File systems and mounting

A partition is an empty container. A file system is the structure that makes it useful: the directories, the names, the permissions, the records of which blocks belong to which file. You create one with mkfs, and then mount it, which means attaching it to a directory so that its content appears in the tree.

Create XFS and ext4 file systems, and prepare swap space.
Mount and unmount file systems by hand, and check what is mounted with df, findmnt and lsblk -f.
Find UUIDs and labels with blkid.

Choosing a file system

XFS or ext4?: Select a file system.

RHEL's default is XFS, and it is the right choice unless you have a reason. The essential practical difference: XFS can be enlarged but not reduced, ext4 can be both.

Creating file systems

[root@servera ~]# mkfs.xfs /dev/sdb1
meta-data=/dev/sdb1              isize=512    agcount=4, agsize=65536 blks
...output omitted...
[root@servera ~]# mkfs.ext4 -L logs /dev/sdb2
...output omitted...
Writing superblocks and filesystem accounting information: 0/8   done
[root@servera ~]# mkswap /dev/sdb3
Setting up swapspace version 1, size = 512 MiB (536866816 bytes)
no label, UUID=89858a47-2255-4fa3-a2a9-64711ac2d782

mkfs destroys whatever was on the partition. -L sets a label. mkswap prepares swap space. Now see what the system knows about them:

[root@servera ~]# blkid /dev/sdb1 /dev/sdb2 /dev/sdb3
/dev/sdb1: UUID="c64946eb-fc27-4485-8020-db4119b686eb" TYPE="xfs" PARTLABEL="data" PARTUUID="e854bd58-..."
/dev/sdb2: LABEL="logs" UUID="cd0b3883-1d6d-4271-a89e-2f48a2dac5b7" TYPE="ext4" PARTLABEL="logs" PARTUUID="79e6c3a7-..."
/dev/sdb3: UUID="89858a47-2255-4fa3-a2a9-64711ac2d782" TYPE="swap" PARTLABEL="swap" PARTUUID="e6f880d9-..."

Each file system has a UUID, a unique identifier written when it was created. (The PARTUUID belongs to the partition, a different thing.)

Mounting

The tree under / is built from file systems. mount DEVICE DIRECTORY attaches one; the directory (the mount point) must exist, and whatever it held is hidden while something is mounted on it.

[root@servera ~]# mkdir -p /data /logs
[root@servera ~]# mount /dev/sdb1 /data
[root@servera ~]# mount /dev/sdb2 /logs
[root@servera ~]# df -hT /data /logs
Filesystem     Type  Size  Used Avail Use% Mounted on
/dev/sdb1      xfs   960M   39M  922M   5% /data
/dev/sdb2      ext4  974M   24K  907M   1% /logs
[root@servera ~]# findmnt -o TARGET,SOURCE,FSTYPE,OPTIONS /data
TARGET SOURCE    FSTYPE OPTIONS
/data  /dev/sdb1 xfs    rw,relatime,seclabel,attr2,inode64,logbufs=8,logbsize=32k,noquota

Command	Shows
`df -h` / `df -hT`	Size, used, available per mounted file system (and type)
`findmnt [DIR]`	The mount tree, with the options in use
`lsblk -f`	Disks and partitions with file system, UUID and mount point
`mount` (no arguments)	Every mount, the raw list
`du -sh DIR`	How much a directory uses (a different question from `df`)

mount -o ro,noexec DEVICE DIR adds options for this mount; mount -o remount,rw DIR changes a mounted one. To detach: umount DIR. It fails with "target is busy" if something is using the file system, which includes your own shell if you are standing in the directory.

Swap

Swap is not mounted. You enable it, and the kernel uses it as overflow for memory:

[root@servera ~]# swapon /dev/sdb3
[root@servera ~]# swapon --show
NAME      TYPE      SIZE USED PRIO
/dev/sdb3 partition 512M   0B   -2
[root@servera ~]# free -h | tail -1
Swap:          511Mi          0B       511Mi

swapoff /dev/sdb3 disables it again. Everything you did with mount and swapon so far is lost at reboot. The next lesson makes it permanent.

Try it

Check your understanding

Which command creates an XFS file system?
mount /dev/sdb1 /data fails with "mount point does not exist". What do you do?
Which command lists mounted file systems with their sizes and types?
Which file system can NOT be shrunk?
Why is blkid useful?
Swap space is activated with ...