Magnus's questions

I am trying to boot Ubuntu Server 64-bit 20.04 from a USB SSD thumb drive on a RPi3B+. I have found that on three separate RPi3B+ units, they ALL fail to find the drive if it is placed in either of the two USB slots NOT right next to the ethernet socket.

I have also tried with a crappy flash drive, which succeeds on booting in an even more inconsistent manner.

Are there any suggestions as to why USB booting of Ubuntu Server 20.04 seems to be so inconsistent on the RPi3B+? Any remedies to this?

It boots for RPi4 from USB device, so I assumed it would also for the RPi3B+. However, I have not had any luck so far. Is there some extra step needed, or do I fundamentally need to boot from SD card and mount the file system separately? (I'm surprised how little comes up when googling this.)

Hot plugging of HDMI seems to work well on Raspberry Pi OS (see here). I tried to achieve the same with Ubuntu Server 20.04 on my RPi3B+ by adding the same lines to /boot/firmware/usercfg.txt:

hdmi_force_hotplug=1
hdmi_group=1
hdmi_mode=16

... but it doesn't seem to work. Is there a way to achieve hot HDMI switching on Ubuntu Server 20.04 for RPi series?

I am trying to install blender 2.8 from here. I've run:

sudo add-apt-repository ppa:thomas-schiex/blender
sudo apt-get update

However, when I run apt search blender, it still shows me the version 2.79 version from the official bionic-updates repo.

I would try and specify the exact repo, but this ppa names the release as bionic main, i.e. super generic terms that overlap with the official names.

I've tried removing the sources.list file so that only the blender package in the ppa repo would get found, but that doesn't work either (Blender is simply not found at all). It seems as though the ppa is not seen by my system.

Edit: Result of apt-cache policy blender:

blender:
  Installed: (none)
  Candidate: 2.79.b+dfsg0-1ubuntu1.18.04.1
  Version table:
     2.79.b+dfsg0-1ubuntu1.18.04.1 500
        500 http://ports.ubuntu.com/ubuntu-ports bionic-updates/universe arm64 Packages
     2.79.b+dfsg0-1 500
        500 http://ports.ubuntu.com/ubuntu-ports bionic/universe arm64 Packages

Edit2: Running sudo apt update confirms that we're hitting the ppa:

...
Hit:9 http://international.download.nvidia.com/jetson/repos/common r32 InRelease
Hit:10 http://international.download.nvidia.com/jetson/repos/t210 r32 InRelease
Hit:12 http://ppa.launchpad.net/thomas-schiex/blender/ubuntu bionic InRelease
Hit:13 http://ports.ubuntu.com/ubuntu-ports bionic InRelease
Hit:14 http://ports.ubuntu.com/ubuntu-ports bionic-updates InRelease
Hit:15 http://ports.ubuntu.com/ubuntu-ports bionic-backports InRelease
Hit:17 http://ports.ubuntu.com/ubuntu-ports bionic-security InRelease
Reading package lists... Done
Building dependency tree
Reading state information... Done
All packages are up to date.

... but it's still showing only version 2.79 as an option.

I'm new to nmcli and trying to figure out how to create an eth0 connection. sudo nmcli con prints the following:

NAME       UUID     TYPE      DEVICE
MyWlan0    ...      wifi      wlan0
MyEth0     ...      ethernet  --

No matter what command I give, my MyEth0 connection will not show eth0 under DEVICE above (just '--') and, as a result, I'll get the following error when I try sudo nmcli con up MyEth0:

Error: Connection activation failed: No suitable device found for this connection (device lo not available because device is strictly unmanaged).

This is how I created the connection:

sudo nmcli con add con-name MyEth0 type ethernet ifname eth0

I've also tried modifying it with the following sorts of commands:

sudo nmcli con mod MyEth0 connection.interface-name eth0

Running sudo nmcli con show MyEth0 connection | grep eth0 does indicate that the connection.interface-name is eth0, but it does not appear this way according to the behavior described above.

Any ideas?

Edit:

Here is the output of sudo lshw -C network:

  *-network:0
       description: Ethernet interface
       physical id: 2
       logical name: eth0
       serial: dc:a6:32:27:84:45
       size: 100Mbit/s
       capacity: 1Gbit/s
       capabilities: ethernet physical tp mii 10bt 10bt-fd 100bt 100bt-fd 1000bt 1000bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=bcmgenet driverversion=v2.0 duplex=full link=yes multicast=yes port=MII speed=100Mbit/s
  *-network:1
       description: Wireless interface
       physical id: 3
       logical name: wlan0
       serial: dc:a6:32:27:84:46
       capabilities: ethernet physical wireless
       configuration: broadcast=yes driver=brcmfmac driverversion=7.45.18 firmware=01-6a2c8ad4 ip=192.168.1.201 multicast=yes wireless=IEEE 802.11

I am running ubuntu server 19.10 on a raspberry pi. My pi is connected to a wifi router and to an ethernet switch, with interface names wlan0 and eth0 respectively.

I want to create a static ip address for my wlan0 connection. I'm used to raspbian and this is my first serious foray into ubuntu server.

When I look into /etc/netplan/, there is a single yml file with the contents:

network:
    ethernets:
        eth0:
            dhcp4: true
            optional: true
    version: 2

Since there's no mention of wlan0 in here, I assumed that ubuntu is not operating a dhcp client for wlan0. I therefore installed dhcpcd5 and configured it to ask for a static ip address (192.168.1.202). However, the wifi router now seems to be giving my device two ip addresses; one is the one I asked for in /etc/dhcpcd.conf, the other must be (I think) via netplan. Both work when I try ssh-ing to the device.

I want to figure out why I am getting two addresses and how to prevent it. When I print-out dhcp-relevant lines from /var/log/syslog, I get the following:

... dhcpcd[1173]: wlan0: carrier acquired
... dhcpcd[1173]: wlan0: IAID eb:25:ec:d9
... dhcpcd[1173]: wlan0: adding address XXX
... dhcpcd[1173]: wlan0: IAID eb:25:ec:d9
... dhcpcd[1173]: wlan0: adding address XXX
... NetworkManager[1162]: <info>  [1573233133.0889] dhcp4 (wlan0): activation: beginning transaction (timeout in 45 seconds)
... dhcpcd[1173]: wlan0: probing address 192.168.1.202/24
... dhcpcd[1173]: wlan0: probing address 192.168.1.202/24
... NetworkManager[1162]: <info>  [1573233133.1426] dhcp4 (wlan0): state changed unknown -> bound
... dhcpcd[1473]: wlan0: using static address 192.168.1.202/24
... dhcpcd[1473]: wlan0: adding route to 192.168.1.0/24
... dhcpcd[1473]: wlan0: no IPv6 Routers available

So it looks like dhcpcd5 is functioning as normal, but then there is this surprise entry from NetworkManager with dhcp4. Furthermore, when I try ifconfig wlan0, I find that my device thinks its wlan0 ip address is NOT the one I requested via dhcpcd5, which is surprising, since my most recent logs suggested that it was adding that route to my wifi router, but then it doesn't seem to write the result to my device locally.

I'd greatly appreciate clarity here, namely:

1a. Why doesn't netplan show info for wlan0 in /etc/netplan/*?

1b. How can I disable netplan's dhcp4 service?

2. General recommendations are welcome as to how best to set up static ip addresses. Is it a bad idea to install dhcpcd5 on ubuntu in place of its native set up? etc.

If I go into /etc/apache2/ and type grep -rl 'LoadModule php' ./* then I'll only see ./mods-available/php7.0.load. If I change r to R then I'll see

./mods-available/php7.0.load
./mods-enabled/php7.0.load

Both of these are plain directories. I'd expect -r to cause grep to recurse all files/dirs in the PWD. What's going on with -r? I'm minded to just always use -R at this point.