14.1 Mount and Unmount Network File Systems Using NFS

14.1 Mount and Unmount Network File Systems Using NFS

Quick Answer: Mount with mount -t nfs server:/export /mnt/path, unmount with umount, and persist with an /etc/fstab line of type nfs (often with _netdev). Prove success with findmnt / df -h. On EX200 you are usually the NFS client—the share already exists on a server named in the task.

Official skill and why it matters

Under Create and configure file systems, Red Hat expects you to mount and unmount network file systems using NFS. Typical lab wording:

• “Mount the NFS share server1.example.com:/share at /mnt/nfsdata and ensure it mounts at boot.”
• “Configure a persistent NFS mount of 172.25.250.10:/exports/docs on /docs.”
• “Unmount the NFS file system from /mnt/temp when finished.”

A manual mount that works in your current shell is incomplete if the task requires boot persistence. Pair every durable NFS requirement with a correct fstab (or, in the next section, autofs when the task asks for on-demand mounts).

NFS client vs server (scope control)

Do not rebuild an NFS server unless the task explicitly says to export. If you only need to use a share, stay on the client side.

Packages and basic connectivity

On RHEL 10 clients, NFS mount helpers usually ship with the base system or the NFS utilities package. If mount.nfs or showmount is missing:

rpm -q nfs-utils
sudo dnf install -y nfs-utils

Confirm the server is reachable before blaming NFS:

ping -c 2 server1.example.com
# or use the IP the task gives
getent hosts server1.example.com

Name resolution failures produce “cannot resolve” style mount errors—fix DNS/hosts first (networking chapter skills).

Discover exports: showmount (when available)

showmount -e server1.example.com
showmount -e 172.25.250.10
showmount -a server1.example.com   # all client mounts (server-side view; if permitted)

showmount -e lists exports advertised by the server (classic NFSv3-style mountd query). Output resembles:

Export list for server1.example.com:
/exports/docs  *
/share         172.25.0.0/16

Caveats for the exam:

• Some environments restrict or firewall mountd/rpcbind so showmount times out even though NFSv4 mounting still works on port 2049.
• If the task gives you server:/path, you do not need discovery—use the path as written.
• Prefer the exact export string from the task or from a successful showmount line; trailing path mistakes (/share vs /shares) fail the mount.

When showmount fails but the task provides the export, proceed with mount -t nfs using that export.

Manual mount: mount -t nfs

Create the local directory first (empty mountpoint):

sudo mkdir -p /mnt/nfsdata
sudo mkdir -p /docs

Mount forms:

# Explicit type (recommended teaching form)
sudo mount -t nfs server1.example.com:/share /mnt/nfsdata

# IP and export path
sudo mount -t nfs 172.25.250.10:/exports/docs /docs

# Often equivalent helpers
sudo mount -t nfs4 server1.example.com:/share /mnt/nfsdata
sudo mount.nfs server1.example.com:/share /mnt/nfsdata

On modern RHEL, nfs frequently negotiates NFSv4 by default. Use nfs4 when a task or troubleshooting requires forcing v4 syntax; otherwise nfs is the standard fstab/mount type you will see in materials.

Useful mount options (client)

Example with options:

sudo mount -t nfs -o rw,soft,timeo=30 server1.example.com:/share /mnt/nfsdata

Unless the task specifies options, start with simple defaults or no extra options, then add _netdev for persistence.

Verify the live mount

findmnt /mnt/nfsdata
findmnt -t nfs,nfs4
mount | grep nfs
df -h /mnt/nfsdata
ls /mnt/nfsdata
# Write test only if the task implies a writable share and policy allows:
touch /mnt/nfsdata/exam-test && rm -f /mnt/nfsdata/exam-test

findmnt is the cleanest proof: source should look like server:/export and FSTYPE nfs or nfs4.

Unmount NFS

cd /          # never stay inside the mountpoint
sudo umount /mnt/nfsdata
sudo umount -l /mnt/nfsdata   # lazy unmount if busy (last resort)

Busy mount errors mean a shell, process, or open file is still using the path:

lsof +f -- /mnt/nfsdata
fuser -vm /mnt/nfsdata

Close those processes or cd out of the directory, then umount again. Do not reboot solely to unmount unless you are out of time and the task requires a clean unmount state.

Persistent NFS: /etc/fstab

Six fields, same as local mounts:

# <file system>                 <mount point>  <type>  <options>           <dump> <pass>
server1.example.com:/share      /mnt/nfsdata    nfs     defaults,_netdev    0      0
172.25.250.10:/exports/docs     /docs           nfs     defaults,_netdev    0      0

Why _netdev matters

Without _netdev (or equivalent network-aware unit dependency), systemd may try to mount NFS before the network is ready, causing boot delays or failed mounts. RHEL/systemd also understands x-systemd.automount and other advanced options, but for classic EX200-style tasks, defaults,_netdev is the reliable teaching default.

Apply and test fstab

sudo mkdir -p /mnt/nfsdata
# edit /etc/fstab, then:
sudo mount -a
# or mount only that line:
sudo mount /mnt/nfsdata
findmnt /mnt/nfsdata
echo $?

mount -a mounts all non-noauto fstab entries that are not yet mounted. A non-zero exit and an error about NFS usually means wrong server path, firewall, or network.

Optional: use noauto only when the task wants a fstab entry you mount manually later—rare for “mount at boot” wording.

server1.example.com:/share  /mnt/nfsdata  nfs  noauto,_netdev  0  0

End-to-end exam workflows

Workflow A — Persistent NFS mount at boot

Task: Mount utility.lab.example.com:/rhel/shared at /mnt/shared permanently.

rpm -q nfs-utils || sudo dnf install -y nfs-utils
# Optional discovery:
showmount -e utility.lab.example.com
sudo mkdir -p /mnt/shared
# Add to /etc/fstab:
# utility.lab.example.com:/rhel/shared  /mnt/shared  nfs  defaults,_netdev  0  0
sudo mount -a
findmnt /mnt/shared
ls /mnt/shared

Workflow B — Temporary mount only

Task: Mount the share for this session; no mention of boot.

sudo mkdir -p /mnt/temp
sudo mount -t nfs 172.25.250.10:/exports/temp /mnt/temp
findmnt /mnt/temp
# When done:
sudo umount /mnt/temp

Do not leave a broken fstab line if persistence was not requested—but leaving a correct fstab is usually fine if it does not break boot.

Workflow C — showmount unavailable

showmount -e server1
# clnt_create: RPC: Program not registered  OR  timeout
# Task said: server1.example.com:/public
sudo mkdir -p /public
sudo mount -t nfs server1.example.com:/public /public
findmnt /public

Trust the task’s export path when discovery fails.

Workflow D — Fix a hanging boot mount

If NFS is wrong in fstab, boots can stall. In recovery or after login:

sudo vim /etc/fstab     # fix server/export or add noauto/_netdev
sudo mount -a
systemctl --failed

Practice labs: always run mount -a before you rely on reboot.

Firewall and SELinux awareness (client-focused)

• Server must allow NFS (firewalld services nfs, rpc-bind, mountd for older styles; NFSv4 often centers on 2049/tcp). If you only control the client and the share is “supposed to work,” focus on client config first.
• Client egress is usually open in exam networks; if mount hangs, check routing and that you used the correct server interface/IP.
• SELinux: mounting NFS is generally allowed for standard mountpoints; context issues appear more often when serving NFS or when labeling local content for export. If a service cannot read mounted files, check contexts later without abandoning a correct mount.

NFS vs local mounts (quick comparison)

Common traps

1. Forgetting to create the mountpoint — mkdir -p first.
2. Wrong export path — /share vs /exports/share; use task or showmount -e.
3. fstab without _netdev — flaky boot ordering.
4. pass field 1 or 2 on NFS — keep 0 0.
5. Staying in the directory during umount — cd / then umount.
6. Assuming showmount must work — NFSv4-only or firewalled mountd still may allow mount -t nfs.
7. Using UUID= for NFS — not the client fstab pattern for remote shares.
8. Mount works, fstab missing — fails after reboot; graders check persistence.
9. Typo in hostname — verify with ping/getent hosts.
10. Confusing this with autofs — static fstab mounts stay mounted; autofs mounts on access (next section).

Relationship to other objectives

Section checkpoint

You should install or confirm nfs-utils when needed, discover exports with showmount -e when available, mount with mount -t nfs server:/export /path, unmount safely with umount, write six-field fstab lines using type nfs and options including _netdev with dump/pass 0 0, validate with mount -a and findmnt, and treat reboot survival as the success criterion whenever the task requires a permanent NFS mount. That is the EX200 standard for mounting and unmounting network file systems using NFS.