Partitions and file systems

The virtual machine disk appears to the system as /dev/sda. The layout depends on the image you choose: Debian and Ubuntu usually install with a single root partition, while AlmaLinux uses LVM. You need to know your layout before you need to extend anything.

What you have now

lsblk -f
df -hT
sudo fdisk -l /dev/sda
sudo blkid

lsblk -f is the most useful: it shows the device tree, file systems, labels, and mount points in a single screen.

If the output contains lines like vg-root and the type lvm, you have LVM. Then these will come in handy:

sudo pvs
sudo vgs
sudo lvs

Plain partitions or LVM

Plain partitions are simpler: one partition — one file system. Only the last partition on the disk can be extended, and only if there is free space after it.

LVM adds a layer: physical volume, volume group, logical volumes. You can carve out multiple volumes, extend them in any order, and add space without reworking the layout. The trade-off is an extra layer you have to remember during recovery.

For a single machine with one disk, the difference is small. Don't redo the layout unnecessarily: moving the root is an operation with a real risk of losing the system.

File systems

Practical takeaway: if you plan for flexibility, stay on ext4. You'll almost never need to shrink a partition on a VPS anyway — the disk on the machine doesn't shrink along with your plan.

Extending after increasing the disk

sudo growpart /dev/sda 1
sudo resize2fs /dev/sda1

For LVM:

sudo growpart /dev/sda 3
sudo pvresize /dev/sda3
sudo lvextend -l +100%FREE /dev/mapper/vg-root
sudo resize2fs /dev/mapper/vg-root      # xfs_growfs / for XFS

Both operations are performed on a running system. For details on moving to a larger tier — the disk has become too small.

An error in the partition number in growpart or parted destroys the partition table, and the machine stops booting. Before any layout change: save sudo sfdisk -d /dev/sda > /root/parts.txt and make sure the data backup is fresh.

Mounting and fstab

Persistent mount points are described in /etc/fstab. Use the UUID, not the device name:

sudo blkid /dev/sda1

UUID=1a2b3c4d-...  /var/lib/postgresql  ext4  defaults,noatime  0  2

Checking before rebooting is mandatory:

sudo mount -a
sudo findmnt --verify

An error in fstab leaves the machine at an emergency console on boot. If this happens, the rescue disk will help: the machine reboots into SystemRescue, root login without a password, the disk is visible as /dev/sda — you mount the partition, edit fstab, and return to normal boot mode.

Swap

A swap file is created in the usual way:

sudo fallocate -l 2G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile
echo "/swapfile none swap sw 0 0" | sudo tee -a /etc/fstab

Swap saves you from a process being suddenly killed when memory runs out, but heavy swapping runs into disk speed and makes the machine very slow. It's wiser to limit the resource demands of services — see resource limits.

Common mistakes

• You didn't check the fstab edit with mount -a. Boot breaks, and you have to troubleshoot through the console in the control panel.
• You mounted a directory over data. The old contents didn't go anywhere, but became invisible — unmount and check.
• You didn't set noatime. It's a small thing, but the extra write on every file read on a loaded machine is noticeable.