
1171 Memory pressure forces reclamation of system memory resources.
1173 An application flushes file data explicitly with
1179 An application closes a file with
1182 The file is locked/unlocked via
1185 In other words, under normal circumstances,
1186 data written by an application may not immediately appear
1187 on the server that hosts the file.
1191 option is specified on a mount point,
1192 any system call that writes data to files on that mount point
1193 causes that data to be flushed to the server
1194 before the system call returns control to user space.
1195 This provides greater data cache coherence among clients,
1196 but at a significant performance cost.
1198 Applications can use the O_SYNC open flag to force application
1199 writes to individual files to go to the server immediately without
1203 .SS "Using file locks with NFS"
1204 The Network Lock Manager protocol is a separate sideband protocol
1205 used to manage file locks in NFS version 2 and version 3.
1206 To support lock recovery after a client or server reboot,
1207 a second sideband protocol --
1208 known as the Network Status Manager protocol --
1211 file locking is supported directly in the main NFS protocol,
1212 and the NLM and NSM sideband protocols are not used.
1214 In most cases, NLM and NSM services are started automatically,
1215 and no extra configuration is required.
1216 Configure all NFS clients with fully-qualified domain names
1217 to ensure that NFS servers can find clients to notify them of server reboots.
1219 NLM supports advisory file locks only.
1220 To lock NFS files, use
1222 with the F_GETLK and F_SETLK commands.
1223 The NFS client converts file locks obtained via
1227 When mounting servers that do not support the NLM protocol,
1228 or when mounting an NFS server through a firewall
1229 that blocks the NLM service port,
1232 mount option. NLM locking must be disabled with the
1234 option when using NFS to mount
1238 contains files used by the NLM implementation on Linux.
1242 option may also be advised to improve the performance
1243 of a proprietary application which runs on a single client
1244 and uses file locks extensively.
1245 .SS "NFS version 4 caching features"
1246 The data and metadata caching behavior of NFS version 4
1247 clients is similar to that of earlier versions.
1248 However, NFS version 4 adds two features that improve
1250 .I change attributes
1252 .IR "file delegation" .
1256 is a new part of NFS file and directory metadata
1257 which tracks data changes.
1258 It replaces the use of a file's modification
1259 and change time stamps
1260 as a way for clients to validate the content
1262 Change attributes are independent of the time stamp
1263 resolution on either the server or client, however.
1267 is a contract between an NFS version 4 client
1268 and server that allows the client to treat a file temporarily
1269 as if no other client is accessing it.
1270 The server promises to notify the client (via a callback request) if another client
1271 attempts to access that file.
1272 Once a file has been delegated to a client, the client can
1273 cache that file's data and metadata aggressively without
1274 contacting the server.
1276 File delegations come in two flavors:
1282 delegation means that the server notifies the client
1283 about any other clients that want to write to the file.
1286 delegation means that the client gets notified about
1287 either read or write accessors.
1289 Servers grant file delegations when a file is opened,
1290 and can recall delegations at any time when another
1291 client wants access to the file that conflicts with
1292 any delegations already granted.
1293 Delegations on directories are not supported.
1295 In order to support delegation callback, the server
1296 checks the network return path to the client during
1297 the client's initial contact with the server.
1298 If contact with the client cannot be established,
1299 the server simply does not grant any delegations to
1301 .SH "SECURITY CONSIDERATIONS"
1302 NFS servers control access to file data,
1303 but they depend on their RPC implementation
1304 to provide authentication of NFS requests.
1305 Traditional NFS access control mimics
1306 the standard mode bit access control provided in local file systems.
1307 Traditional RPC authentication uses a number
1308 to represent each user
1309 (usually the user's own uid),
1310 a number to represent the user's group (the user's gid),
1311 and a set of up to 16 auxiliary group numbers
1312 to represent other groups of which the user may be a member.
1314 Typically, file data and user ID values appear unencrypted
1315 (i.e. "in the clear") on the network.
1316 Moreover, NFS versions 2 and 3 use
1317 separate sideband protocols for mounting,
1318 locking and unlocking files,
1319 and reporting system status of clients and servers.
1320 These auxiliary protocols use no authentication.
1322 In addition to combining these sideband protocols with the main NFS protocol,
1323 NFS version 4 introduces more advanced forms of access control,
1324 authentication, and in-transit data protection.
1325 The NFS version 4 specification mandates NFSv4 ACLs,
1326 RPCGSS authentication, and RPCGSS security flavors
1327 that provide per-RPC integrity checking and encryption.
1328 Because NFS version 4 combines the
1329 function of the sideband protocols into the main NFS protocol,
1330 the new security features apply to all NFS version 4 operations
1331 including mounting, file locking, and so on.
1332 RPCGSS authentication can also be used with NFS versions 2 and 3,
1333 but does not protect their sideband protocols.
1337 mount option specifies the RPCGSS security mode
1338 that is in effect on a given NFS mount point.
1341 provides cryptographic proof of a user's identity in each RPC request.
1342 This provides strong verification of the identity of users
1343 accessing data on the server.
1344 Note that additional configuration besides adding this mount option
1345 is required in order to enable Kerberos security.
1348 man page for details.
1350 Two additional flavors of Kerberos security are supported:
1356 security flavor provides a cryptographically strong guarantee
1357 that the data in each RPC request has not been tampered with.
1360 security flavor encrypts every RPC request
1361 to prevent data exposure during network transit; however,
1362 expect some performance impact
1363 when using integrity checking or encryption.
1364 Similar support for other forms of cryptographic security (such as lipkey and SPKM3)
1367 The NFS version 4 protocol allows
1368 clients and servers to negotiate among multiple security flavors
1369 during mount processing.
1370 However, Linux does not yet implement such negotiation.
1371 The Linux client specifies a single security flavor at mount time
1372 which remains in effect for the lifetime of the mount.
1373 If the server does not support this flavor,
1374 the initial mount request is rejected by the server.
1375 .SS "Using non-privileged source ports"
1376 NFS clients usually communicate with NFS servers via network sockets.
1377 Each end of a socket is assigned a port value, which is simply a number
1378 between 1 and 65535 that distinguishes socket endpoints at the same
1380 A socket is uniquely defined by a tuple that includes the transport
1381 protocol (TCP or UDP) and the port values and IP addresses of both
1384 The NFS client can choose any source port value for its sockets,
1385 but usually chooses a
1388 A privileged port is a port value less than 1024.
1389 Only a process with root privileges may create a socket
1390 with a privileged source port.
1392 The exact range of privileged source ports that can be chosen is
1393 set by a pair of sysctls to avoid choosing a well-known port, such as
1394 the port used by ssh.
1395 This means the number of source ports available for the NFS client,
1396 and therefore the number of socket connections that can be used
1398 is practically limited to only a few hundred.
1400 As described above, the traditional default NFS authentication scheme,
1401 known as AUTH_SYS, relies on sending local UID and GID numbers to identify
1402 users making NFS requests.
1403 An NFS server assumes that if a connection comes from a privileged port,
1404 the UID and GID numbers in the NFS requests on this connection have been
1405 verified by the client's kernel or some other local authority.
1406 This is an easy system to spoof, but on a trusted physical network between
1407 trusted hosts, it is entirely adequate.
1409 Roughly speaking, one socket is used for each NFS mount point.
1410 If a client could use non-privileged source ports as well,
1411 the number of sockets allowed,
1412 and thus the maximum number of concurrent mount points,
1413 would be much larger.
1415 Using non-privileged source ports may compromise server security somewhat,
1416 since any user on AUTH_SYS mount points can now pretend to be any other
1417 when making NFS requests.
1418 Thus NFS servers do not support this by default.
1419 They explicitly allow it usually via an export option.
1421 To retain good security while allowing as many mount points as possible,
1422 it is best to allow non-privileged client connections only if the server
1423 and client both require strong authentication, such as Kerberos.
1424 .SS "Mounting through a firewall"
1425 A firewall may reside between an NFS client and server,
1426 or the client or server may block some of its own ports via IP
1428 It is still possible to mount an NFS server through a firewall,
1431 command's automatic service endpoint discovery mechanisms may not work; this
1432 requires you to provide specific endpoint details via NFS mount options.
1434 NFS servers normally run a portmapper or rpcbind daemon to advertise
1435 their service endpoints to clients. Clients use the rpcbind daemon to determine:
1437 What network port each RPC-based service is using
1439 What transport protocols each RPC-based service supports
1441 The rpcbind daemon uses a well-known port number (111) to help clients find a service endpoint.
1442 Although NFS often uses a standard port number (2049),
1443 auxiliary services such as the NLM service can choose
1444 any unused port number at random.
1446 Common firewall configurations block the well-known rpcbind port.
1447 In the absense of an rpcbind service,
1448 the server administrator fixes the port number
1449 of NFS-related services so that the firewall
1450 can allow access to specific NFS service ports.
1451 Client administrators then specify the port number
1452 for the mountd service via the
1457 It may also be necessary to enforce the use of TCP or UDP
1458 if the firewall blocks one of those transports.
1459 .SS "NFS Access Control Lists"
1460 Solaris allows NFS version 3 clients direct access
1461 to POSIX Access Control Lists stored in its local file systems.
1462 This proprietary sideband protocol, known as NFSACL,
1463 provides richer access control than mode bits.
1464 Linux implements this protocol
1465 for compatibility with the Solaris NFS implementation.
1466 The NFSACL protocol never became a standard part
1467 of the NFS version 3 specification, however.
1469 The NFS version 4 specification mandates a new version
1470 of Access Control Lists that are semantically richer than POSIX ACLs.
1471 NFS version 4 ACLs are not fully compatible with POSIX ACLs; as such,
1472 some translation between the two is required
1473 in an environment that mixes POSIX ACLs and NFS version 4.
1481 option is not fully supported.
1482 Generic options, such as
1484 can be modified using the
1487 but NFS-specific options are not all supported.
1488 The underlying transport or NFS version
1489 cannot be changed by a remount, for example.
1490 Performing a remount on an NFS file system mounted with the
1492 option may have unintended consequences.
1495 option is a mixture of a generic option,
1497 and an NFS-specific option
1500 Before 2.4.7, the Linux NFS client did not support NFS over TCP.
1502 Before 2.4.20, the Linux NFS client used a heuristic
1503 to determine whether cached file data was still valid
1504 rather than using the standard close-to-open cache coherency method
1507 Starting with 2.4.22, the Linux NFS client employs
1508 a Van Jacobsen-based RTT estimator to determine retransmit
1509 timeout values when using NFS over UDP.
1511 Before 2.6.0, the Linux NFS client did not support NFS version 4.
1513 Before 2.6.8, the Linux NFS client used only synchronous reads and writes
1515 .BR rsize " and " wsize
1516 settings were smaller than the system's page size.
1518 The Linux NFS client does not yet support
1519 certain optional features of the NFS version 4 protocol,
1520 such as security negotiation, server referrals, and named attributes.
1535 .BR rpc.svcgssd (8),
1538 RFC 768 for the UDP specification.
1540 RFC 793 for the TCP specification.
1542 RFC 1094 for the NFS version 2 specification.
1544 RFC 1813 for the NFS version 3 specification.
1546 RFC 1832 for the XDR specification.
1548 RFC 1833 for the RPC bind specification.
1550 RFC 2203 for the RPCSEC GSS API protocol specification.
1552 RFC 3530 for the NFS version 4 specification.