Skip to main content

How to encrypt a single Linux filesystem

Sure, you can manually encrypt a filesystem. But, you can also automate it with Ansible.
Image
How to encrypt a single Linux filesystem

Photo by PhotoMIX Company on Pexels

There are a few different reasons that you might want to encrypt a filesystem, such as protecting sensitive information while it's at rest, not having to worry about encrypting individual files on the filesystem, or other reasons. To manually encrypt a filesystem in Red Hat Enterprise Linux (RHEL), you can use the cryptsetup command. This article will walk you through how to use Ansible to do this for you for a RHEL 8 server.

Before we dive into using Ansible to automate that process, let's first go through the steps to manually create the encrypted filesystem so that we better understand what we're asking Ansible to do. There are native commands in RHEL that enable you to create an encrypted filesystem, and we'll use those in our walkthrough.

[ You might also enjoy reading: Configuring LUKS: Linux Unified Key Setup ]

Manually create an encrypted partition

To start with, we'll look at the device on which I'll put the partition:

[root@ansibleclient ~]# fdisk /dev/vdc

Welcome to fdisk (util-linux 2.32.1).
Changes will remain only in memory until you decide to write them.
Be careful before using the write command.


Command (m for help): p
Disk /dev/vdc: 30 GiB, 32212254720 bytes, 62914560 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x803e8b19

Device     Boot Start     End Sectors Size Id Type
/dev/vdc1        2048 6291455 6289408   3G 83 Linux

Command (m for help):

We can see that my /dev/vdc already has a partition on it, but there is still space available for another partition. I’ll create my /dev/vdc2 partition:

Command (m for help): n
Partition type
   p   primary (1 primary, 0 extended, 3 free)
   e   extended (container for logical partitions)
Select (default p):

Using default response p.
Partition number (2-4, default 2):
First sector (6291456-62914559, default 6291456):
Last sector, +sectors or +size{K,M,G,T,P} (6291456-62914559, default 62914559): +7G

Created a new partition 2 of type 'Linux' and of size 7 GiB.

Command (m for help): p
Disk /dev/vdc: 30 GiB, 32212254720 bytes, 62914560 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x803e8b19

Device     Boot   Start      End  Sectors Size Id Type
/dev/vdc1          2048  6291455  6289408   3G 83 Linux
/dev/vdc2       6291456 20971519 14680064   7G 83 Linux

Command (m for help): w
The partition table has been altered.
Syncing disks.

[root@ansibleclient ~]# partprobe /dev/vdc
[root@ansibleclient ~]#

I now have a partition /dev/vdc2 of size 7G. Next, I format that partition for luks:

[root@ansibleclient ~]# cryptsetup luksFormat /dev/vdc2

WARNING!
========
This will overwrite data on /dev/vdc2 irrevocably.

Are you sure? (Type uppercase yes): YES
Enter passphrase for /dev/vdc2:
Verify passphrase:
[root@ansibleclient ~]#

To open the encrypted volume, I use the luksOpen argument for cryptsetup, and I tell it the name I want my target to be manualluks:

[root@ansibleclient ~]# cryptsetup luksOpen /dev/vdc2 manualluks
Enter passphrase for /dev/vdc2:
[root@ansibleclient ~]# ls /dev/mapper/
control  examplevg-examplelv  manualluks  mycrypt  rhel-root  rhel-swap
[root@ansibleclient ~]#

After it's been opened, I can actually put it to use. In this example, I'll put a volume group there:

[root@ansibleclient ~]# vgcreate manual_luks_vg /dev/mapper/manualluks
  Physical volume "/dev/mapper/manualluks" successfully created.
  Volume group "manual_luks_vg" successfully created
[root@ansibleclient ~]# vgdisplay manual_luks_vg
  --- Volume group ---
  VG Name               manual_luks_vg
  System ID            
  Format                lvm2
  Metadata Areas        1
  Metadata Sequence No  1
  VG Access             read/write
  VG Status             resizable
  MAX LV                0
  Cur LV                0
  Open LV               0
  Max PV                0
  Cur PV                1
  Act PV                1
  VG Size               6.98 GiB
  PE Size               4.00 MiB
  Total PE              1787
  Alloc PE / Size       0 / 0   
  Free  PE / Size       1787 / 6.98 GiB
  VG UUID               bjZ7FM-9jNw-pdfs-Dd5y-5IsF-tEdK-CpVqH4
   
[root@ansibleclient ~]#

I have a volume group, manual_luks_vg, so I'm now able to put a logical volume inside:

[root@ansibleclient ~]# lvcreate -n manual_luks_logvol -L +5G manual_luks_vg
  Logical volume "manual_luks_logvol" created.
[root@ansibleclient ~]# lvdisplay manual_luks_vg
  --- Logical volume ---
  LV Path                /dev/manual_luks_vg/manual_luks_logvol
  LV Name                manual_luks_logvol
  VG Name                manual_luks_vg
  LV UUID                nR5UKo-jRvR-97L0-60YF-dbSp-D0pc-l8W3Td
  LV Write Access        read/write
  LV Creation host, time ansibleclient.usersys.redhat.com, 2020-12-03 10:15:03 -0500
  LV Status              available
  # open                 0
  LV Size                5.00 GiB
  Current LE             1280
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     8192
  Block device           253:5
   
[root@ansibleclient ~]#

The lvcreate command specified the name for my new logical volume, manual_luks_logvol, its size, 5G, and that the logical volume should be in the volume group of manual_luks_vg.

At this point, I have a logical volume, but I haven't formatted it yet for ext or xfs. Typing mkfs and then hitting Tab shows me that there are a number of options for me to format this partition:

# mkfs
mkfs         mkfs.cramfs  mkfs.ext2    mkfs.ext3    mkfs.ext4    mkfs.minix   mkfs.xfs

Here, I’ll use mkfs.xfs:

[root@ansibleclient ~]# mkfs.xfs /dev/manual_luks_vg/manual_luks_logvol
meta-data=/dev/manual_luks_vg/manual_luks_logvol isize=512    agcount=4, agsize=327680 blks
         =                       sectsz=512   attr=2, projid32bit=1
         =                       crc=1        finobt=1, sparse=1, rmapbt=0
         =                       reflink=1
data     =                       bsize=4096   blocks=1310720, imaxpct=25
         =                       sunit=0      swidth=0 blks
naming   =version 2              bsize=4096   ascii-ci=0, ftype=1
log      =internal log           bsize=4096   blocks=2560, version=2
         =                       sectsz=512   sunit=0 blks, lazy-count=1
realtime =none                   extsz=4096   blocks=0, rtextents=0

I have it formatted, but not mounted. To mount it, I’ll create a new directory and then run the mount command:

[root@ansibleclient ~]# mkdir /manual_luks
[root@ansibleclient ~]# mount /dev/manual_luks_vg/manual_luks_logvol /manual_luks

To verify that worked, I can use mount by itself and then write to a new file there:

[root@ansibleclient ~]# mount | grep luks
/dev/mapper/manual_luks_vg-manual_luks_logvol on /manual_luks type xfs (rw,relatime,seclabel,attr2,inode64,noquota)
[root@ansibleclient ~]# date > /manual_luks/testing
[root@ansibleclient ~]# cat /manual_luks/testing
Thu Dec  3 10:24:42 EST 2020
[root@ansibleclient ~]#

To enable the system to mount the encrypted partition at boot, I need to update my /etc/crypttab file. The format for the file is the name of your luks device, the physical partition, and then the file whose only contents are the password for that luks device:

# cat /etc/crypttab
manualluks /dev/vdc2 /root/manualluks.txt

In the /root/manualluks.txt, I have just the plaintext password for my luks device.

I use the luksAddKey argument to add the key to the device:

# cryptsetup luksAddKey /dev/vdc2 /root/manualluks.txt

To mount the filesystem at boot time, edit the /etc/fstab file so there is an entry for the logical volume and its mount point:

/dev/manual_luks_vg/manual_luks_logvol /manual_luks xfs defaults 0 0

After you've done the manual steps for creating the partition and writing to it, give the system a reboot to verify that the settings are persistent and the system reboots as expected.

Now that we understand what we need to do to manually create an encrypted partition, we know what we need to do to automate that process.

Automate the creation of an encrypted partition

The script hosted at https://people.redhat.com/pgervase/sysadmin/partition.yml gives one example of how to use Ansible to take a blank disk and go through the steps to create an encrypted partition, mount it, and then write to it. Like so many things with technology, there are several different ways to accomplish this, but this approach will also show some examples of variables, getting facts, and using a block and rescue.

---
- name: pb to create partition
  hosts: all
  become: true
  vars:
    target_size: 3GiB
    target_device: /dev/vdc
    myvg: examplevg
    mylv: examplelv
    keyfile: /root/mylukskey.yml
    mycrypt: mycrypt

At the top of the playbook, I place some basic information and declare a few variables. Rather than having the parameters hardcoded in the playbook, by having them defined as variables, I can override them when I run the play and make the tasks able to be used for other purposes.

 tasks:
    - name: block for doing basic setup and verification for target system
      block:
        - name: get facts for "{{ target_device }}"
          parted:
            device: "{{ target_device }}"
          register: target_facts

        - name: print facts for "{{ target_device }}"
          debug:
            msg: "{{ target_facts }}"

        - name: check to see if there are any facts for /dev/vdb1. this means there are existing partitions that we would overwrite, so fail
          debug:
            msg: "{{ target_facts }}.partitions"
          failed_when: ansible_devices.vdb.partitions.vdb1 is defined   ### if vdb1 is defined, there's already a partition there, so abort.

        - name: print size for the disk
          debug:
            msg: "the size is {{ target_facts['disk']['size'] }} kib"

        - name: copy keyfile to remote system
          copy:
            src: mylukskey.yml
            dest: "{{ keyfile }}"

        - name: make sure cryptsetup is installed
          yum:
            name: cryptsetup
            state: installed

The first few tasks that get run are going to get information about my targeted system and make sure that I'm not going to overwrite an existing partition. I then copy the keyfile onto my remote system. This keyfile contains the passphrase which will be used when I create the LUKS container. Not all systems will have the cryptsetup package installed, so the next thing to do is install that RPM if it's not already installed.

   - name: block to attempt to get info on what my destination device will become
      block:
        - name: task to attempt to get info on what my destination device will be
          parted:
            device: "{{ target_device}}"
            number: 1
            state: info
          register: info_output
        - name: print info_output
          debug:
            msg: "{{ info_output }}"

    - name: block to attempt parted
      block:
        - name: use parted in block to create new partition
          parted:
            device: "{{ target_device }}"
            number: 1
            state: present  
            part_end: "{{ target_size }}"
          register: parted_output

      rescue:
        - name: parted failed
          fail:
            msg: 'parted failed:  {{ parted_output }}'

At this point, I have a system that is ready and appropriate to be partitioned. For my own logging purposes, I have a task that prints out the information that parted gives back for my target device, /dev/sdb. The partitions here should be blank because I've already failed when ansible_devices.vdb.partitions.vdb1 is defined, so this is simply for verification. Next, I use parted to create my partition. To catch any errors in this step—maybe my destination device is too small, or something else happened—I use a block and rescue to register the output of parted and then display that in the fail part of my rescue section.

    - name: block for LUKS and filesystem tasks
      block:
        - name: create LUKS container with passphrase
          luks_device:
            device: "{{ target_device }}1"
            state: present
            name: "{{ mycrypt }}"
            keyfile: "{{ keyfile }}"

        - name: open luks container
          luks_device:
            device: "{{ target_device }}1"
            state: opened
            name: "{{ mycrypt }}"
            keyfile: "{{ keyfile }}"

        - name: create a new volgroup in that partition
          lvg:
            vg: "{{ myvg }}"
            pvs: "/dev/mapper/{{ mycrypt }}"

        - name: create a logvol in my new vg
          lvol:
            vg: "{{ myvg }}"
            lv: "{{ mylv }}"
            size: +100%FREE`

       - name: create a filesystem
          filesystem:
            fstype: xfs
            dev: "/dev/mapper/{{ myvg }}-{{ mylv }}"

Now that I have a partition and cryptsetup installed, I need to do the LUKS and filesystem part of my setup. The first step is to use the luks_device module, along with the keyfile that I copied over. After I have the LUKS container, I create the volume group, then the logical volume, and then the filesystem.

       - name: mount device
          mount:
            path: /mnt
            src: "/dev/mapper/{{ myvg }}-{{ mylv }}"
            state: mounted
            fstype: xfs

    - name: put some content in my new filesystem
      copy:
        content: "this is secure content!"
        dest: /mnt/newcontent.txt

    - name: set content in /etc/crypttab so I can mount the partition on reboot
      copy:
        content: "{{ mycrypt }} {{ target_device }}1 {{ keyfile }}"
        dest: /etc/crypttab
        owner: root
        group: root
        mode: 0644

After I have a filesystem there, I mount the filesystem and write a test file to verify that everything is working correctly. The final step is to create the /etc/crypttab file so that the system can mount my filesystem when it gets rebooted.

[ Want to learn more about security? Check out the IT security and compliance checklist. ] 

Wrap up

The process of manually configuring an encrypted partition is not particularly difficult, or even time-consuming. However, such tasks are perfect for Ansible to handle for you, helping to ensure consistent, secure, and reproducible configurations.

Further information about LUKS devices can be found at:

Topics:   Linux   Linux administration   Security  
Author’s photo

Peter Gervase

I am a Senior Principal Security Architect at Verizon. Before that, I worked at Red Hat in various roles such as consulting and in the Solutions Architect where I specialized in Smart Management, Ansible, and OpenShift. More about me

Try Red Hat Enterprise Linux

Download it at no charge from the Red Hat Developer program.