Customer says the Azure VM's C: or / drive is full: expanding the OS disk

#Before you expand

1. Find what's using the space first. Expanding a disk to hide a runaway log just delays the next ticket.
   • Windows: Get-ChildItem C:\ -Recurse -ErrorAction SilentlyContinue | Sort-Object Length -Descending | Select-Object -First 20 FullName, Length, or use TreeSize/WinDirStat.
   • Linux: sudo du -xh / --max-depth=2 | sort -h | tail -20
2. Snapshot the disk. It's quick and gives you a way back.

RG=rg-prod; VM=vm-app-01
DISK=$(az vm show -g $RG -n $VM --query "storageProfile.osDisk.managedDisk.id" -o tsv)
az snapshot create -g $RG -n "${VM}-os-pre-expand-$(date +%F)" --source $DISK

#Grow the managed disk

Most disk types can be expanded without deallocating (live resize), as long as the disk isn't larger than 4 TiB before or after. If live resize isn't supported for your configuration, deallocate first.

az disk update --ids $DISK --size-gb 256
az disk show --ids $DISK --query "{size:diskSizeGb, state:diskState}" -o table

Warning

Disks can only grow, never shrink. Choose the size deliberately, and remember that a bigger Premium SSD also moves into a higher performance tier.

#Extend the partition: Windows

# Rescan, then grow C: to all available space
"rescan" | diskpart
$max = (Get-PartitionSupportedSize -DriveLetter C).SizeMax
Resize-Partition -DriveLetter C -Size $max
Get-Volume -DriveLetter C

If Resize-Partition fails, a recovery partition sits directly after C:. You can move or delete it (after a backup) or extend through Disk Management. This is common on older images.

#Extend the partition: Linux

lsblk                                   # confirm the disk shows the new size
sudo growpart /dev/sda 1                # grow partition 1 (check the number with lsblk)
# ext4
sudo resize2fs /dev/sda1
# XFS (RHEL default) - pass the mount point
sudo xfs_growfs /
df -h /

With LVM (common on RHEL images), grow the physical volume, then the logical volume and the filesystem together:

sudo growpart /dev/sda 4
sudo pvresize /dev/sda4
sudo lvextend -r -l +100%FREE /dev/mapper/rootvg-rootlv   # -r also resizes the filesystem

#Verify and close the ticket

• df -h or Get-Volume shows the new free space.
• The application is writing again. Check its logs for "disk full" errors stopping after the change.
• Delete the snapshot once the change is confirmed stable, because it costs money.

#Prevent it next time

• Alert at 85% used on every volume (VM insights LogicalDisk / FreeSpacePercentage).
• Move logs and data to data disks so the OS disk only holds the OS.
• Set up log rotation (logrotate on Linux, IIS log cleanup on Windows) as part of the image.