Difference between revisions of "NFS"

From ArchWiki
Jump to: navigation, search
m (typo - extra word)
m (NetworkManager dispatcher: use dash instead of underline; consistency with other NetworkManager dispatcher scripts)
 
(279 intermediate revisions by 99 users not shown)
Line 1: Line 1:
[[Category:Networking]]
+
[[Category:File systems]]
[[it:NFSv4]]
+
[[Category:Network sharing]]
[[zh-CN:NFSv4]]
+
[[ar:NFS]]
{{note|this article covers NFSv4, for the older version 3 see [[NFSv3]]}}
+
[[cs:NFS]]
'''Network File System (NFS)''', is an open standard network file sharing protocol.
+
[[de:Network File System]]
 +
[[es:NFS]]
 +
[[fr:NFS]]
 +
[[it:NFS]]
 +
[[ja:NFS]]
 +
[[ru:NFS]]
 +
[[zh-CN:NFS]]
 +
{{Related articles start}}
 +
{{Related|NFS/Troubleshooting}}
 +
{{Related articles end}}
 +
From [[Wikipedia: Network File System|Wikipedia]]:
 +
: ''Network File System (NFS) is a distributed file system protocol originally developed by Sun Microsystems in 1984, allowing a user on a client computer to access files over a network in a manner similar to how local storage is accessed.''
  
==Installing==
+
== Installation ==
Both client and server only require {{Pkg|nfs-utils}} from the [[official repositories]].
+
  
==Configuring==
+
Both client and server only require the [[install|installation]] of the {{Pkg|nfs-utils}} package.
===Time synchronization===
+
In order for NFS to function properly, both server and client must have closely matching time values.  If the clocks on the clients differ from the server too much, then basic functions like file copy operations may hang for a very long time leaving the system unusable until they resume.  The clocks do not have to match to micro/nano second accuracies, but ideally they should be within 1 second of each other.
+
  
The [[NTP]] system is recommended to sync both the server and the clients to the highly accurate NTP servers available on the Internet. For a small system like a home network, the ntpdate utility may be used to sync both servers and clients to the same time. For a larger installation, it may be desirable to install an OpenNTP server (see [[NTP]]) onto the same machine acting as the NFS server, and then all clients on the network would sync time values from the server. This has the advantage of lowering the stress on the external NTP servers, and in assuring that the NFS clients will use the exact time that the NFS server has, even if the NFS server experiences some drift.
+
It is '''highly''' recommended to use a time sync daemon such as {{Pkg|ntp}} on all nodes to keep client/server clocks in sync.  Without accurate clocks on all nodes, NFS can introduce unwanted delays. The [[Network Time Protocol daemon]] is recommended to sync both the server and the clients to the highly accurate NTP servers available on the Internet.
 +
 
 +
{{Note|{{Pkg|nfs-utils}} for Arch Linux ARM starting with update 1.3.2-4 (possibly earlier) has been reported by one user to behave differently from the x86_64 or i686 package. See the discussion page for a recipe for client mounts.}}
 +
 
 +
==Configuration==
  
 
===Server===
 
===Server===
The server configuration involves the {{ic|/etc/idmapd.conf}} file and the {{ic|/etc/exports}} file to export shares. Further tweaking can be done by editing {{ic|/etc/conf.d/nfs-server.conf}}.
 
  
====Server ID mapping====
+
NFS needs to see the list of shares (referred to as "exports" from here on out), which are defined in {{ic|/etc/exports}} in order to serve-up the content.  This can be any directory on the file system. In the interest of security, it is recommended to use an NFS export root which will keep users limited to that mount point onlyThe following example illustrates this concept.
The {{ic|/etc/idmapd.conf}} file needs to be edited. You'll need to at the very least specify your Domain there. Example:
+
  [General]
+
+
Verbosity = 1
+
Pipefs-Directory = /var/lib/nfs/rpc_pipefs
+
'''Domain = archlinux.org'''
+
+
[Mapping]
+
+
Nobody-User = nobody
+
Nobody-Group = nobody
+
  
====Exports====
+
Any NFS shares defined in {{ic|/etc/exports}} are relative to the NFS root.  In this example, the NFS root will be {{ic|/srv/nfs4}} and we are sharing {{ic|/mnt/music}}.
All the NFS shares are defined in {{ic|/etc/exports}}. Add directories which you want to share and ip addresses or hostnames of client machines that will be allowed to mount them:
+
/mnt/music 192.168.0.12(rw,no_subtree_check)
+
  
You can also share it to a whole subnet:
+
  # mkdir -p /srv/nfs4/music /mnt/music
  /mnt/music 192.168.0.0/24(rw,no_subtree_check)
+
{{Note|The old NFSv3-style 192.168.0.*-scheme is ''no longer'' supported.}}
+
  
A typical NFSv4 export would look like this:
+
Read/Write permissions must be set on the music directory so clients may write to it.  
/mnt      192.168.0.12(rw,fsid=0,no_subtree_check)
+
 
  /mnt/music 192.168.0.12(rw,no_subtree_check)
+
Now mount the actual target share, {{ic|/mnt/music}} to the NFS share via the mount command:
{{Note|The {{ic|1=fsid=0}} is required for the root filesystem being exported. {{ic|/mnt}} is the NFS root here (due to the {{ic|1=fsid=0}} entry). Everything else that you want to be shared over NFS must be accessible under {{ic|/mnt}}. Setting an NFS root is required. For exporting directories outside the NFS root, see below.}}
+
 
 +
  # mount --bind /mnt/music /srv/nfs4/music
 +
 
 +
To make it stick across server reboots, add the bind mount to {{ic|fstab}}:
 +
{{hc|/etc/fstab|
 +
/mnt/music /srv/nfs4/music  none  bind  0  0
 +
}}
 +
{{Note|[[ZFS]] filesystems require special handling of bindmounts, see [[ZFS#Bindmount]].}}
 +
 
 +
Add directories to be shared and an ip address or hostname(s) of client machines that will be allowed to mount them in {{ic|exports}}:
 +
{{hc|/etc/exports|<nowiki>
 +
/srv/nfs4/ 192.168.1.0/24(rw,fsid=root,no_subtree_check)
 +
/srv/nfs4/music 192.168.1.0/24(rw,no_subtree_check,nohide) # note the nohide option which is applied to mounted directories on the file system.
 +
</nowiki>}}
 +
 
 +
Users need-not open the share to the entire subnet; one can specify a single IP address or hostname as well.
  
 
For more information about all available options see {{ic|man 5 exports}}.
 
For more information about all available options see {{ic|man 5 exports}}.
  
=====Exporting directories outside your NFS root=====
+
{{Note|Modifying {{ic|/etc/exports}} while the server is running will require a re-export for changes to take effect as noted by the upstream comments in {{ic|/etc/exports}}:
To do this, you will need to use bind mounts. For example, to bind {{ic|/home/john}} to {{ic|/mnt/john}}:
+
  # exportfs -rav
  # mount --bind /home/john /mnt/john
+
}}
To make it stick across server reboots add the bind mount to {{ic|/etc/fstab}}:
+
/home/john    /mnt/john    none    bind  0 0
+
  
Then, {{ic|/mnt/john}} needs to be added to {{ic|/etc/exports}}:
+
==== Starting the server ====
/mnt      192.168.0.12(rw,fsid=0,no_subtree_check)
+
/mnt/music 192.168.0.12(rw,no_subtree_check)
+
/mnt/john  192.168.0.12(rw,no_subtree_check,'''nohide''')
+
The {{ic|nohide}} option is '''required''', because the kernel NFS server automatically hides mounted directories.
+
  
====Starting the server====
+
[[Start]] and [[enable]] {{ic|nfs-server.service}}. The {{ic|rpcbind.service}} is also needed for older V2 and V3 exports. To run a V4-only setup, be sure to explicitly disable V2 and V3 using [https://bbs.archlinux.org/viewtopic.php?id=193629]: {{hc|/etc/conf.d/nfs-server.conf|2=
To start the NFS server:
+
NFSD_OPTS="-N 2 -N 3"
# rc.d start rpcbind nfs-common nfs-server
+
}}
Or add them to your {{ic|/etc/rc.conf}}.
+
otherwise the {{ic|rpcbind.service}} is required.
DAEMONS=(... rpcbind nfs-common nfs-server ...)
+
  
To start the NFS server when you are using systemd, use:
+
==== Miscellaneous ====
# systemctl start nfsd.service rpc-idmapd.service rpc-mountd.service rpcbind.service
+
===== Optional configuration =====
  
===Client===
+
{{ic|/etc/conf.d/nfs-server.conf}} holds optional configurations for options to pass to rpc.nfsd, rpc.mountd, or rpc.svcgssd.  Users setting up a simple configuration may not need to edit this file.
The client configuration only involves the {{ic|/etc/idmapd.conf}} file. If your client also acts as a server for other machines on the network, then you will still have to configure the files covered in the server section.  
+
  
====Client ID mapping====
+
===== Static ports for NFSv3 =====
The {{ic|/etc/idmapd.conf}} file needs to be edited on all clients. '''The Domain entry should be identical to the one on the server''' (see the ''Server ID mapping'' section). Example:
+
[General]
+
+
Verbosity = 1
+
Pipefs-Directory = /var/lib/nfs/rpc_pipefs
+
'''Domain = archlinux.org'''
+
+
[Mapping]
+
+
Nobody-User = nobody
+
Nobody-Group = nobody
+
+
[Translation]
+
Method = nsswitch
+
  
====Starting the client====
+
Users needing support for NFSv3 clients, may wish to consider using static ports. By default, for NFSv3 operation {{ic|rpc.statd}} and {{ic|lockd}} use random ephemeral ports; in order to allow NFSv3 operations through a firewall static ports need to be defined. Edit {{ic|/etc/conf.d/nfs-common.conf}} to set {{ic|STATD_OPTS}}:
To start the NFS client:
+
# rc.d start rpcbind nfs-common
+
Or add them to your {{ic|/etc/rc.conf}}.
+
DAEMONS=(... rpcbind nfs-common ...)
+
  
{{note|On a client only setup make sure rpc.idmapd is running. The nfs-common daemon usually auto-detects whether rpc.idmapd has to be started, but it might fail if there aren't any nfs4 mount entries in {{ic|/etc/fstab}} or if {{ic|/etc/exports}} is empty (which both might be the case if you are using [[autofs]] to mount the nfs4 shares).
+
{{hc|/etc/conf.d/nfs-common.conf|2=
In this case set '''NEED_IDMAPD&#61;&quot;yes&quot;''' in {{ic|/etc/conf.d/nfs-common.conf}}. }}
+
STATD_OPTS="-p 32765 -o 32766 -T 32803"
 +
}}
  
==Mounting NFS shares on the client==
+
The {{ic|rpc.mountd}} should consult {{ic|/etc/services}} and bind to the same static port 20048 under normal operation; however, if it needs to be explicity defined edit {{ic|/etc/conf.d/nfs-server.conf}} to set {{ic|MOUNTD_OPTS}}:
Show the server's exported filesystems:
+
showmount -e server
+
Then just mount as normal:
+
# rc.d start rpcbind nfs-common
+
# mount -t nfs4 server:/ /mnt/server/
+
# mount -t nfs4 server:/music /mnt/music/
+
# mount -t nfs4 server:/john /mnt/john
+
Replacing 'server' with the hostname or IP address of your NFS server and of course 'server', 'music' and 'john' with the names of whatever directories you exported on the server.
+
{{note|The root of the path on the server is the NFS root specified; all paths must be specified relative to it.}}
+
  
===Auto mounting===
+
{{hc|/etc/conf.d/nfs-server.conf|2=
*With Initscripts: If you want to auto mount the NFS shares on boot, you will have to make sure that the network(or any other networking daemon that you use), rpcbind, nfs-common daemons are started up and also in that order. Do NOT background the daemons since the order in which they start up is important. Additionally you will also want to start netfs daemon which handles the clean unmount of NFS shares while shutting down the client machine. The netfs daemon can be backgrounded without any issues.  
+
MOUNTD_OPTS="-p 20048"
DAEMONS=(... network rpcbind nfs-common @netfs ...)
+
}}
  
*With Systemd: [[Systemd#Remote_filesystem_mounts|Systemd/RemoteFilesystem page]] and make sure to enable rpc-idmapd.service for user id mapping.
+
After making these changes, several services need to be restarted; the first writes the configuration options out to {{ic|/run/sysconfig/nfs-utils}} (see {{ic|/usr/lib/systemd/scripts/nfs-utils_env.sh}}), the second restarts {{ic|rpc.statd}} with the new ports, the last reloads {{ic|lockd}} (kernel module) with the new ports. [[Restart]] these services now: {{ic|nfs-config}}, {{ic|rpcbind}}, {{ic|rpc-statd}}, and {{ic|nfs-server}}.
  
After you have added the daemons, auto mounting of NFS shares can be handled in one of two ways:
+
After the restarts, use {{ic|rpcinfo -p}} on the server to examine the static ports are as expected. Using {{ic|rpcinfo -p <server IP>}} from the client should reveal the exact same static ports.
====Using fstab====
+
Using [[fstab]] is useful when you have a server which is always on, and the NFS shares are available whenever your client boots up. Edit your {{ic|/etc/fstab}} file, and add an appropriate line in there reflecting your setup.
+
server:/ /mnt/nfsshare nfs4 defaults 0 0
+
{{note| where ''server'' is the server hostname or IP address}}
+
If you wish to specify a packet size for read and write packets, specify them in your {{ic|/etc/fstab}} entry. Read the NFS man page for further information, including all available mount options.
+
  
====Using autofs====
+
===== NFSv2 compatibility =====
Using [[autofs]] is useful when you have multiple machines that you want to connect via NFS and they could both be clients as well as servers. The reason this method is preferable over the earlier one is that if one of the machine(server) is switched off, the client will not throw errors about being unable to find NFS shares. Please see the relevant section on the [[autofs#NFS_Network_mounts]] page for setting up NFS shares.
+
  
== Mounting from Windows ==
+
Users needing to support clients using NFSv2 (for example U-Boot), should set NFSD_OPTS="-V 2" in /etc/conf.d/nfs-server.conf.
{{note|only the Enterprise and Ultimate versions of Windows 7 include "Client for NFS"}}
+
NFS shares can be mounted from windows if the "Client for NFS" service is actived (which it is not by default).
+
To install the service go to "Programs and features" either through the control panel or by typing it in the search box from the start menu and click on "Turn Windows features on or off". Locate the "Services for NFS" and activate it as well as both subservices ("Administrative tools" and "Client for NFS").
+
  
Some global options can be set by opening the "Services for Network File System" (locate it with the search box) and right clicking on the client->properties.
+
===== Firewall configuration =====
  
{{Warning|under Windows the share is addressed by it's full path on the server, not just the path relative to the nfsroot!! If in doubt run {{ic|showmount -e servername}} from cmd.exe}}
+
To enable access through a firewall, tcp and udp ports 111, 2049, and 20048 need to be opened when using the default configuration; use {{ic|rpcinfo -p}} to examine the exact ports in use on the server. To configure this for [[iptables]], execute this commands:
  
==Troubleshooting==
+
# iptables -A INPUT -p tcp -m tcp --dport 111 -j ACCEPT
===exportfs: /etc/exports:2: syntax error: bad option list===
+
# iptables -A INPUT -p tcp -m tcp --dport 2049 -j ACCEPT
Delete all space from the option list in {{ic|/etc/exports}}
+
# iptables -A INPUT -p tcp -m tcp --dport 20048 -j ACCEPT
 +
# iptables -A INPUT -p udp -m udp --dport 111 -j ACCEPT
 +
# iptables -A INPUT -p udp -m udp --dport 2049 -j ACCEPT
 +
# iptables -A INPUT -p udp -m udp --dport 20048 -j ACCEPT
  
===mount.nfs4: No such device===
+
To have this configuration load on every system start, edit {{ic|/etc/iptables/iptables.rules}} to include the following lines:
Check that you have loaded the {{ic|nfs}} module
+
lsmod | grep nfs
+
and if previous returns empty or only nfsd-stuff, do
+
modprobe nfs
+
  
===mount.nfs4: access denied by server while mounting===
+
{{hc|/etc/iptables/iptables.rules|<nowiki>
Check that the permissions on your client's folder are correct. Try using 755.
+
-A INPUT -p tcp -m tcp --dport 111 -j ACCEPT
 +
-A INPUT -p tcp -m tcp --dport 2049 -j ACCEPT
 +
-A INPUT -p tcp -m tcp --dport 20048 -j ACCEPT
 +
-A INPUT -p udp -m udp --dport 111 -j ACCEPT
 +
-A INPUT -p udp -m udp --dport 2049 -j ACCEPT
 +
-A INPUT -p udp -m udp --dport 20048 -j ACCEPT
 +
</nowiki>}}
  
=== permissions issues ===
+
The previous commands can be saved by executing:
If you find that you cannot set the permissions on files properly, make sure the user/group you are chowning are on both the client and server.
+
# iptables-save > /etc/iptables/iptables.rules
If that does not help, try modifying these lines in {{ic|/etc/conf.d/nfs-common.conf}}
+
{{bc|<nowiki>
+
# /etc/conf.d/nfs-common.conf
+
  
# Do you want to start the statd daemon? It is not needed for NFSv4.
+
{{note|This command will '''override''' the current iptables start configuration with the current iptables configuration!}}
NEED_STATD="no"
+
If using NFSv3 and the above listed static ports for {{ic|rpc.statd}} and {{ic|lockd}} these also need to be added to the configuration:
  
# Do you want to start the idmapd daemon? It is only needed for NFSv4.
+
{{hc|/etc/iptables/iptables.rules|<nowiki>
NEED_IDMAPD="yes"
+
-A INPUT -p tcp -m tcp --dport 32765 -j ACCEPT
 +
-A INPUT -p tcp -m tcp --dport 32803 -j ACCEPT
 +
-A INPUT -p udp -m udp --dport 32765 -j ACCEPT
 +
-A INPUT -p udp -m udp --dport 32803 -j ACCEPT
 
</nowiki>}}
 
</nowiki>}}
Restart the nfs-common daemon for the changes to take effect.
 
I restarted all the other daemons as well, just to be sure.
 
  
=== group/gid permissions issues ===
+
If using V4-only setup, only tcp port 2049 need to be opened. Therefore only one line need.
If NFS shares mount fine, and are fully accessible to the owner, but not to group members; check the number of groups that user belongs to. NFS has a limit of 16 on the number of groups a user can belong to. If you have users with more then this, you need to enable the {{ic|--manage-gids}} start-up flag for {{ic|rpc.mountd}} on the NFS server.
+
 
 +
{{hc|/etc/iptables/iptables.rules|<nowiki>
 +
-A INPUT -p tcp -m tcp --dport 2049 -j ACCEPT
 +
</nowiki>}}
 +
 
 +
To apply changes, [[Restart]] {{ic|iptables.service}}.
 +
 
 +
=== Client ===
 +
 
 +
[[Start]] {{ic|rpcbind.service}},{{ic|nfs-client.target}} and {{ic|remote-fs.target}} and [[enable]] them to start at boot.
 +
 
 +
Users intending to use NFS4 with Kerberos, also need to [[start]] and [[enable]] {{ic|rpc-gssd.service}}. Setting up {{ic|/etc/krb5.keytab /etc/krb5.conf}} are beyond the scope of this article.
 +
==== Error from systemd ====
 +
Users experiencing the following may consider turning off the service using system's masking feature: "Dependency failed for pNFS block layout mapping daemon."
 +
 
 +
Example:
 +
# systemctl mask nfs-blkmap.service
 +
 
 +
==== Manual mounting ====
 +
 
 +
For NFSv3 use this command to show the server's exported file systems:
 +
$ showmount -e servername
 +
 
 +
For NFSv4 mount the root NFS directory and look around for available mounts:
 +
$ mount server:/ /mountpoint/on/client
 +
 
 +
Then mount omitting the server's NFS export root:
 +
# mount -t nfs -o vers=4 servername:/music /mountpoint/on/client
 +
 
 +
If mount fails try including the server's export root (required for Debian/RHEL/SLES, some distributions need {{ic|-t nfs4}} instead of {{ic|-t nfs}}):
 +
# mount -t nfs -o vers=4 servername:/full/path/to/music /mountpoint/on/client
 +
 
 +
{{Note|Server name needs to be a valid hostname (not just IP address). Otherwise mounting of remote share will hang.}}
 +
 
 +
==== Mount using /etc/fstab ====
 +
 
 +
Using [[fstab]] is useful for a server which is always on, and the NFS shares are available whenever the client boots up. Edit {{ic|/etc/fstab}} file, and add an appropriate line reflecting the setup. Again, the server's NFS export root is omitted.
 +
 
 +
{{hc|/etc/fstab|<nowiki>
 +
servername:/music  /mountpoint/on/client  nfs4  rsize=8192,wsize=8192,timeo=14,_netdev 0 0
 +
</nowiki>}}
 +
 
 +
{{Note|Consult the ''NFS'' and ''mount'' man pages for more mount options.}}
 +
 
 +
Some additional mount options to consider are include:
 +
 
 +
; rsize and wsize: The {{ic|rsize}} value is the number of bytes used when reading from the server. The {{ic|wsize}} value is the number of bytes used when writing to the server. The default for both is 1024, but using higher values such as 8192 can improve throughput.  This is not universal.  It is recommended to test after making this change, see [[#Performance tuning]].
 +
 
 +
; timeo: The {{ic|timeo}} value is the amount of time, in tenths of a second, to wait before resending a transmission after an RPC timeout. After the first timeout, the timeout value is doubled for each retry for a maximum of 60 seconds or until a major timeout occurs. If connecting to a slow server or over a busy network, better performance can be achieved by increasing this timeout value.
 +
 
 +
; _netdev: The {{ic|_netdev}} option tells the system to wait until the network is up before trying to mount the share. systemd assumes this for NFS, but anyway it is good practice to use it for all types of networked file systems
 +
 
 +
{{Note|Setting the sixth field (fs_passno) to a nonzero value may lead to unexpected behaviour, e.g. hangs when the systemd automount waits for a check which will never happen.}}
 +
 
 +
==== Mount using /etc/fstab with systemd ====
 +
 
 +
Another method is using the systemd {{ic|automount}} service. This is a better option than {{ic|_netdev}}, because it remounts the network device quickly when the connection is broken and restored. As well, it solves the problem from autofs, see the example below:
 +
 
 +
{{hc|1=/etc/fstab|2=
 +
servername:/home  ''/mountpoint/on/client''  nfs  noauto,x-systemd.automount,x-systemd.device-timeout=10,timeo=14,x-systemd.idle-timeout=1min 0 0 
 +
}}
 +
 
 +
One might have to reboot the client to make systemd aware of the changes to fstab. Alternatively, try [[Systemd#Using_units|reloading]] systemd and restarting {{ic|''mountpoint-on-client''.automount}} to reload the {{ic|/etc/fstab}} configuration.
 +
 
 +
{{Tip|
 +
* The {{ic|noauto}} mount option will not mount the NFS share until it is accessed: use {{ic|auto}} for it to be available immediately. <br> If experiencing any issues with the mount failing due to the network not being up/available, [[enable]] {{ic|NetworkManager-wait-online.service}}. It will ensure that {{ic|network.target}} has all the links available prior to being active.
 +
* The {{ic|users}} mount option would allow user mounts, but be aware it implies further options as {{ic|noexec}} for example.
 +
* The {{ic|<nowiki>x-systemd.idle-timeout=1min</nowiki>}} option will unmount the NFS share automatically after 1 minute of non-use. Good for laptops which might suddenly disconnect from the network.
 +
* If shutdown/reboot holds too long because of NFS,  [[enable]] {{ic|NetworkManager-wait-online.service}} to ensure that NetworkManager is not exited before the NFS volumes are unmounted. }}
 +
 
 +
{{Note|Users trying to automount a NFS-share via systemd which is mounted the same way on the server may experience a freeze when handling larger amounts of data.}}
 +
 
 +
==== Mount using autofs ====
 +
 
 +
Using [[autofs]] is useful when multiple machines want to connect via NFS; they could both be clients as well as servers. The reason this method is preferable over the earlier one is that if the server is switched off, the client will not throw errors about being unable to find NFS shares. See [[autofs#NFS network mounts]] for details.
 +
 
 +
== Tips and tricks ==
 +
 
 +
=== Performance tuning ===
 +
 
 +
In order to get the most out of NFS, it is necessary to tune the {{ic|rsize}} and {{ic|wsize}} mount options to meet the requirements of the network configuration.
 +
 
 +
=== Automounting shares with systemd-networkd ===
 +
 
 +
Users making use of systemd-networkd might notice nfs mounts the fstab are not mounted when booting; errors like the following are common:
 +
mount[311]: mount.nfs4: Network is unreachable
 +
 
 +
The solution is simple; force systemd to wait for the network to be completely configured by [[enabling]] {{ic|systemd-networkd-wait-online.service}}. In theory this slows down the boot-process because less services run in parallel.
 +
 
 +
=== Automatic mount handling ===
 +
 
 +
This trick is useful for laptops that require nfs shares from a local wireless network. If the nfs host becomes unreachable, the nfs share will be unmounted to hopefully prevent system hangs when using the hard mount option. See https://bbs.archlinux.org/viewtopic.php?pid=1260240#p1260240
 +
 
 +
Make sure that the NFS mount points are correctly indicated in {{ic|/etc/fstab}}:
 +
 
 +
{{hc|$ cat /etc/fstab|<nowiki>
 +
lithium:/mnt/data          /mnt/data         nfs noauto,noatime,rsize=32768,wsize=32768 0 0
 +
lithium:/var/cache/pacman  /var/cache/pacman nfs noauto,noatime,rsize=32768,wsize=32768 0 0</nowiki>
 +
}}
 +
 
 +
{{Note|You must use hostnames in {{ic|/etc/fstab}} for this to work, not IP addresses.}}
 +
 
 +
The {{ic|noauto}} mount option tells systemd not to automatically mount the shares at boot. systemd would otherwise attempt to mount the nfs shares that may or may not exist on the network causing the boot process to appear to stall on a blank screen.
 +
 
 +
In order to mount NFS shares with non-root users the {{ic|user}} option has to be added.
 +
 
 +
Create the {{ic|auto_share}} script that will be used by ''cron'' or ''systemd/Timers'' to use ICMP ping to check if the NFS host is reachable:
 +
 
 +
{{hc|/usr/local/bin/auto_share|<nowiki>
 +
#!/bin/bash
 +
 
 +
function net_umount {
 +
  umount -l -f $1 &>/dev/null
 +
}
 +
 
 +
function net_mount {
 +
  mountpoint -q $1 || mount $1
 +
}
 +
 
 +
NET_MOUNTS=$(sed -e '/^.*#/d' -e '/^.*:/!d' -e 's/\t/ /g' /etc/fstab | tr -s " ")$'\n'b
 +
 
 +
printf %s "$NET_MOUNTS" | while IFS= read -r line
 +
do
 +
  SERVER=$(echo $line | cut -f1 -d":")
 +
  MOUNT_POINT=$(echo $line | cut -f2 -d" ")
 +
 
 +
  # Check if server already tested
 +
  if [[ "${server_ok[@]}" =~ "${SERVER}" ]]; then
 +
    # The server is up, make sure the share are mounted
 +
    net_mount $MOUNT_POINT
 +
  elif [[ "${server_notok[@]}" =~ "${SERVER}" ]]; then
 +
    # The server could not be reached, unmount the share
 +
    net_umount $MOUNT_POINT
 +
  else
 +
    # Check if the server is reachable
 +
    ping -c 1 "${SERVER}" &>/dev/null
 +
 
 +
    if [ $? -ne 0 ]; then
 +
      server_notok[${#Unix[@]}]=$SERVER
 +
      # The server could not be reached, unmount the share
 +
      net_umount $MOUNT_POINT
 +
    else
 +
      server_ok[${#Unix[@]}]=$SERVER
 +
      # The server is up, make sure the share are mounted
 +
      net_mount $MOUNT_POINT
 +
    fi
 +
  fi
 +
done
 +
</nowiki>}}
 +
 
 +
{{Note|If you want to test using a TCP probe instead of ICMP ping (default is tcp port 2049 in NFS4) then replace the line:
 +
  # Check if the server is reachable
 +
  ping -c 1 "${SERVER}" &>/dev/null
 +
 
 +
with:
 +
 
 +
  # Check if the server is reachable
 +
  timeout 1 bash -c ": < /dev/tcp/${SERVER}/2049"
 +
 
 +
in the {{ic|auto_share}} script above.}}
 +
 
 +
# chmod +x /usr/local/bin/auto_share
 +
 
 +
Create a cron entry or a systemd/Timers timer to check every minute if the server of the shares are reachable.
 +
 
 +
==== Cron ====
 +
 
 +
{{hc|# crontab -e|<nowiki>
 +
* * * * * /usr/local/bin/auto_share
 +
</nowiki>}}
 +
 
 +
==== systemd/Timers ====
 +
 
 +
{{hc|# /etc/systemd/system/auto_share.timer|<nowiki>
 +
[Unit]
 +
Description=Check the network mounts
 +
 
 +
[Timer]
 +
OnCalendar=*-*-* *:*:00
 +
 
 +
[Install]
 +
WantedBy=timer.target
 +
</nowiki>}}
 +
 
 +
{{hc|# /etc/systemd/system/auto_share.service|<nowiki>
 +
[Unit]
 +
Description=Check the network mounts
 +
 
 +
[Service]
 +
Type=simple
 +
ExecStart=/usr/local/bin/auto_share
 +
</nowiki>}}
 +
 
 +
# systemctl enable auto_share.timer
 +
 
 +
==== Mount at startup via systemd ====
 +
 
 +
A systemd unit file can also be used to mount the NFS shares at startup. The unit file is not necessary if NetworkManager is installed and configured on the client system. See [[#NetworkManager dispatcher]].
 +
 
 +
{{hc|/etc/systemd/system/auto_share.service|<nowiki>
 +
[Unit]
 +
Description=NFS automount
 +
After=syslog.target network.target
 +
 
 +
[Service]
 +
Type=oneshot
 +
RemainAfterExit=yes
 +
ExecStart=/usr/local/bin/auto_share
 +
 
 +
[Install]
 +
WantedBy=multi-user.target
 +
</nowiki>}}
 +
 
 +
Now [[enable]] the {{ic|auto_share.service}}.
 +
 
 +
==== NetworkManager dispatcher ====
 +
 
 +
In addition to the method described previously, [[NetworkManager#Network_services_with_NetworkManager_dispatcher|NetworkManager]] can also be configured to run a script on network status change: [[Enable]] and [[start]] the {{ic|NetworkManager-dispatcher.service}}.
 +
 
 +
The easiest method for mount shares on network status change is to just symlink to the {{ic|auto_share}} script:
 +
 
 +
# ln -s /usr/local/bin/auto_share /etc/NetworkManager/dispatcher.d/30-nfs.sh
 +
 
 +
However, in that particular case unmounting will happen only after the network connection has already been disabled, which is unclean and may result in effects like freezing of KDE Plasma applets.
 +
 
 +
The following script safely unmounts the NFS shares before the relevant network connection is disabled by listening for the {{ic|pre-down}} and {{ic|vpn-pre-down}} events:
 +
 
 +
{{Note|This script ignores mounts with the noauto option.}}
 +
 
 +
{{hc|/etc/NetworkManager/dispatcher.d/30-nfs.sh|<nowiki>
 +
#!/bin/bash
 +
 
 +
# Find the connection UUID with "nmcli con show" in terminal.
 +
# All NetworkManager connection types are supported: wireless, VPN, wired...
 +
WANTED_CON_UUID="CHANGE-ME-NOW-9c7eff15-010a-4b1c-a786-9b4efa218ba9"
 +
 
 +
if [[ "$CONNECTION_UUID" == "$WANTED_CON_UUID" ]]; then
 +
   
 +
    # Script parameter $1: NetworkManager connection name, not used
 +
    # Script parameter $2: dispatched event
 +
   
 +
    case "$2" in
 +
        "up")
 +
            mount -a -t nfs4,nfs
 +
            ;;
 +
        "pre-down");&
 +
        "vpn-pre-down")
 +
            umount -l -a -t nfs4,nfs >/dev/null
 +
            ;;
 +
    esac
 +
fi
 +
</nowiki>}}
 +
 
 +
Make the script executable with [[chmod]] and create a symlink inside {{ic|/etc/NetworkManager/dispatcher.d/pre-down}} to catch the {{ic|pre-down}} events:
 +
# ln -s /etc/NetworkManager/dispatcher.d/30-nfs.sh /etc/NetworkManager/dispatcher.d/pre-down.d/30-nfs.sh
 +
 
 +
The above script can be modified to mount different shares (even other than NFS) for different connections.
 +
 
 +
See also: [[NetworkManager#Use dispatcher to handle mounting of CIFS shares]].
 +
 
 +
== Troubleshooting ==
 +
 
 +
There is a dedicated article [[NFS Troubleshooting]].
 +
 
 +
== See also ==
  
/etc/conf.d/nfs-server.conf
+
* See also [[Avahi]], a Zeroconf implementation which allows automatic discovery of NFS shares.
+
* HOWTO: [[Diskless network boot NFS root]]
# Options for rpc.mountd.
+
* [http://publib.boulder.ibm.com/infocenter/pseries/v5r3/index.jsp?topic=/com.ibm.aix.prftungd/doc/prftungd/nfs_perf.htm NFS Performance Management]
# If you have a port-based firewall, you might want to set up
+
* [http://blogs.msdn.com/sfu/archive/2008/04/14/all-well-almost-about-client-for-nfs-configuration-and-performance.aspx Microsoft Services for Unix NFS Client info]
# a fixed port here using the --port option.
+
* [https://blogs.oracle.com/jag/entry/nfs_on_snow_leopard NFS on Snow Leopard]
# See rpc.mountd(8) for more details.
+
+
MOUNTD_OPTS="--manage-gids"
+

Latest revision as of 16:40, 19 May 2016

Related articles

From Wikipedia:

Network File System (NFS) is a distributed file system protocol originally developed by Sun Microsystems in 1984, allowing a user on a client computer to access files over a network in a manner similar to how local storage is accessed.

Installation

Both client and server only require the installation of the nfs-utils package.

It is highly recommended to use a time sync daemon such as ntp on all nodes to keep client/server clocks in sync. Without accurate clocks on all nodes, NFS can introduce unwanted delays. The Network Time Protocol daemon is recommended to sync both the server and the clients to the highly accurate NTP servers available on the Internet.

Note: nfs-utils for Arch Linux ARM starting with update 1.3.2-4 (possibly earlier) has been reported by one user to behave differently from the x86_64 or i686 package. See the discussion page for a recipe for client mounts.

Configuration

Server

NFS needs to see the list of shares (referred to as "exports" from here on out), which are defined in /etc/exports in order to serve-up the content. This can be any directory on the file system. In the interest of security, it is recommended to use an NFS export root which will keep users limited to that mount point only. The following example illustrates this concept.

Any NFS shares defined in /etc/exports are relative to the NFS root. In this example, the NFS root will be /srv/nfs4 and we are sharing /mnt/music.

# mkdir -p /srv/nfs4/music /mnt/music

Read/Write permissions must be set on the music directory so clients may write to it.

Now mount the actual target share, /mnt/music to the NFS share via the mount command:

# mount --bind /mnt/music /srv/nfs4/music

To make it stick across server reboots, add the bind mount to fstab:

/etc/fstab
/mnt/music /srv/nfs4/music  none   bind   0   0
Note: ZFS filesystems require special handling of bindmounts, see ZFS#Bindmount.

Add directories to be shared and an ip address or hostname(s) of client machines that will be allowed to mount them in exports:

/etc/exports
/srv/nfs4/ 192.168.1.0/24(rw,fsid=root,no_subtree_check)
/srv/nfs4/music 192.168.1.0/24(rw,no_subtree_check,nohide) # note the nohide option which is applied to mounted directories on the file system.

Users need-not open the share to the entire subnet; one can specify a single IP address or hostname as well.

For more information about all available options see man 5 exports.

Note: Modifying /etc/exports while the server is running will require a re-export for changes to take effect as noted by the upstream comments in /etc/exports:
# exportfs -rav

Starting the server

Start and enable nfs-server.service. The rpcbind.service is also needed for older V2 and V3 exports. To run a V4-only setup, be sure to explicitly disable V2 and V3 using [1]:
/etc/conf.d/nfs-server.conf
NFSD_OPTS="-N 2 -N 3"

otherwise the rpcbind.service is required.

Miscellaneous

Optional configuration

/etc/conf.d/nfs-server.conf holds optional configurations for options to pass to rpc.nfsd, rpc.mountd, or rpc.svcgssd. Users setting up a simple configuration may not need to edit this file.

Static ports for NFSv3

Users needing support for NFSv3 clients, may wish to consider using static ports. By default, for NFSv3 operation rpc.statd and lockd use random ephemeral ports; in order to allow NFSv3 operations through a firewall static ports need to be defined. Edit /etc/conf.d/nfs-common.conf to set STATD_OPTS:

/etc/conf.d/nfs-common.conf
STATD_OPTS="-p 32765 -o 32766 -T 32803"

The rpc.mountd should consult /etc/services and bind to the same static port 20048 under normal operation; however, if it needs to be explicity defined edit /etc/conf.d/nfs-server.conf to set MOUNTD_OPTS:

/etc/conf.d/nfs-server.conf
MOUNTD_OPTS="-p 20048"

After making these changes, several services need to be restarted; the first writes the configuration options out to /run/sysconfig/nfs-utils (see /usr/lib/systemd/scripts/nfs-utils_env.sh), the second restarts rpc.statd with the new ports, the last reloads lockd (kernel module) with the new ports. Restart these services now: nfs-config, rpcbind, rpc-statd, and nfs-server.

After the restarts, use rpcinfo -p on the server to examine the static ports are as expected. Using rpcinfo -p <server IP> from the client should reveal the exact same static ports.

NFSv2 compatibility

Users needing to support clients using NFSv2 (for example U-Boot), should set NFSD_OPTS="-V 2" in /etc/conf.d/nfs-server.conf.

Firewall configuration

To enable access through a firewall, tcp and udp ports 111, 2049, and 20048 need to be opened when using the default configuration; use rpcinfo -p to examine the exact ports in use on the server. To configure this for iptables, execute this commands:

# iptables -A INPUT -p tcp -m tcp --dport 111 -j ACCEPT
# iptables -A INPUT -p tcp -m tcp --dport 2049 -j ACCEPT
# iptables -A INPUT -p tcp -m tcp --dport 20048 -j ACCEPT
# iptables -A INPUT -p udp -m udp --dport 111 -j ACCEPT
# iptables -A INPUT -p udp -m udp --dport 2049 -j ACCEPT
# iptables -A INPUT -p udp -m udp --dport 20048 -j ACCEPT

To have this configuration load on every system start, edit /etc/iptables/iptables.rules to include the following lines:

/etc/iptables/iptables.rules
-A INPUT -p tcp -m tcp --dport 111 -j ACCEPT
-A INPUT -p tcp -m tcp --dport 2049 -j ACCEPT
-A INPUT -p tcp -m tcp --dport 20048 -j ACCEPT
-A INPUT -p udp -m udp --dport 111 -j ACCEPT
-A INPUT -p udp -m udp --dport 2049 -j ACCEPT
-A INPUT -p udp -m udp --dport 20048 -j ACCEPT

The previous commands can be saved by executing:

# iptables-save > /etc/iptables/iptables.rules
Note: This command will override the current iptables start configuration with the current iptables configuration!

If using NFSv3 and the above listed static ports for rpc.statd and lockd these also need to be added to the configuration:

/etc/iptables/iptables.rules
-A INPUT -p tcp -m tcp --dport 32765 -j ACCEPT
-A INPUT -p tcp -m tcp --dport 32803 -j ACCEPT
-A INPUT -p udp -m udp --dport 32765 -j ACCEPT
-A INPUT -p udp -m udp --dport 32803 -j ACCEPT

If using V4-only setup, only tcp port 2049 need to be opened. Therefore only one line need.

/etc/iptables/iptables.rules
-A INPUT -p tcp -m tcp --dport 2049 -j ACCEPT

To apply changes, Restart iptables.service.

Client

Start rpcbind.service,nfs-client.target and remote-fs.target and enable them to start at boot.

Users intending to use NFS4 with Kerberos, also need to start and enable rpc-gssd.service. Setting up /etc/krb5.keytab /etc/krb5.conf are beyond the scope of this article.

Error from systemd

Users experiencing the following may consider turning off the service using system's masking feature: "Dependency failed for pNFS block layout mapping daemon."

Example:

# systemctl mask nfs-blkmap.service

Manual mounting

For NFSv3 use this command to show the server's exported file systems:

$ showmount -e servername

For NFSv4 mount the root NFS directory and look around for available mounts:

$ mount server:/ /mountpoint/on/client

Then mount omitting the server's NFS export root:

# mount -t nfs -o vers=4 servername:/music /mountpoint/on/client

If mount fails try including the server's export root (required for Debian/RHEL/SLES, some distributions need -t nfs4 instead of -t nfs):

# mount -t nfs -o vers=4 servername:/full/path/to/music /mountpoint/on/client
Note: Server name needs to be a valid hostname (not just IP address). Otherwise mounting of remote share will hang.

Mount using /etc/fstab

Using fstab is useful for a server which is always on, and the NFS shares are available whenever the client boots up. Edit /etc/fstab file, and add an appropriate line reflecting the setup. Again, the server's NFS export root is omitted.

/etc/fstab
servername:/music   /mountpoint/on/client   nfs4   rsize=8192,wsize=8192,timeo=14,_netdev	0 0
Note: Consult the NFS and mount man pages for more mount options.

Some additional mount options to consider are include:

rsize and wsize
The rsize value is the number of bytes used when reading from the server. The wsize value is the number of bytes used when writing to the server. The default for both is 1024, but using higher values such as 8192 can improve throughput. This is not universal. It is recommended to test after making this change, see #Performance tuning.
timeo
The timeo value is the amount of time, in tenths of a second, to wait before resending a transmission after an RPC timeout. After the first timeout, the timeout value is doubled for each retry for a maximum of 60 seconds or until a major timeout occurs. If connecting to a slow server or over a busy network, better performance can be achieved by increasing this timeout value.
_netdev
The _netdev option tells the system to wait until the network is up before trying to mount the share. systemd assumes this for NFS, but anyway it is good practice to use it for all types of networked file systems
Note: Setting the sixth field (fs_passno) to a nonzero value may lead to unexpected behaviour, e.g. hangs when the systemd automount waits for a check which will never happen.

Mount using /etc/fstab with systemd

Another method is using the systemd automount service. This is a better option than _netdev, because it remounts the network device quickly when the connection is broken and restored. As well, it solves the problem from autofs, see the example below:

/etc/fstab
servername:/home   /mountpoint/on/client  nfs  noauto,x-systemd.automount,x-systemd.device-timeout=10,timeo=14,x-systemd.idle-timeout=1min 0 0

One might have to reboot the client to make systemd aware of the changes to fstab. Alternatively, try reloading systemd and restarting mountpoint-on-client.automount to reload the /etc/fstab configuration.

Tip:
  • The noauto mount option will not mount the NFS share until it is accessed: use auto for it to be available immediately.
    If experiencing any issues with the mount failing due to the network not being up/available, enable NetworkManager-wait-online.service. It will ensure that network.target has all the links available prior to being active.
  • The users mount option would allow user mounts, but be aware it implies further options as noexec for example.
  • The x-systemd.idle-timeout=1min option will unmount the NFS share automatically after 1 minute of non-use. Good for laptops which might suddenly disconnect from the network.
  • If shutdown/reboot holds too long because of NFS, enable NetworkManager-wait-online.service to ensure that NetworkManager is not exited before the NFS volumes are unmounted.
Note: Users trying to automount a NFS-share via systemd which is mounted the same way on the server may experience a freeze when handling larger amounts of data.

Mount using autofs

Using autofs is useful when multiple machines want to connect via NFS; they could both be clients as well as servers. The reason this method is preferable over the earlier one is that if the server is switched off, the client will not throw errors about being unable to find NFS shares. See autofs#NFS network mounts for details.

Tips and tricks

Performance tuning

In order to get the most out of NFS, it is necessary to tune the rsize and wsize mount options to meet the requirements of the network configuration.

Automounting shares with systemd-networkd

Users making use of systemd-networkd might notice nfs mounts the fstab are not mounted when booting; errors like the following are common:

mount[311]: mount.nfs4: Network is unreachable

The solution is simple; force systemd to wait for the network to be completely configured by enabling systemd-networkd-wait-online.service. In theory this slows down the boot-process because less services run in parallel.

Automatic mount handling

This trick is useful for laptops that require nfs shares from a local wireless network. If the nfs host becomes unreachable, the nfs share will be unmounted to hopefully prevent system hangs when using the hard mount option. See https://bbs.archlinux.org/viewtopic.php?pid=1260240#p1260240

Make sure that the NFS mount points are correctly indicated in /etc/fstab:

$ cat /etc/fstab
lithium:/mnt/data           /mnt/data	        nfs noauto,noatime,rsize=32768,wsize=32768 0 0
lithium:/var/cache/pacman   /var/cache/pacman	nfs noauto,noatime,rsize=32768,wsize=32768 0 0
Note: You must use hostnames in /etc/fstab for this to work, not IP addresses.

The noauto mount option tells systemd not to automatically mount the shares at boot. systemd would otherwise attempt to mount the nfs shares that may or may not exist on the network causing the boot process to appear to stall on a blank screen.

In order to mount NFS shares with non-root users the user option has to be added.

Create the auto_share script that will be used by cron or systemd/Timers to use ICMP ping to check if the NFS host is reachable:

/usr/local/bin/auto_share
#!/bin/bash

function net_umount {
  umount -l -f $1 &>/dev/null
}

function net_mount {
  mountpoint -q $1 || mount $1
}

NET_MOUNTS=$(sed -e '/^.*#/d' -e '/^.*:/!d' -e 's/\t/ /g' /etc/fstab | tr -s " ")$'\n'b

printf %s "$NET_MOUNTS" | while IFS= read -r line
do
  SERVER=$(echo $line | cut -f1 -d":")
  MOUNT_POINT=$(echo $line | cut -f2 -d" ")

  # Check if server already tested
  if [[ "${server_ok[@]}" =~ "${SERVER}" ]]; then
    # The server is up, make sure the share are mounted
    net_mount $MOUNT_POINT
  elif [[ "${server_notok[@]}" =~ "${SERVER}" ]]; then
    # The server could not be reached, unmount the share
    net_umount $MOUNT_POINT
  else
    # Check if the server is reachable
    ping -c 1 "${SERVER}" &>/dev/null

    if [ $? -ne 0 ]; then
      server_notok[${#Unix[@]}]=$SERVER
      # The server could not be reached, unmount the share
      net_umount $MOUNT_POINT
    else
      server_ok[${#Unix[@]}]=$SERVER
      # The server is up, make sure the share are mounted
      net_mount $MOUNT_POINT
    fi
  fi
done
Note: If you want to test using a TCP probe instead of ICMP ping (default is tcp port 2049 in NFS4) then replace the line:
 # Check if the server is reachable
 ping -c 1 "${SERVER}" &>/dev/null

with:

 # Check if the server is reachable
 timeout 1 bash -c ": < /dev/tcp/${SERVER}/2049"
in the auto_share script above.
# chmod +x /usr/local/bin/auto_share

Create a cron entry or a systemd/Timers timer to check every minute if the server of the shares are reachable.

Cron

# crontab -e
* * * * * /usr/local/bin/auto_share

systemd/Timers

# /etc/systemd/system/auto_share.timer
[Unit]
Description=Check the network mounts

[Timer]
OnCalendar=*-*-* *:*:00

[Install]
WantedBy=timer.target
# /etc/systemd/system/auto_share.service
[Unit]
Description=Check the network mounts

[Service]
Type=simple
ExecStart=/usr/local/bin/auto_share
# systemctl enable auto_share.timer

Mount at startup via systemd

A systemd unit file can also be used to mount the NFS shares at startup. The unit file is not necessary if NetworkManager is installed and configured on the client system. See #NetworkManager dispatcher.

/etc/systemd/system/auto_share.service
[Unit]
Description=NFS automount
After=syslog.target network.target

[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/usr/local/bin/auto_share

[Install]
WantedBy=multi-user.target

Now enable the auto_share.service.

NetworkManager dispatcher

In addition to the method described previously, NetworkManager can also be configured to run a script on network status change: Enable and start the NetworkManager-dispatcher.service.

The easiest method for mount shares on network status change is to just symlink to the auto_share script:

# ln -s /usr/local/bin/auto_share /etc/NetworkManager/dispatcher.d/30-nfs.sh

However, in that particular case unmounting will happen only after the network connection has already been disabled, which is unclean and may result in effects like freezing of KDE Plasma applets.

The following script safely unmounts the NFS shares before the relevant network connection is disabled by listening for the pre-down and vpn-pre-down events:

Note: This script ignores mounts with the noauto option.
/etc/NetworkManager/dispatcher.d/30-nfs.sh
#!/bin/bash

# Find the connection UUID with "nmcli con show" in terminal.
# All NetworkManager connection types are supported: wireless, VPN, wired...
WANTED_CON_UUID="CHANGE-ME-NOW-9c7eff15-010a-4b1c-a786-9b4efa218ba9"

if [[ "$CONNECTION_UUID" == "$WANTED_CON_UUID" ]]; then
    
    # Script parameter $1: NetworkManager connection name, not used
    # Script parameter $2: dispatched event
    
    case "$2" in
        "up")
            mount -a -t nfs4,nfs 
            ;;
        "pre-down");&
        "vpn-pre-down")
            umount -l -a -t nfs4,nfs >/dev/null
            ;;
    esac
fi

Make the script executable with chmod and create a symlink inside /etc/NetworkManager/dispatcher.d/pre-down to catch the pre-down events:

# ln -s /etc/NetworkManager/dispatcher.d/30-nfs.sh /etc/NetworkManager/dispatcher.d/pre-down.d/30-nfs.sh

The above script can be modified to mount different shares (even other than NFS) for different connections.

See also: NetworkManager#Use dispatcher to handle mounting of CIFS shares.

Troubleshooting

There is a dedicated article NFS Troubleshooting.

See also