.\"@(#)sm-notify.8"
.\"
.\" Copyright (C) 2004 Olaf Kirch <okir@suse.de>
.\"
.\" Rewritten by Chuck Lever <chuck.lever@oracle.com>, 2009.
.\" Copyright 2009 Oracle.  All rights reserved.
.\"
.TH SM-NOTIFY 8 "1 November 2009
.SH NAME
sm-notify \- send reboot notifications to NFS peers
.SH SYNOPSIS
.BI "/usr/sbin/sm-notify [-dfn] [-m " minutes "] [-v " name "] [-p " notify-port "] [-P " path "]
.SH DESCRIPTION
File locks are not part of persistent file system state.
Lock state is thus lost when a host reboots.
.PP
Network file systems must also detect when lock state is lost
because a remote host has rebooted.
After an NFS client reboots, an NFS server must release all file locks
held by applications that were running on that client.
After a server reboots, a client must remind the
server of file locks held by applications running on that client.
.PP
For NFS version 2 and version 3, the
.I Network Status Monitor
protocol (or NSM for short)
is used to notify NFS peers of reboots.
On Linux, two separate user-space components constitute the NSM service:
.TP
.B sm-notify
A helper program that notifies NFS peers after the local system reboots
.TP
.B rpc.statd
A daemon that listens for reboot notifications from other hosts, and
manages the list of hosts to be notified when the local system reboots
.PP
The local NFS lock manager alerts its local
.B rpc.statd
of each remote peer that should be monitored.
When the local system reboots, the
.B sm-notify
command notifies the NSM service on monitored peers of the reboot.
When a remote reboots, that peer notifies the local
.BR rpc.statd ,
which in turn passes the reboot notification
back to the local NFS lock manager.
.SH NSM OPERATION IN DETAIL
The first file locking interaction between an NFS client and server causes
the NFS lock managers on both peers to contact their local NSM service to
store information about the opposite peer.
On Linux, the local lock manager contacts
.BR rpc.statd .
.PP
.B rpc.statd
records information about each monitored NFS peer on persistent storage.
This information describes how to contact a remote peer
in case the local system reboots,
how to recognize which monitored peer is reporting a reboot,
and how to notify the local lock manager when a monitored peer
indicates it has rebooted.
.PP
An NFS client sends a hostname, known as the client's
.IR caller_name ,
in each file lock request.
An NFS server can use this hostname to send asynchronous GRANT
calls to a client, or to notify the client it has rebooted.
.PP
The Linux NFS server can provide the client's
.I caller_name
or the client's network address to
.BR rpc.statd .
For the purposes of the NSM protocol,
this name or address is known as the monitored peer's
.IR mon_name .
In addition, the local lock manager tells
.B rpc.statd
what it thinks its own hostname is.
For the purposes of the NSM protocol,
this hostname is known as
.IR my_name .
.PP
There is no equivalent interaction between an NFS server and a client
to inform the client of the server's
.IR caller_name .
Therefore NFS clients do not actually know what
.I mon_name
an NFS server might use in an SM_NOTIFY request.
The Linux NFS client records the server's hostname used on the mount command
to identify rebooting NFS servers.
.SS Reboot notification
When the local system reboots, the
.B sm-notify
command reads the list of monitored peers from persistent storage and
sends an SM_NOTIFY request to the NSM service on each listed remote peer.
It uses the
.I mon_name
string as the destination.
To identify which host has rebooted, the
.B sm-notify
command normally sends the results of
.BR gethostname (3)
as the
.I my_name
string.
The remote
.B rpc.statd
matches incoming SM_NOTIFY requests using this string,
or the caller's network address,
to one or more peers on its own monitor list.
.PP
If
.B rpc.statd
does not find a peer on its monitor list that matches
an incoming SM_NOTIFY request,
the notification is not forwarded to the local lock manager.
In addition, each peer has its own
.IR "NSM state number" ,
a 32-bit integer that is bumped after each reboot by the
.B sm-notify
command.
.B rpc.statd
uses this number to distinguish between actual reboots
and replayed notifications.
.PP
Part of NFS lock recovery is rediscovering
which peers need to be monitored again.
The
.B sm-notify
command clears the monitor list on persistent storage after each reboot.
.SH OPTIONS
.TP
.B -d
Keeps
.B sm-notify
attached to its controlling terminal and running in the foreground
so that notification progress may be monitored directly.
.TP
.B -f
Send notifications even if
.B sm-notify
has already run since the last system reboot.
.TP
.BI -m " retry-time
Specifies the length of time, in minutes, to continue retrying
notifications to unresponsive hosts.
If this option is not specified,
.B sm-notify
attempts to send notifications for 15 minutes.
Specifying a value of 0 causes
.B sm-notify
to continue sending notifications to unresponsive peers
until it is manually killed.
.IP
Notifications are retried if sending fails,
the remote does not respond,
the remote's NSM service is not registered,
or if there is a DNS failure
which prevents the remote's
.I mon_name
from being resolved to an address.
.IP
Hosts are not removed from the notification list until a valid
reply has been received.
However, the SM_NOTIFY procedure has a void result.
There is no way for
.B sm-notify
to tell if the remote recognized the sender and has started
appropriate lock recovery.
.TP
.B -n
Prevents
.B sm-notify
from updating the local system's NSM state number.
.TP
.BI -p " port
Specifies the source port number
.B sm-notify
should use when sending reboot notifications.
If this option is not specified, a randomly chosen ephemeral port is used.
.IP
This option can be used to traverse a firewall between client and server.
.TP
.BI "\-P, " "" \-\-state\-directory\-path " pathname
Specifies the pathname of the parent directory
where NSM state information resides.
If this option is not specified,
.B sm-notify
uses
.I /var/lib/nfs
by default.
.IP
After starting,
.B sm-notify
attempts to set its effective UID and GID to the owner
and group of this directory.
.TP
.BI -v " ipaddr " | " hostname
Specifies the network address from which to send reboot notifications,
and the
.I mon_name
argument to use when sending SM_NOTIFY requests.
If this option is not specified,
.B sm-notify
uses a wildcard address as the transport bind address,
and uses the results of
.BR gethostname (3)
as the
.I mon_name
argument.
.IP
The
.I ipaddr
form can be expressed as either an IPv4 or an IPv6 presentation address.
.IP
This option can be useful in multi-homed configurations where
the remote requires notification from a specific network address.
.SH SECURITY
The
.B sm-notify
command must be started as root to acquire privileges needed
to access the state information database.
It drops root privileges
as soon as it starts up to reduce the risk of a privilege escalation attack.
.PP
During normal operation,
the effective user ID it chooses is the owner of the state directory.
This allows it to continue to access files in that directory after it
has dropped its root privileges.
To control which user ID
.B rpc.statd
chooses, simply use
.BR chown (1)
to set the owner of
the state directory.
.SH ADDITIONAL NOTES
Lock recovery after a reboot is critical to maintaining data integrity
and preventing unnecessary application hangs.
.PP
To help
.B rpc.statd
match SM_NOTIFY requests to NLM requests, a number of best practices
should be observed, including:
.IP
The UTS nodename of your systems should match the DNS names that NFS
peers use to contact them
.IP
The UTS nodenames of your systems should always be fully qualified domain names
.IP
The forward and reverse DNS mapping of the UTS nodenames should be
consistent
.IP
The hostname the client uses to mount the server should match the server's
.I mon_name
in SM_NOTIFY requests it sends
.IP
The use of network addresses as a
.I mon_name
or a
.I my_name
string should be avoided when
interoperating with non-Linux NFS implementations.
.PP
Unmounting an NFS file system does not necessarily stop
either the NFS client or server from monitoring each other.
Both may continue monitoring each other for a time in case subsequent
NFS traffic between the two results in fresh mounts and additional
file locking.
.PP
On Linux, if the
.B lockd
kernel module is unloaded during normal operation,
all remote NFS peers are unmonitored.
This can happen on an NFS client, for example,
if an automounter removes all NFS mount
points due to inactivity.
.SS IPv6 and TI-RPC support
TI-RPC is a pre-requisite for supporting NFS on IPv6.
If TI-RPC support is built into the
.B sm-notify
command ,it will choose an appropriate IPv4 or IPv6 transport
based on the network address returned by DNS for each remote peer.
It should be fully compatible with remote systems
that do not support TI-RPC or IPv6.
.PP
Currently, the
.B sm-notify
command supports sending notification only via datagram transport protocols.
.SH FILES
.TP 2.5i
.I /var/lib/nfs/sm
directory containing monitor list
.TP 2.5i
.I /var/lib/nfs/sm.bak
directory containing notify list
.TP 2.5i
.I /var/lib/nfs/state
NSM state number for this host
.TP 2.5i
.I /proc/sys/fs/nfs/nsm_local_state
kernel's copy of the NSM state number
.SH SEE ALSO
.BR rpc.statd (8),
.BR nfs (5),
.BR uname (2),
.BR hostname (7)
.PP
RFC 1094 - "NFS: Network File System Protocol Specification"
.br
RFC 1813 - "NFS Version 3 Protocol Specification"
.br
OpenGroup Protocols for Interworking: XNFS, Version 3W - Chapter 11
.SH AUTHORS
Olaf Kirch <okir@suse.de>
.br
Chuck Lever <chuck.lever@oracle.com>