I’ve always referred to the version of Debian running on a Raspberry Pi as Raspbian and so it was until the recent 8 GB RAM Raspberry Pi was launched.  But it turns out that Raspbian was an independent version of Debian created by people at raspbian.org.  As they say “Note: Raspbian is not affiliated with the Raspberry Pi Foundation. Raspbian was created by a small, dedicated team of developers that are fans of the Raspberry Pi hardware, the educational goals of the Raspberry Pi Foundation and, of course, the Debian Project.”

From now on, it seems the correct name is Raspberry Pi OS and no longer Raspbian. The change was announced at the bottom of this Raspberry Pi Foundation blog post. This has something to do with the fact that Raspbian is mainly 32-bit while the Raspberry Pi OS is both 32-bit and 64-bit though the latter is still at beta. You can read more about the recent changes to Raspberry Pi OS on this blog entry.

Most Linux development is done on Ubuntu running under Hyper-V on my Windows 10 PC. If you have lots of RAM (and I have a full 64 GB), it’s very convenient. I run Snagit on Windows and this makes it very easy to grab screenshots of the Ubuntu window.

I also have a “Raspberry-pi” running under Hyper-V.  There’s a Raspbian desktop that you can download and run in Hyper-V, VirtualBox or VMWare though I’ve only done Hyper-V. Don’t forget when you are running a Raspberry Pi this way that its x86 based not ARM. That does affect the available software, so it doesn’t behave exactly like a real Pi though often close enough.

Today though I started getting low disk space from my virtual Ubuntu. That’s the problem with virtual machines. When you first setup a Virtual hard disk, you never know just how much disk space you will need.

There’s a terminal command that shows how much space you have left.

df -h --total

This produced this

david@david-Virtual-Machine:~$ df -h --total
Filesystem      Size  Used Avail Use% Mounted on
udev            942M     0  942M   0% /dev
tmpfs           193M  1.4M  192M   1% /run
/dev/sda1        11G  9.9G  603M  95% /
tmpfs           964M     0  964M   0% /dev/shm
tmpfs           5.0M     0  5.0M   0% /run/lock
tmpfs           964M     0  964M   0% /sys/fs/cgroup
/dev/sda15      105M  3.6M  101M   4% /boot/efi
tmpfs           193M   16K  193M   1% /run/user/121
tmpfs           193M   24K  193M   1% /run/user/1000
total            14G  9.9G  4.1G  71% -

This was after I’d extended my virtual hard disk.  You can see I now have 4.1 GB free.

The pretty picture is from a utility filelight. You install it in the usual way

sudo apt install filelight

Or if you prefer a more visual insight, install qdirstat.

sudo apt install qdirstat

This is like WinDirStat on Windows but qdirstat seems to run many times faster. It took a couple of seconds to produce this image below. WinDirStat would take 10-30 minutes.

So how did I expand my Hyper-V hard drive?

First you have to get rid of any checkpoints. Save your Hyper-V session if open then delete the checkpoint.

Right click on the checkpoint for the selected VM and click delete. This will take a minute or two and you’ll see it have a Merging status. You may need to shutdown the VM.

After that you can go into the settings and it will let you edit the virtual hard drive and change the size.

Well to be fair, it was me that made it unbootable. I’d been reading this Wiki page on configuring the Pi. I’d told it to give the GPU almost a GB of RAM. The Pi is a 4 GB Pi 4B.  I did it last night and so this morning, I found it not working at all well.

My first thoughts were I’d messed up the Operating system and so I took the SD card put it in a holder and booted up my old laptop which has Ubuntu 18.04 LTS on it. This page on switchdoc suggested I could do a repair with these commands.

First lsblk to view attached devices. There was a /dev/sdc2 . There was also a /dev/sdc1

Then

sudo fsck -fy /dev/sdc1

That gave information about the drive but not the disk. That took

sudo fsck -fy /dev/sdc2

That took a few seconds and listed information, but still my drive wouldn’t boot back in the Pi.

Then I remembered I’d changed the boot config.txt and it was back with the SD card in the laptop and rebooted that. It showed two devices on the desktop (no need to mount anything) and clicking boot gave me a directory listing of /boot. I edited config.txt and changed the GPU Mem value to 256MB.

That fixed it and my PI is now booting quite happily again. I am now going to make a backup copy! It wouldn’t the end of the world if I had made it permanently unbootable, I’d just burn the OS again. It’s just the time wasted and minor hassles copying files, downloading and reinstalling software. Best avoided if possible!