Three installs compared¶

Exercise three-installs-compared-annex.pdf · PDF

"Linux is better." "No, Windows is better." You have heard that argument, maybe in your own class. Who wins?

Nobody — because the question is wrong. In three days you installed three operating systems with your own hands: deb-lab, ubu-srv and win11. Now you have something better than opinions: evidence. Today you compare the three like a judge in a trial, where every claim needs a number and the place where you read it.

Evidence from the host¶

Your three VMs are three files on the host's disk — and the host is the class server, not your PC. Before starting any of them, ask libvirt — the service behind virt-manager that manages the VMs — about those files. Open a terminal on your Ubuntu PC (Ctrl+Alt+T) and log in to the server with the user and password from today's slip:

ssh <user>@<server>    # the same login as virt-manager
virsh -c qemu:///session vol-info --pool default deb-lab.qcow2
virsh -c qemu:///session vol-info --pool default ubu-srv.qcow2
virsh -c qemu:///session vol-info --pool default win11.qcow2
virsh -c qemu:///session snapshot-list deb-lab

After ssh, the prompt changes to <user>@si: you are typing on the server. qemu:///session is your own libvirt there, the one that holds your VMs. Type exit when you finish.

virsh is the command-line twin of virt-manager. Each vol-info answers with two lines worth reading:

The allocation is usually much smaller than the capacity. By default a qcow2 disk file starts small and grows as the guest writes data, so a 64 GiB disk that has received about 20 GiB takes about 20 GiB on the host, not 64.

The last command lists the snapshots of one VM: clean must be there. Press the Up key to bring the last command back, change the name to ubu-srv, then win11.

Evidence from inside¶

You already measured most of it in the last three sessions. Missing numbers for Windows? Start win11 alone and shut down any VM you are not using: the server is shared by the whole class. Then read:

• RAM in use: Ctrl+Shift+Esc → Performance → Memory → In use.
• Disk used: File Explorer → This PC → drive C: (used = total − free).

Then fill the table on the Exercise sheet. Your numbers will differ a little from your neighbour's — updates and open programs change them. The pattern is what matters:

Every difference has a cause¶

A table of numbers is only half the work. The other half is asking why for every row. Four causes explain almost everything:

flowchart TB
    subgraph CAUSES["Four causes"]
        D["<b>A desktop or not</b>"]
        L["<b>Free or proprietary</b>"]
        R["<b>Who writes the drivers</b>"]
        T["<b>Who the maker designs for</b>"]
    end
    D -->|"explains"| D1["RAM at rest, disk used,<br>kind of installer, what you see after login"]
    L -->|"explains"| L1["product key, licence,<br>online account, privacy questions"]
    R -->|"explains"| R1["the VirtIO driver<br>loaded from a CD"]
    T -->|"explains"| T1["TPM 2.0 and UEFI required,<br>old PCs refused"]

Better for what?¶

Now turn the comparison around, so that nobody leaves with "Linux good, Windows bad". Which of the three could your grandparents use alone? Which one runs the commercial programs a design studio buys? Which did you find easiest to install?

Each system is the right answer to some question. Nobody chooses an operating system because it is "the best". They choose it because it is the best for this job, on this machine, for these people, with this budget.

Three jobs, three systems¶

Choose one system for each case and give two facts from your table:

Another choice can also be right — if your two facts are true and relevant. The point is the argument.

Finish your reports¶

Your reports for Practicals 2.4 and 2.5 are documentation, not a list of clicks. Anyone can find the clicks; your report has the decisions and their reasons. Check that each one has:

• the sections of the report template, in order;
• the VM's specification as a table (vCPU, RAM, disk, firmware, ISO);
• each decision with its reason, in your own words;
• the real output of your checks, not a classmate's;
• one problem you had and how you solved it;
• the snapshot clean, and why it exists.

Key takeaways¶

• Compare systems with evidence: a number and the place where you read it.
• virsh vol-info shows a disk's capacity (what the guest sees) and its allocation (what the file really takes).
• Four causes explain the table: a desktop or not, free or proprietary, who writes the drivers, who the maker designs for.
• ubu-srv is the lightest because it has no desktop, not because Ubuntu is lighter than Debian.
• Choose a system by the job, the machine, the people and the budget, with facts.

Quick check¶

Remember how the desktop arrived on deb-lab: a thousand packages from a mirror. Next you learn to ask for one more yourself: Package managers.