xref: /freebsd/usr.sbin/nfsd/nfsd.8 (revision d9a42747950146bf03cda7f6e25d219253f8a57a)
1.\" Copyright (c) 1989, 1991, 1993
2.\"	The Regents of the University of California.  All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\"    notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\"    notice, this list of conditions and the following disclaimer in the
11.\"    documentation and/or other materials provided with the distribution.
12.\" 3. Neither the name of the University nor the names of its contributors
13.\"    may be used to endorse or promote products derived from this software
14.\"    without specific prior written permission.
15.\"
16.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
17.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
20.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26.\" SUCH DAMAGE.
27.\"
28.\"	@(#)nfsd.8	8.4 (Berkeley) 3/29/95
29.\" $FreeBSD$
30.\"
31.Dd March 22, 2023
32.Dt NFSD 8
33.Os
34.Sh NAME
35.Nm nfsd
36.Nd remote
37NFS server
38.Sh SYNOPSIS
39.Nm
40.Op Fl ardute
41.Op Fl n Ar num_servers
42.Op Fl h Ar bindip
43.Op Fl p Ar pnfs_setup
44.Op Fl m Ar mirror_level
45.Op Fl V Ar virtual_hostname
46.Op Fl Fl maxthreads Ar max_threads
47.Op Fl Fl minthreads Ar min_threads
48.Sh DESCRIPTION
49The
50.Nm
51utility runs on a server machine to service NFS requests from client machines.
52At least one
53.Nm
54must be running for a machine to operate as a server.
55.Pp
56Unless otherwise specified, eight servers per CPU for UDP transport are
57started.
58.Pp
59When
60.Nm
61is run in an appropriately configured vnet jail, the server is restricted
62to TCP transport and no pNFS service.
63Therefore, the
64.Fl t
65option must be specified and none of the
66.Fl u ,
67.Fl p
68and
69.Fl m
70options can be specified when run in a vnet jail.
71See
72.Xr jail 8
73for more information.
74.Pp
75The following options are available:
76.Bl -tag -width Ds
77.It Fl r
78Register the NFS service with
79.Xr rpcbind 8
80without creating any servers.
81This option can be used along with the
82.Fl u
83or
84.Fl t
85options to re-register NFS if the rpcbind server is restarted.
86.It Fl d
87Unregister the NFS service with
88.Xr rpcbind 8
89without creating any servers.
90.It Fl V Ar virtual_hostname
91Specifies a hostname to be used as a principal name, instead of
92the default hostname.
93.It Fl n Ar threads
94Specifies how many servers to create.
95This option is equivalent to specifying
96.Fl Fl maxthreads
97and
98.Fl Fl minthreads
99with their respective arguments to
100.Ar threads .
101.It Fl Fl maxthreads Ar threads
102Specifies the maximum servers that will be kept around to service requests.
103.It Fl Fl minthreads Ar threads
104Specifies the minimum servers that will be kept around to service requests.
105.It Fl h Ar bindip
106Specifies which IP address or hostname to bind to on the local host.
107This option is recommended when a host has multiple interfaces.
108Multiple
109.Fl h
110options may be specified.
111.It Fl a
112Specifies that nfsd should bind to the wildcard IP address.
113This is the default if no
114.Fl h
115options are given.
116It may also be specified in addition to any
117.Fl h
118options given.
119Note that NFS/UDP does not operate properly when
120bound to the wildcard IP address whether you use -a or do not use -h.
121.It Fl p Ar pnfs_setup
122Enables pNFS support in the server and specifies the information that the
123daemon needs to start it.
124This option can only be used on one server and specifies that this server
125will be the MetaData Server (MDS) for the pNFS service.
126This can only be done if there is at least one
127.Fx
128system configured
129as a Data Server (DS) for it to use.
130.Pp
131The
132.Ar pnfs_setup
133string is a set of fields separated by ',' characters:
134Each of these fields specifies one DS.
135It consists of a server hostname, followed by a ':'
136and the directory path where the DS's data storage file system is mounted on
137this MDS server.
138This can optionally be followed by a '#' and the mds_path, which is the
139directory path for an exported file system on this MDS.
140If this is specified, it means that this DS is to be used to store data
141files for this mds_path file system only.
142If this optional component does not exist, the DS will be used to store data
143files for all exported MDS file systems.
144The DS storage file systems must be mounted on this system before the
145.Nm
146is started with this option specified.
147.br
148For example:
149.sp
150nfsv4-data0:/data0,nfsv4-data1:/data1
151.sp
152would specify two DS servers called nfsv4-data0 and nfsv4-data1 that comprise
153the data storage component of the pNFS service.
154These two DSs would be used to store data files for all exported file systems
155on this MDS.
156The directories
157.Dq /data0
158and
159.Dq /data1
160are where the data storage servers exported
161storage directories are mounted on this system (which will act as the MDS).
162.br
163Whereas, for the example:
164.sp
165nfsv4-data0:/data0#/export1,nfsv4-data1:/data1#/export2
166.sp
167would specify two DSs as above, however nfsv4-data0 will be used to store
168data files for
169.Dq /export1
170and nfsv4-data1 will be used to store data files for
171.Dq /export2 .
172.sp
173When using IPv6 addresses for DSs
174be wary of using link local addresses.
175The IPv6 address for the DS is sent to the client and there is no scope
176zone in it.
177As such, a link local address may not work for a pNFS client to DS
178TCP connection.
179When parsed,
180.Nm
181will only use a link local address if it is the only address returned by
182.Xr getaddrinfo 3
183for the DS hostname.
184.It Fl m Ar mirror_level
185This option is only meaningful when used with the
186.Fl p
187option.
188It specifies the
189.Dq mirror_level ,
190which defines how many of the DSs will
191have a copy of a file's data storage file.
192The default of one implies no mirroring of data storage files on the DSs.
193The
194.Dq mirror_level
195would normally be set to 2 to enable mirroring, but
196can be as high as NFSDEV_MAXMIRRORS.
197There must be at least
198.Dq mirror_level
199DSs for each exported file system on the MDS, as specified in the
200.Fl p
201option.
202This implies that, for the above example using "#/export1" and "#/export2",
203mirroring cannot be done.
204There would need to be two DS entries for each of "#/export1" and "#/export2"
205in order to support a
206.Dq mirror_level
207of two.
208.Pp
209If mirroring is enabled, the server must use the Flexible File
210layout.
211If mirroring is not enabled, the server will use the File layout
212by default, but this default can be changed to the Flexible File layout if the
213.Xr sysctl 8
214vfs.nfsd.default_flexfile
215is set non-zero.
216.It Fl t
217Serve TCP NFS clients.
218.It Fl u
219Serve UDP NFS clients.
220.It Fl e
221Ignored; included for backward compatibility.
222.El
223.Pp
224For example,
225.Dq Li "nfsd -u -t -n 6"
226serves UDP and TCP transports using six daemons.
227.Pp
228A server should run enough daemons to handle
229the maximum level of concurrency from its clients,
230typically four to six.
231.Pp
232The
233.Nm
234utility listens for service requests at the port indicated in the
235NFS server specification; see
236.%T "Network File System Protocol Specification" ,
237RFC1094,
238.%T "NFS: Network File System Version 3 Protocol Specification" ,
239RFC1813,
240.%T "Network File System (NFS) Version 4 Protocol" ,
241RFC7530,
242.%T "Network File System (NFS) Version 4 Minor Version 1 Protocol" ,
243RFC5661,
244.%T "Network File System (NFS) Version 4 Minor Version 2 Protocol" ,
245RFC7862,
246.%T "File System Extended Attributes in NFSv4" ,
247RFC8276 and
248.%T "Parallel NFS (pNFS) Flexible File Layout" ,
249RFC8435.
250.Pp
251If
252.Nm
253detects that
254NFS is not loaded in the running kernel, it will attempt
255to load a loadable kernel module containing NFS support using
256.Xr kldload 2 .
257If this fails, or no NFS KLD is available,
258.Nm
259will exit with an error.
260.Pp
261If
262.Nm
263is to be run on a host with multiple interfaces or interface aliases, use
264of the
265.Fl h
266option is recommended.
267If you do not use the option NFS may not respond to
268UDP packets from the same IP address they were sent to.
269Use of this option
270is also recommended when securing NFS exports on a firewalling machine such
271that the NFS sockets can only be accessed by the inside interface.
272The
273.Nm ipfw
274utility
275would then be used to block NFS-related packets that come in on the outside
276interface.
277.Pp
278If the server has stopped servicing clients and has generated a console message
279like
280.Dq Li "nfsd server cache flooded..." ,
281the value for vfs.nfsd.tcphighwater needs to be increased.
282This should allow the server to again handle requests without a reboot.
283Also, you may want to consider decreasing the value for
284vfs.nfsd.tcpcachetimeo to several minutes (in seconds) instead of 12 hours
285when this occurs.
286.Pp
287Unfortunately making vfs.nfsd.tcphighwater too large can result in the mbuf
288limit being reached, as indicated by a console message
289like
290.Dq Li "kern.ipc.nmbufs limit reached" .
291If you cannot find values of the above
292.Nm sysctl
293values that work, you can disable the DRC cache for TCP by setting
294vfs.nfsd.cachetcp to 0.
295.Pp
296The
297.Nm
298utility has to be terminated with
299.Dv SIGUSR1
300and cannot be killed with
301.Dv SIGTERM
302or
303.Dv SIGQUIT .
304The
305.Nm
306utility needs to ignore these signals in order to stay alive as long
307as possible during a shutdown, otherwise loopback mounts will
308not be able to unmount.
309If you have to kill
310.Nm
311just do a
312.Dq Li "kill -USR1 <PID of master nfsd>"
313.Sh EXIT STATUS
314.Ex -std
315.Sh SEE ALSO
316.Xr nfsstat 1 ,
317.Xr kldload 2 ,
318.Xr nfssvc 2 ,
319.Xr nfsv4 4 ,
320.Xr pnfs 4 ,
321.Xr pnfsserver 4 ,
322.Xr exports 5 ,
323.Xr stablerestart 5 ,
324.Xr gssd 8 ,
325.Xr ipfw 8 ,
326.Xr jail 8 ,
327.Xr mountd 8 ,
328.Xr nfsiod 8 ,
329.Xr nfsrevoke 8 ,
330.Xr nfsuserd 8 ,
331.Xr rpcbind 8
332.Sh HISTORY
333The
334.Nm
335utility first appeared in
336.Bx 4.4 .
337.Sh BUGS
338If
339.Nm
340is started when
341.Xr gssd 8
342is not running, it will service AUTH_SYS requests only.
343To fix the problem you must kill
344.Nm
345and then restart it, after the
346.Xr gssd 8
347is running.
348.Pp
349For a Flexible File Layout pNFS server,
350if there are Linux clients doing NFSv4.1 or NFSv4.2 mounts, those
351clients might need the
352.Xr sysctl 8
353vfs.nfsd.flexlinuxhack
354to be set to one on the MDS as a workaround.
355.Pp
356Linux 5.n kernels appear to have been patched such that this
357.Xr sysctl 8
358does not need to be set.
359