How to Add a Dedicated Disk for Swap in Ubuntu (Hyper-V & KVM)
When a Linux VM or server starts running tight on RAM, adding swap space keeps services from crashing under load spikes. If your root disk is already nearly full, the cleanest solution is attaching a separate virtual or physical drive dedicated to swap. Here are both ways to set it up on Ubuntu in Hyper-V (the exact same Linux commands work on KVM, Proxmox, and VMware).
Quick answer: attach a second disk to the VM, find its device name with lsblk (usually /dev/sdb), initialize it with sudo mkswap /dev/sdb, enable it with sudo swapon /dev/sdb, and add its UUID to /etc/fstab so it mounts automatically after reboot.
Step 1. Attach a Second Disk in Hyper-V and Find It in Ubuntu
In your Hyper-V virtual machine settings, add a new virtual hard disk of the desired size (32 GB in my example). I also removed the unused virtual DVD drive so the device list stays clean.
Boot the VM, log into Ubuntu, and list attached block devices:
sudo lsblk -o NAME,SIZE,TYPE,MOUNTPOINT # or classic output: sudo fdisk -l
You will see your OS disk (/dev/sda) alongside the brand-new unpartitioned drive (/dev/sdb):
Disk /dev/sda: 50 GiB, 53687091200 bytes, 104857600 sectors Disk /dev/sdb: 32 GiB, 34359738368 bytes, 67108864 sectors
Warning: double-check the device name – everything on the target drive will be wiped.
Method 1 (Recommended): Use the Entire Second Disk for Swap
If the whole disk is dedicated strictly to swap, there is no reason to format ext4 and create a file on top of it – initializing the entire disk as swap is faster and cleaner:
sudo mkswap /dev/sdb sudo swapon /dev/sdb
Verify that the new swap area is active:
swapon --show free -h
Example output from swapon --show:
NAME TYPE SIZE USED PRIO /dev/dm-1 partition 1G 0B -1 /dev/sdb partition 32G 0B -2
Make It Persist After Reboot (Using UUID)
To enable the swap disk automatically on every boot, add it to /etc/fstab. Using its permanent UUID is much safer than hardcoding /dev/sdb, which can shift if you add or remove virtual disks later. Grab the UUID first:
sudo blkid /dev/sdb
Copy the UUID="..." value and append this line to /etc/fstab (the nofail option prevents the VM from stalling at boot if the disk is ever detached):
UUID=your-uuid-from-blkid none swap sw,nofail 0 0
(If your VM disk layout never changes, /dev/sdb none swap sw,nofail 0 0 also works.)
Method 2: Create a Swap File on the Second Disk (If You Need Shared Storage)
If you want to store normal files on the second disk alongside swap, format it as ext4, mount it, and allocate a swap file:
sudo mkfs.ext4 /dev/sdb sudo mkdir -p /swapdisk sudo mount /dev/sdb /swapdisk sudo fallocate -l 16G /swapdisk/swapfile sudo chmod 600 /swapdisk/swapfile sudo mkswap /swapdisk/swapfile sudo swapon /swapdisk/swapfile
Setting chmod 600 is required so non-root users cannot read swapped memory pages. To persist both mounts across reboots, add both lines to /etc/fstab (disk mount first, then swap file):
/dev/sdb /swapdisk ext4 defaults,nofail 0 2 /swapdisk/swapfile none swap sw,nofail 0 0
Quick Server Tuning Tip: vm.swappiness
Ubuntu defaults to vm.swappiness=60, which starts swapping fairly early. On servers where swap is mainly a safety net against OOM Killer during traffic spikes, setting it to 10 is a solid default:
sudo sysctl vm.swappiness=10 echo 'vm.swappiness=10' | sudo tee /etc/sysctl.d/99-swappiness.conf
On budget Linux VPS plans with modest RAM (like entry-level servers at Fornex with 7% off or Vultr), a few gigabytes of NVMe-backed swap regularly saves databases and web servers from sudden crashes. And if you want a faster way to browse and edit configs like /etc/fstab, see my guide on installing Midnight Commander (mc) on Ubuntu.
Disclosure: the Fornex and Vultr links are affiliate links – your price stays standard or discounted, while the blog earns a small commission.
For a dedicated swap disk on Hyper-V or KVM, stick with Method 1 using mkswap and a UUID entry in fstab – fewer moving parts and rock-solid across reboots.

