eolas/neuron/58bd372f-64fb-4d7a-a8dc-9bae966b3483/Filesystems.md
2025-01-01 16:02:50 +00:00

103 lines
3.9 KiB
Markdown

---
tags:
- disks
- computer-architecture
- operating-systems
---
# Filesystems
We cannot yet mount or interact with the partitions we have created. This is
because we have not added a filesystem to each partition.
> A filesytem is a form of [database](Basic_database_concepts.md); it
> supplies the structure to transform a simple block device into the
> sophisticated hierarchy of files and subdirectories that users can understand.
Linux recognises many types of filesystems. The native Linux filesystem is the
**ext4** (Fourth Extended Filesystem). Another common filesystem is **FAT**
(File Allocation Table). Instances of this include _MSDOS_,_EXFAT_,_FAT-32_.
They originate from Microsoft systems
## Creating a filesystem
Remember we have two partitions on our external drive: `sda1` and `sda2`. We are
going to use the `mkfs` utility to create an EXT4 system on both.
```bash
mkfs -t ext4 /dev/sda1
mkfs -t ext4 /dev/sda2
```
## Mounting a filesystem
We can now mount our filesystems. When we mount, we must specify the following
criteria with the request:
- The name of the device we want to mount.
- This will be the name or the partition. However the names (`sda` etc)
assigned by the OS can change. In these cases and with GPT-based partitions
you can use the UUID.
- To see a list of devices and the corresponding filesystems and UUIDs on your
system, you can use the **`blkid`** ('block id') program.
```
/dev/nvme0n1p3: UUID="c53577b5-92ef-4a0a-9a19-e488bfdfa39c" BLOCK_SIZE="4096" TYPE="ext4" PARTUUID="e152b9f4-7ce8-e74b-94db-2731c6fce53d"
/dev/nvme0n1p1: UUID="9920-636A" BLOCK_SIZE="512" TYPE="vfat" PARTUUID="50592521-d386-194a-a362-bc8562ed6c82"
/dev/nvme0n1p2: UUID="2ee6b834-0857-49dc-b8ba-a24d46d228ae" BLOCK_SIZE="4096" TYPE="ext4" PARTUUID="e08cc442-ef51-7b4f-9d55-e236c55c933c"
/dev/sda2: UUID="abac6e2e-e3bf-40d3-a5ba-317c53eb27ce" BLOCK_SIZE="4096" TYPE="ext4" PARTUUID="4ef1b0e8-3d5b-c940-a3b1-0f85cddeca42"
/dev/sda1: UUID="ba1e40c5-9b29-4309-a559-99bf8f68116f" BLOCK_SIZE="4096" TYPE="ext4" PARTUUID="b4983358-6036-df40-a1f8-793976f3dfb1"
```
- The filesystem type (optional)
- The **mount point**
- This is the place within the existing filesystem where you want to mount the
partition.
- When you mount to a directory, this directory _becomes_ the disk you have
mounted, you will not see it as a subdirectory within the the mount point,
you will just see the contents of the disk itself
```bash
mkdir mountpoint
mount -t ext4 /dev/sda1 /mnt
touch test.txt
```
Our `sda1` partition is now mounted at `mountpoint`. We can go ahead and create
files. If we now look within the graphical file manager when we click on the
`sda1` volume, we will see the new file we have created in `mountpoint`.
![](static/mount-directory.png)
## fstab
In most cases you want your filesystem to mount automatically on boot and always
to the same mount point. You can do this via the specialised `fstab` file on
Linux systems within the `/etc/` directory.
This is my current `fstab`:
```
# <file system> <dir> <type> <options> <dump> <pass>
# /dev/nvme0n1p2
UUID=2ee6b834-0857-49dc-b8ba-a24d46d228ae / ext4 rw,relatime 0 1
# /dev/nvme0n1p3
UUID=c53577b5-92ef-4a0a-9a19-e488bfdfa39c /home ext4 rw,relatime 0 2
/swapfile none swap sw 0 0
```
It shows my root and home filesystems and my
[swap](Swap_space.md) file. Note that we use the UUID
to name the partition rather than its name in `/dev/`. The order of the
parameters is as follows:
- Device name or UUID
- The mount point
- The filesystem type (in the example there are two types (`ext4` and `swap`))
- Options
- Now largely irrelevant backup information, set to )
- Filesystem integrity test order
- Runs the `fsck` ('filesystem check') command against each filesystem
- Put `1` against the root partition for this to be checked first
- Put `0` for no checks to take place