d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Modify nfsd.c to add support for the experimental nfs server.

Browse Source 
This includes the addition of a new flag "-4" that will force
use of the experimental server with nfsv4 support in it. This
commit also adds two new man pages to the repository that are
NFSv4 specific. One describes the file used by the server to
restart nfsv4 services safely. The other is a brief overview
of nfsv4 and its setup.

Reviewed by:	dfr
Approved by:	kib (mentor)
This commit is contained in:
Rick Macklem 2009-05-24 01:18:56 +00:00
parent e17bb250c9
commit 20d0a5e544
Notes: svn2git 2020-12-20 02:59:44 +00:00 
svn path=/head/; revision=192674
5 changed files with 544 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
usr.sbin/nfsd/Makefile
View File

@ -2,7 +2,7 @@
# $FreeBSD$
PROG= nfsd
MAN= nfsd.8
MAN= nfsd.8 nfsv4.4 stablerestart.5
WARNS?= 6

36
usr.sbin/nfsd/nfsd.8
View File

@ -38,7 +38,7 @@
server
.Sh SYNOPSIS
.Nm
.Op Fl ardut
.Op Fl ardut4
.Op Fl n Ar num_servers
.Op Fl h Ar bindip
.Sh DESCRIPTION
@ -100,6 +100,18 @@ clients.
Serve
.Tn UDP NFS
clients.
.It Fl 4
Forces
.Nm
to try and start the experimental server that includes NFSv4 support in it.
If this flag isn't specified, the experimental server will only be started
if it is linked into the kernel and the regular one isn't.
.br
ie. The kernel is built with the following:
.Bd -literal -offset indent -compact
# options NFSSERVER
options NFSD
.Ed
.El
.Pp
For example,
@ -120,8 +132,11 @@ utility listens for service requests at the port indicated in the
.Tn NFS
server specification; see
.%T "Network File System Protocol Specification" ,
RFC1094 and
.%T "NFS: Network File System Version 3 Protocol Specification" .
RFC1094,
.%T "NFS: Network File System Version 3 Protocol Specification" ,
RFC1813 and
.%T "Network File System (NFS) Version 4 Protocol" ,
RFC3530.
.Pp
If
.Nm
@ -178,13 +193,28 @@ just do a
.Xr nfsstat 1 ,
.Xr kldload 2 ,
.Xr nfssvc 2 ,
.Xr nfsv4 4 ,
.Xr exports 5 ,
.Xr gssd 8 ,
.Xr ipfw 8 ,
.Xr mountd 8 ,
.Xr nfsiod 8 ,
.Xr nfsrevoke 8 ,
.Xr nfsuserd 8 ,
.Xr rpcbind 8
.Sh HISTORY
The
.Nm
utility first appeared in
.Bx 4.4 .
.Sh BUGS
If
.Nm
is started when
.Xr gssd 8
is not running, it will service AUTH_SYS requests only. To fix the problem
you must kill
.Nm
and then restart it, after the
.Xr gssd 8
is running.

133
usr.sbin/nfsd/nfsd.c
View File

@ -48,6 +48,7 @@ static const char rcsid[] =
#include <sys/syslog.h>
#include <sys/wait.h>
#include <sys/mount.h>
#include <sys/fcntl.h>
#include <sys/linker.h>
#include <sys/module.h>
@ -59,6 +60,7 @@ static const char rcsid[] =
#include <nfs/rpcv2.h>
#include <nfs/nfsproto.h>
#include <nfsserver/nfs.h>
#include <nfs/nfssvc.h>
#include <err.h>
#include <errno.h>
@ -77,11 +79,14 @@ int debug = 1;
int debug = 0;
#endif
#define NFSD_STABLERESTART "/var/db/nfs-stablerestart"
#define MAXNFSDCNT 256
#define DEFNFSDCNT 4
pid_t children[MAXNFSDCNT]; /* PIDs of children */
int nfsdcnt; /* number of children */
int new_syscall;
int run_v4server = 0; /* Force running of nfsv4 server */
int nfssvc_nfsd; /* Set to correct NFSSVC_xxx flag */
void cleanup(int);
void child_cleanup(int);
@ -112,6 +117,7 @@ void usage(void);
* -d - unregister with rpcbind
* -t - support tcp nfs clients
* -u - support udp nfs clients
* -4 - forces it to run a server that supports nfsv4
* followed by "n" which is the number of nfsds' to fork off
*/
int
@ -131,20 +137,15 @@ main(int argc, char **argv)
int tcp6sock, ip6flag, tcpflag, tcpsock;
int udpflag, ecode, s, srvcnt;
int bindhostc, bindanyflag, rpcbreg, rpcbregcnt;
int stablefd, nfssvc_addsock;
char **bindhost = NULL;
pid_t pid;
if (modfind("nfsserver") < 0) {
/* Not present in kernel, try loading it */
if (kldload("nfsserver") < 0 || modfind("nfsserver") < 0)
errx(1, "NFS server is not available");
}
nfsdcnt = DEFNFSDCNT;
unregister = reregister = tcpflag = maxsock = 0;
bindanyflag = udpflag = connect_type_cnt = bindhostc = 0;
#define GETOPT "ah:n:rdtu"
#define USAGE "[-ardtu] [-n num_servers] [-h bindip]"
#define GETOPT "ah:n:rdtu4"
#define USAGE "[-ardtu4] [-n num_servers] [-h bindip]"
while ((ch = getopt(argc, argv, GETOPT)) != -1)
switch (ch) {
case 'a':
@ -179,6 +180,9 @@ main(int argc, char **argv)
case 'u':
udpflag = 1;
break;
case '4':
run_v4server = 1;
break;
default:
case '?':
usage();
@ -203,6 +207,25 @@ main(int argc, char **argv)
}
}
/*
* If the "-4" option was specified OR only the nfsd module is
* found in the server, run "nfsd".
* Otherwise, try and run "nfsserver".
*/
if (run_v4server > 0) {
if (modfind("nfsd") < 0) {
/* Not present in kernel, try loading it */
if (kldload("nfsd") < 0 || modfind("nfsd") < 0)
errx(1, "NFS server is not available");
}
} else if (modfind("nfsserver") < 0 && modfind("nfsd") >= 0) {
run_v4server = 1;
} else if (modfind("nfsserver") < 0) {
/* Not present in kernel, try loading it */
if (kldload("nfsserver") < 0 || modfind("nfsserver") < 0)
errx(1, "NFS server is not available");
}
ip6flag = 1;
s = socket(AF_INET6, SOCK_DGRAM, IPPROTO_UDP);
if (s == -1) {
@ -328,15 +351,47 @@ main(int argc, char **argv)
openlog("nfsd", LOG_PID, LOG_DAEMON);
/*
* Figure out if the kernel supports the new-style
* NFSSVC_NFSD. Old kernels will return ENXIO because they
* don't recognise the flag value, new ones will return EINVAL
* because argp is NULL.
* For V4, we open the stablerestart file and call nfssvc()
* to get it loaded. This is done before the daemons do the
* regular nfssvc() call to service NFS requests.
* (This way the file remains open until the last nfsd is killed
* off.)
* Note that this file is not created by this daemon and can
* only be relocated by recompiling the daemon, in order to
* minimize accidentally starting up with the wrong file.
* If should be created as an empty file Read and Write for
* root before the first time you run NFS v4 and should never
* be re-initialized if at all possible. It should live on a
* local, non-volatile storage device that does not do hardware
* level write-back caching. (See SCSI doc for more information
* on how to prevent write-back caching on SCSI disks.)
*/
new_syscall = FALSE;
if (nfssvc(NFSSVC_NFSD, NULL) < 0 && errno == EINVAL)
if (run_v4server > 0) {
stablefd = open(NFSD_STABLERESTART, O_RDWR, 0);
if (stablefd < 0) {
syslog(LOG_ERR, "Can't open %s\n", NFSD_STABLERESTART);
exit(1);
}
if (nfssvc(NFSSVC_STABLERESTART, (caddr_t)&stablefd) < 0) {
syslog(LOG_ERR, "Can't read stable storage file\n");
exit(1);
}
nfssvc_addsock = NFSSVC_NFSDADDSOCK;
nfssvc_nfsd = NFSSVC_NFSDNFSD;
new_syscall = TRUE;
new_syscall = FALSE;
} else {
nfssvc_addsock = NFSSVC_ADDSOCK;
nfssvc_nfsd = NFSSVC_NFSD;
/*
* Figure out if the kernel supports the new-style
* NFSSVC_NFSD. Old kernels will return ENXIO because they
* don't recognise the flag value, new ones will return EINVAL
* because argp is NULL.
*/
new_syscall = FALSE;
if (nfssvc(NFSSVC_NFSD, NULL) < 0 && errno == EINVAL)
new_syscall = TRUE;
}
if (!new_syscall) {
/* If we use UDP only, we start the last server below. */
@ -413,7 +468,7 @@ main(int argc, char **argv)
addsockargs.sock = sock;
addsockargs.name = NULL;
addsockargs.namelen = 0;
if (nfssvc(NFSSVC_ADDSOCK, &addsockargs) < 0) {
if (nfssvc(nfssvc_addsock, &addsockargs) < 0) {
syslog(LOG_ERR, "can't Add UDP socket");
nfsd_exit(1);
}
@ -481,7 +536,7 @@ main(int argc, char **argv)
addsockargs.sock = sock;
addsockargs.name = NULL;
addsockargs.namelen = 0;
if (nfssvc(NFSSVC_ADDSOCK, &addsockargs) < 0) {
if (nfssvc(nfssvc_addsock, &addsockargs) < 0) {
syslog(LOG_ERR,
"can't add UDP6 socket");
nfsd_exit(1);
@ -711,7 +766,7 @@ main(int argc, char **argv)
addsockargs.sock = msgsock;
addsockargs.name = (caddr_t)&inetpeer;
addsockargs.namelen = len;
nfssvc(NFSSVC_ADDSOCK, &addsockargs);
nfssvc(nfssvc_addsock, &addsockargs);
(void)close(msgsock);
} else if (FD_ISSET(tcpsock, &v6bits)) {
len = sizeof(inet6peer);
@ -733,7 +788,7 @@ main(int argc, char **argv)
addsockargs.sock = msgsock;
addsockargs.name = (caddr_t)&inet6peer;
addsockargs.namelen = len;
nfssvc(NFSSVC_ADDSOCK, &addsockargs);
nfssvc(nfssvc_addsock, &addsockargs);
(void)close(msgsock);
}
}
@ -861,19 +916,47 @@ nfsd_exit(int status)
void
start_server(int master)
{
char principal[128];
char hostname[128];
char principal[MAXHOSTNAMELEN + 5];
struct nfsd_nfsd_args nfsdargs;
int status;
int status, error;
char hostname[MAXHOSTNAMELEN + 1], *cp;
struct addrinfo *aip, hints;
status = 0;
if (new_syscall) {
gethostname(hostname, sizeof(hostname));
snprintf(principal, sizeof(principal), "nfs@%s", hostname);
gethostname(hostname, sizeof (hostname));
snprintf(principal, sizeof (principal), "nfs@%s", hostname);
if ((cp = strchr(hostname, '.')) == NULL ||
*(cp + 1) == '\0') {
/* If not fully qualified, try getaddrinfo() */
memset((void *)&hints, 0, sizeof (hints));
hints.ai_flags = AI_CANONNAME;
error = getaddrinfo(hostname, NULL, &hints, &aip);
if (error == 0) {
if (aip->ai_canonname != NULL &&
(cp = strchr(aip->ai_canonname, '.')) !=
NULL && *(cp + 1) != '\0')
snprintf(principal, sizeof (principal),
"nfs@%s", aip->ai_canonname);
freeaddrinfo(aip);
}
}
nfsdargs.principal = principal;
nfsdargs.minthreads = nfsdcnt;
nfsdargs.maxthreads = nfsdcnt;
if (nfssvc(NFSSVC_NFSD, &nfsdargs) < 0) {
error = nfssvc(nfssvc_nfsd, &nfsdargs);
if (error < 0 && errno == EAUTH) {
/*
* This indicates that it could not register the
* rpcsec_gss credentials, usually because the
* gssd daemon isn't running.
* (only the experimental server with nfsv4)
*/
syslog(LOG_ERR, "No gssd, using AUTH_SYS only");
principal[0] = '\0';
error = nfssvc(nfssvc_nfsd, &nfsdargs);
}
if (error < 0) {
syslog(LOG_ERR, "nfssvc: %m");
status = 1;
}

309
usr.sbin/nfsd/nfsv4.4 Normal file
View File

@ -0,0 +1,309 @@
.\" Copyright (c) 2009 Rick Macklem, University of Guelph
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" $FreeBSD$
.\"
.Dd April 30, 2009
.Dt NFSV4 4
.Os
.Sh NAME
.Nm nfsv4
.Nd NFS Version 4 Protocol
.Sh SYNOPSIS
experimental client and server with NFSv4 support
.Sh DESCRIPTION
The experimental nfs client and server provides support for the
.Tn NFSv4
specification; see
.%T "Network File System (NFS) Version 4 Protocol \\*(tNRFC\\*(sP 3530" .
The protocol is somewhat similar to NFS Version 3, but differs in significant
ways. It uses a single Compound RPC that concatenates operations to-gether.
Each of these operations are similar to the RPCs of NFS Version 3.
The operations in the compound are performed in order, until one of
them fails (returns an error) and then the RPC terminates at that point.
.Pp
It has
integrated locking support, which implies that the server is no longer
stateless. As such, the
.Tn NFSv4
server remains in recovery mode for a Grace period (always greater than the
lease duration the server uses) after a reboot.
During this Grace period, clients may recover state but not perform other
open/lock state changing operations.
To provide for correct recovery semantics, a small file described by
.Xr stablerestart 5
is used by the server during the recovery phase. If this file is missing,
the server will not start.
If this file is lost, it should be recovered from backups, since creating
an empty
.Xr stablerestart 5
file will result in the server starting without providing a Grace Period
for recovery.
Note that recovery only occurs when the server
machine is rebooted, not when the
.Xr nfsd 8
are just restarted.
.Pp
It provides several optional features not in NFS Version 3:
.sp
.Bd -literal -offset indent -compact
- NFS Version 4 ACLs
- Referrals, which redirect subtrees to other servers
(not yet implemented)
- Delegations, which allow a client to operate on a file locally
.Ed
.Pp
The
.Tn NFSv4
protocol does not use a separate mount protocol and assumes that the
server provides a single file system tree structure, rooted at the point
in the local file system tree specified by the
.sp 1
.Bd -literal -offset indent -compact
V4: <rootdir>
.Ed
.sp 1
line in
.Xr exports 5 .
The
.Xr nfsd 8
allows a limited subset of operations to be performed on non-exported subtrees
of the local file system, so that traversal of the tree to the exported
subtrees is possible.
As such, the ``<rootdir>'' can be in a non-exported file system. However,
the entire tree that is rooted at that point must be in local file systems
that are of types that can be NFS exported.
Since the
.Nm
file system is rooted at ``<rootdir>'', setting this to anything other
than ``/'' will result in clients being required to use different mount
paths for
.Nm
than for NFS Version 2 or 3.
Unlike NFS Version 2 and 3, Version 4 allows a client mount to span across
multiple server file systems, although not all clients are capable of doing
this.
.Pp
.Nm
uses names for users and groups instead of numbers. On the wire, they
take the form:
.sp
.Bd -literal -offset indent -compact
<user>@<dns.domain>
.Ed
.sp
where ``<dns.domain>'' is not the same as the DNS domain used
for host name lookups, but is usually set to the same string. Most systems set this ``<dns.domain>''
to the domain name part of the machine's
.Xr hostname 1
by default. However, this can normally be overridden by a command line
option or configuration file for the daemon used to do the name<->number
mapping.
On FreeBSD, the mapping daemon is called
.Xr nfsuserd 8
and has a command line option that overrides the domain component of the
machine's hostname.
This can be set in
.Xr rc.conf 5
via the nfsv4_userd_flags variable.
For use of
.Nm ,
either client or server, this daemon must be enabled by setting
.sp
.Bd -literal -offset indent -compact
nfsv4_enable="YES"
.Ed
.sp
in
.Xr rc.conf 5 .
If this ``<dns.domain>'' is not set correctly or the daemon is not running, ``ls -l'' will typically
report a lot of ``nobody'' and ``nogroup'' ownerships.
.Pp
Although numbers are no longer used in the
.Nm
protocol, they will still be in the RPC authentication fields when running
using AUTH_SYS (sec=sys), which is the default.
As such, in this case both the user/group name and number spaces must
be consistent between the client and server.
.Pp
However, if you run
.Nm
with RPCSEC_GSS (sec=krb5, krb5i, krb5p), only names and KerberosV tickets
will go on the wire.
.Sh SERVER SETUP
.Pp
To set up the experimental nfs server that supports
.Nm
you will need to either build a kernel with:
.sp
.Bd -literal -offset indent -compact
options NFSD
.Ed
and not
.Bd -literal -offset indent -compact
options NFSSERVER
.Ed
.sp
or start
.Xr mountd 8
and
.Xr nfsd 8
with the ``-4'' option to force use of the experimental server.
This will occur if
.sp
.Bd -literal -offset indent -compact
nfsv4_enable="YES"
.Ed
.sp
is set in
.Xr rc.conf 5 ,
as above.
.Pp
You will also need to add a:
.sp
.Bd -literal -offset indent -compact
V4: <rootdir>
.Ed
.sp
line to the
.Xr exports 5
file, to tell the server where the
.Nm
tree is rooted.
.Pp
If the file systems you are exporting are only being accessed via
.Nm
there are a couple of
.Xr sysctl 8
variables that you can change, which might improve performance.
.Bl -tag -width Ds
.It Cm vfs.newnfs.issue_delegations
when set non-zero, allows the server to issue Open Delegations to
clients. These delegations permit the client to manipulate the file
locally on the client. Unfortunately, at this time, client use of
delegations is limited, so performance gains may not be observed.
This can only be enabled when the file systems being exported to
.Nm
clients are not being accessed locally on the server and, if being
accessed via NFS Version 2 or 3 clients, these clients cannot be
using the NLM.
.It Cm vfs.newnfs.enable_locallocks
can be set to 0 to disable acquisition of local byte range locks.
Disabling local locking can only be done if neither local accesses
to the exported file systems nor the NLM is operating on them.
.El
.sp
Note that Samba server access would be considered ``local access'' for the above
discussion.
.Pp
To build a kernel with the experimental
.Nm
linked into it, the
.sp
.Bd -literal -offset indent -compact
options NFSD
.Ed
.sp
must be specified in the kernel's
.Xr config 5
file.
.Sh CLIENT MOUNTS
.Pp
To do an
.Nm
mount, specify the ``nfsv4'' option on the
.Xr mount_nfs 8
command line.
This will force use of the experimental client plus set ``tcp'' and
.Nm .
.Pp
If the
.Nm
server that is being mounted on supports delegations, you can set
.sp
.Bd -literal -offset indent -compact
nfsv4_callbackdaemon_enable="YES"
.Ed
.sp
via
.Xr rc.conf 5
so that the client side callback daemon
.Xr nfscbd 8
is started upon boot.
Without a functioning callback path, a server will never issue Delegations
to a client.
.sp
By default, the callback address will be set to the IP address acquired via
rtalloc() in the kernel and port# 7745.
To override the default port#, a command line option for
.Xr nfscbd 8
can be set via the variable
.sp
.Bd -literal -offset indent -compact
nfsv4_callbackdaemon_flag
.Ed
.sp
using
.Xr rc.conf 5 .
.sp
To get callbacks to work when behind a NAT gateway, a port for the callback
service will need to be set up on the NAT gateway and then the address
of the NAT gateway (host IP plus port#) will need to be set by assigning the
.Xr sysctl 8
variable vfs.newnfs.callback_addr to a string of the form:
.sp
N.N.N.N.N.N
.sp
where the first 4 Ns are the host IP address and the last two are the
port# in network byte order (all decimal #s in the range 0-255).
.Pp
To build a kernel with the experimental
.Nm
client linked into it, the option
.sp
.Bd -literal -offset indent -compact
options NFSCL
.Ed
.sp
must be specified in the kernel's
.Xr config 5
file.
.Sh FILES
.Bl -tag -width /var/db/nfs-stablerestart -compact
.It Pa /var/db/nfs-stablerestart
NFS V4 stable restart file
.El
.Sh SEE ALSO
.Xr stablerestart 5
.Xr mountd 8
.Xr nfscbd 8
.Xr nfsd 8
.Xr nfsdumpstate 8
.Xr nfsrevoke 8
.Xr nfsuserd 8
.Sh BUGS
At this time, there is no recall of delegations for local file system
operations. As such, delegations should only be enabled for file systems
that are being used soley as NFS export volumes and are not being accessed
via local system calls nor services such as Samba.

93
usr.sbin/nfsd/stablerestart.5 Normal file
View File

@ -0,0 +1,93 @@
.\" Copyright (c) 2009 Rick Macklem, University of Guelph
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" $FreeBSD$
.\"
.Dd Sept 7, 2007
.Dt STABLERESTART 5
.Os
.Sh NAME
.Nm nfs-stablerestart
.Nd handles restart edge conditions for the
.Tn NFS
V4 server
.Sh SYNOPSIS
.Nm nfs-stablerestart
.Sh DESCRIPTION
The
.Nm
file holds information that allows the
.Tn NFS
V4 server to restart without always returning the NFSERR_NOGRACE error, as described in the
.Tn NFS V4
server specification; see
.%T "Network File System (NFS) Version 4 Protocol \\*(tNRFC\\*(sP 3530, Section 8.6.3" .
.Pp
The first record in the file, as defined by struct nfsf_rec in
/usr/include/fs/nfs/nfsrvstate.h, holds the lease duration of the
last incarnation of the server and the number of boot times that follows.
Following this are the number of previous boot times listed in the
first record.
The lease duration is used to set the Grace Period.
The boot times
are used to avoid the unlikely occurrence of a boot time being reused,
due to a TOD clock going backwards. This record and the previous boot times with this boot time added is re-written at the
end of the Grace Period.
.Pp
The rest of the file are appended records, as defined by
struct nfst_rec in /usr/include/fs/nfs/nfsrvstate.h and are used
represent one of two things. There are records which indicate that a
client successfully aquired state and records that indicate a client's state was revoked.
State revoke records indicate that state information
for a client was discarded, due to lease expiry and an otherwise
conflicting open or lock request being made by a different client.
These records can be used
to determine if clients might have done either of the
edge conditions.
.Pp
If a client might have done either edge condition or this file is
empty or corrupted, the server returns NFSERR_NOGRACE for any reclaim
request from the client.
.Pp
For correct operation of the server, it must be ensured that the file
is written to stable storage by the time a write op with IO_SYNC specified
has returned. This might require hardware level caching to be disabled for
a local disk drive that holds the file, or similar.
.Sh FILES
.Bl -tag -width /var/db/nfs-stablerestart -compact
.It Pa /var/db/nfs-stablerestart
NFS V4 stable restart file
.El
.Sh SEE ALSO
.Xr nfsv4 4
.Xr nfsd 8
.Sh BUGS
If the file is empty, the NFS V4 server has no choice but to return
NFSERR_NOGRACE for all Reclaim requests. Although correct, this is
a highly undesirable occurrence, so the file should not be lost if
at all possible. Nfsd will not create the file if it does not
exist and will simply log a failure to start, in the hopes that the
file can be recovered from a backup. To move the file, you must edit
the nfsd sources and recompile it. This was done to discourage
accidental relocation of the file.