Managing Virtual Machines

Managing Virtual Machines

The primary tool for managing KVM virtual machines from the command line is virsh, which talks to the libvirt daemon. Everything you can do in virt-manager's GUI you can do with virsh, making it suitable for automation and headless server management. This article covers the daily operations: starting/stopping VMs, checking status, modifying resources, and connecting to VM consoles.

virsh basics

# List VMs:
virsh list          # Running VMs only
virsh list --all    # All VMs including stopped

# Get VM info:
virsh dominfo ubuntu-test
virsh domstats ubuntu-test    # Resource usage statistics

virsh list --all output

 Id   Name          State
-------------------------------
 1    ubuntu-prod   running
 -    ubuntu-test   shut off
 -    ubuntu-dev    paused

VM lifecycle commands

# Start / stop / restart:
virsh start ubuntu-test         # Start a stopped VM
virsh shutdown ubuntu-test      # Graceful shutdown (sends ACPI signal)
virsh destroy ubuntu-test       # Force stop (like pulling the power cord)
virsh reboot ubuntu-test        # Reboot
virsh suspend ubuntu-test       # Pause (freeze in memory)
virsh resume ubuntu-test        # Resume from suspended state

# Connect to VM console (serial console):
virsh console ubuntu-test    # Exit with Ctrl+]

# Get VM's IP address:
virsh domifaddr ubuntu-test

⚠️ WARNING: virsh destroy is a hard power-off — it stops the VM immediately without flushing disk buffers, which can corrupt the filesystem. Always use virsh shutdown first and only use destroy if the VM is unresponsive.

Modifying VM resources

# Add CPU/RAM (requires VM to be stopped):
virsh setmaxmem ubuntu-test 4G --config    # Set max allowed RAM
virsh setmem ubuntu-test 4G --config       # Set current RAM
virsh setvcpus ubuntu-test 4 --config --maximum
virsh setvcpus ubuntu-test 4 --config

# Edit VM XML configuration directly:
virsh edit ubuntu-test    # Opens in $EDITOR

# Add a disk to a running VM:
virsh attach-disk ubuntu-test /var/lib/libvirt/images/extra.qcow2 vdb   --cache none --persistent

# Resize an existing disk image (VM must be stopped):
sudo qemu-img resize /var/lib/libvirt/images/ubuntu-test.qcow2 +20G
# Then expand the partition inside the VM (growpart + resize2fs)

Autostart and scheduling

# Start VM automatically when host boots:
virsh autostart ubuntu-prod

# Disable autostart:
virsh autostart ubuntu-prod --disable

# List VMs with autostart status:
virsh list --all --autostart

# Suspend all VMs before a maintenance window:
for vm in $(virsh list --name); do virsh suspend $vm; done
# Resume all:
for vm in $(virsh list --name); do virsh resume $vm; done

Conclusion

Daily VM management with virsh: virsh list --all for status overview, virsh start/shutdown for power operations, and virsh console when you need to access a VM that has lost network connectivity. Enable virsh autostart for production VMs that should survive host reboots. Keep virsh destroy as a last resort, not a first response to a stuck VM.