autofs mounts

autofs is a nice traditional way to mount the nfs directories on-demand on the nfs clients. was originally used to mount the user home directories to the clients to make them sort of roaming user profiles, but this doesnt stop us from using it to configure nfs mounts in other ways. using autofs saves unnecessary resource allocation to mount the resources on a client machine, which may never be accessed in that session. it also saves the open handles on the nfs server, which is another plus.

ps: the examples below have been done on a centos 8.3 gnu/linux machine. adjust the path locations accordingly as per your flavor of gnu/linux. rest of the options should be good.

install autofs

dnf -y install autofs

define the mounts in the master configuration file named auto.master in /etc

directory listing for auto.master in /etc in a CentOS 8.3 machine

explanation with example

nfs server ip:
nfs export: /nfs-mount-00

client ip:
client directory where we need to mount the nfs share: /nfs-mount-client

step 1: add the map file path in auto.master

the text i added to the auto.master has been bolded for easy reading below

[root@centos-c ~]# cat /etc/auto.master
 # Sample auto.master file
 # This is a 'master' automounter map and it has the following format:
 # mount-point [map-type[,format]:]map [options]
 # For details of the format look at auto.master(5).
 /misc   /etc/auto.misc
 /nfs-mount-client  /etc/auto.nfs-mount-client  --ghost
 # NOTE: mounts done from a hosts map will be mounted with the
 # "nosuid" and "nodev" options unless the "suid" and "dev"
 # options are explicitly given.
 /net    -hosts
 # Include /etc/auto.master.d/*.autofs
 # The included files must conform to the format of this file.
 # If you have fedfs set up and the related binaries, either
 # built as part of autofs or installed from another package,
 # uncomment this line to use the fedfs program map to access
 # your fedfs mounts.
 #/nfs4  /usr/sbin/fedfs-map-nfs4 nobind
 # Include central master map if it can be found using
 # nsswitch sources.
 # Note that if there are entries for /net or /misc (as
 # above) in the included master map any keys that are the
 # same will not be seen as the first read key seen takes
 # precedence.
 [root@centos-c ~]#

step 2: define the map file (added above)

[root@centos-c ~]# cat /etc/auto.nfs-mount-client 
data -fstype=nfs
[root@centos-c ~]#

step 3: reload the autofs daemon

systemctl reload autofs.service

that it! now when you browse to the mounted directory on the nfs client, the nfs will be mounted automatically.

Good things to know

  • the default settings in autofs do not show the mounted directories. this might give you the illusion that the directory is not mounted at all even when you are in the mount point! you need to remember the name of the directory to cd into that. if you dont want to remember all this, then you may use the –ghost option in the map file options defined in the auto.master configuration file, as shown in the example above. i prefer that and like to see the mounted directories.
  • another way would be to turn on the browse_mode in the autofs.conf file. it is the same as adding the –ghost option manually.
  • if you’re using a systemd based linux, you might not be interested in autofs at all! yeah, use the systemd automount instead. it is built in to the system already and no more remembering the autofs mapping and stuff. no installing the additional autofs package either.


man 5 autofs


selinux prevents mounted directories for serving content for httpd

error message

(13)Permission denied: [client <ipaddress>:<port>] AH00035: access to /index.html denied (filesystem path '/opt/rh/httpd24/root/var/www/html/index.html') because search permissions are missing on a component of the path


run this on the host directory serving the content

chcon  --user system_u --type httpd_sys_content_t -Rv <dir>

if you’ve worked with selinux before, you will realise that chcon is a temporary but a faster way to apply contexts to a file. they will be reset in the next relabel. good to use in places where you do not care too much, for example containers.

if you wish to apply contexts to the file/directory permanently, probably to a directory which is being shared by several containers, file servers, where it doesn’t make much sense to have an extra command to set the security context each time a new container is spun up, then use the following command. you run it one time, and the virtual machines or containers will be able to serve the content correctly.

semanage fcontext -a -s system_u -t httpd_sys_content_t <dir>

the above command will make an entry for the file’s context in the /etc/selinux/targeted/contexts/files/file_contexts.local file, which is used to relabel the files in the future. the above command does not apply the context immediately. you will have to do it manually.

restorecon -RvF <dir>


selinux needs the correct context to serve the content

courtesy –


setup and use nfs server on rhel 8

default nfs version in rhel 8 is version 4

install nfs-server

dnf -y install nfs-server nfs-utils

most probably the nfs-utils package is there on your server already.

ensure nfs server starts up automatically

systemctl enable --now nfs-server

firewall stuff

firewall-cmd --add-service=nfs --add-service=rpc-bind --add-servicemountd --permanent
firewall-cmd --reload

add directories to share on the nfs server

file controlling the exports is /etc/exports

check out the man page for exports for more details

man exports

example content for /etc/exports file (pasted from the exports man page)

# sample /etc/exports file
/               master(rw) trusty(rw,no_root_squash)
/projects       proj*.local.domain(rw)
/usr            *.local.domain(ro) @trusted(rw)
/home/joe       pc001(rw,all_squash,anonuid=150,anongid=100)
/pub            *(ro,insecure,all_squash)
/srv/www        -sync,rw server @trusted @external(ro)
/foo            2001:db8:9:e54::/64(rw)
/build          buildhost[0-9].local.domain(rw)

The first line exports the entire filesystem to machines master and trusty.  In addition to write access, all uid squashing is turned off for host trusty. The second and third entry show examples for wildcard hostnames and netgroups (this is the entry `@trusted'). The fourth line shows  the  entry  for the PC/NFS client discussed above. Line 5 exports the public FTP directory to every host in the world, executing all requests under the nobody account. The insecure option in this entry also allows clients with NFS implementations that don't use a reserved port for NFS.  The sixth line exports a directory read-write to the machine 'server' as well as the `@trusted' netgroup, and read-only to netgroup `@external', all three mounts with the `sync' option enabled. The seventh line exports a directory to both an IPv6 and an IPv4 subnet. The eighth line demonstrates a character class wildcard match.

mount nfs shared directories on your machine

you’ve got 2 options – root mount, and directory mount.

root mount – if you mount the / of the nfs server, then all the shared directories with you will appear in this mount.

otherwise you may check what directories are shared with you using:

showmount --exports <server_address>

now mount them individually using the mount command

notes for adding an entry in the fstab

version may be specified using the nfsvers= argument

important – remember to add the _netdev to the options to mount the nfs share after the networking has started on the machine, else…

example /etc/fstab entry for a nfs share

nfshost:/   /nfs-share   nfs   defaults,_netdev   0 0

resolving permissions issue while using nfs

try chown‘ing the nfs shared directories to nobody

monitoring nfs server

nfsstat and nfsiostat