James Newton's questions

I have two repos for a web site: one for the backend, one for the frontend. I want to copy files from the frontend to the backend directory, so that everything can be served from one place.

When I run the build command on the frontend, a folder named dist/ is created. I want to copy all these files to a folder with the relative path ../backend/public. This public folder may already contain some other files.

My current script uses hard-coded names for index.html and and assets/ directory, but more files may be added over time, so I would like a generic solution that will copy everything that has not changed in dist/ to ../backend/public.

#!/bin/bash
cd ../backend/public
rm -rf index.html assets

cp ../../frontend/dist/index.html ./
cp -r ../../frontend/dist/assets ./assets

I want to write a command in ~/.bashrcthat will create a trigger file that will open a different target file in a different folder when I double-click on the trigger file. I want the target file to open in VS Code.

The files might appear in a directory structure like this:

parentDir/
  archiveDir/
  yesterdaysDateDir/
  todaysDateDir/
    fileIWantToOpen.js

  index/
    fileIWantToDoubleClick

I understand that I can create an executable .desktop file to do this. I have found the specifications for .desktop files, but I have got lost in all the possible details.

I think that what I want might look something like this...

[Desktop Entry]
Type=Application
Terminal=true
Name=fileIWantToDoubleClick
Icon=utilities-terminal
Exec="code '../todaysDate/fileIWantToOpen.js'"
Categories=Application;

... but I am out of my depth. Any help in understanding what to do would be appreciated.

Creating a simple symbolic link does not work for me: I want to click on a file in my index directory and have it open a specific file in a specific folder, rather than launch a clone with a different name from the index directory.

I use a Dell XPS 13 9350 laptop. I recently purchased an HP EliteDisplay E223d monitor. Both the laptop and the monitor have native resolution of 1920 x 1080 pixels. The monitor is connected to the laptop directly via a Thunderbolt cable.*

When I run sudo lshw -C video the output is:

*-display                 
   description: VGA compatible controller
   product: Skylake GT2 [HD Graphics 520]
   vendor: Intel Corporation
   physical id: 2
   bus info: pci@0000:00:02.0
   version: 07
   width: 64 bits
   clock: 33MHz
   capabilities: pciexpress msi pm vga_controller bus_master cap_list rom
   configuration: driver=i915 latency=0
   resources: irq:139 memory:db000000-dbffffff memory:90000000-9fffffff ioport:f000(size=64) memory:c0000-dffff

A Dell page that refers to the Skylake GT2 [HD Graphics 520] video card shows that it is capable of handling up to four times the native resolution of each of my screens:

DP 1.2/eDP 1.3 max. resolution  3840 x 2160 @ 60 Hz

While a fresh installation of Ubuntu 20.04 is loading, the entire second monitor shows the purple background colour, but as soon as Ubuntu is running, the second monitor displays with a maximum resolution of 1024 x 768 px:

When used with other operating systems, the monitor has worked at full resolution.

Another issue, which may have a different cause, is that the monitor often tells me:

⚠️ Input Signal Not Found
Check the video cable and video source.
Display going to sleep.

When used with other operating systems, the monitor has worked at full resolution.

What steps should I take to get my second monitor to work correctly with Ubuntu 20.04?

From page 3 of the specifications for my laptop, I see

*Thunderbolt 3 (USB Type‑C) port Supports USB 3.1 Gen 2, DisplayPort 1.2 , Thunderbolt 3 and also allows you to connect to external display using dongles. Provides data transfer rates up to 10 Gbps for USB 3.1 Gen 2 and up to 40 Gbps for Thunderbolt 3.

Does this mean that a dongle is required, or that dongles are also supported? The monitor (sometimes) functioned at full resolution under Ubuntu 16.04, without a dongle.

On Ubuntu 16.04, I have umake installed at /usr/bin/umake. When I run umake --version it reports

18.05

I just used snap install ubuntu-make --classic to install the latest version, and this is installed at /snap/bin/ubuntu-make.umake. When I run ubuntu-make.umake --version it reports

20.04+snap678

I had expected the newer version to overwrite the original and to create an alias for itself at /usr/bin/umake, but now I have two copies of umake in two different places. I have tried sudo apt remove umake, but this does nothing useful:

Reading package lists... Done
Building dependency tree       
Reading state information... Done
E: Unable to locate package umake

Sanity check: Does it make sense to remove the older version and to create a symbolic link for the newer version at usr/bin/umake? If so, how do I uninstall version 18.05?

In Ubuntu 16.04, I (eventually) want to drop a text file onto a desktop icon and get a Node.js app to run, treating the dropped file and outputting a new file.

I have discovered .desktop files, and am doing my best to understand how to get the process to work, using a barebones set-up. I currently have three files: two files in a directory at /home/blackslate/Utilities/ and one on my desktop.

Utilities files

A text file named convert.js:

const fs = require('fs');
const crypto = require('crypto');
const random = crypto.randomBytes(4).toString('hex');
const filename = random + '.txt'

fs.appendFile(filename, 'Hello ' + random + '!', function (err) {
  if (err) throw err;
  console.log(filename + ' saved');
});

An executable file named convert.sh:

#!/bin/sh
node ./convert.js

gnome-terminal -e "bash -c 'echo hello world; sleep 3'"

Desktop file

An executable file named convert.desktop:

[Desktop Entry]
Exec=gnome-terminal -e "/home/blackslate/Utilities/convert.sh \"%u\""
Icon=utilities-terminal
Type=Application
Name=Convert

Here's what happens:

1. If I call node ./convert.js from a Terminal window, I see that:

   • A new file is created and its name is shown in the Terminal window

2. If I call ./convert.sh from a Terminal window, I see that:

   • A new file is created and its name is shown in the same Terminal window
   • A second Terminal window is opened and says hello world before closing 3 seconds later

3. If I click on the Convert desktop file, I see that:

   • One Terminal window opens, showing a bunch of text, and then it closes immediately
   • A second Terminal window is opened and says hello world before closing 3 seconds later
   • No new text file is created

It is obvious that the second command in the convert.sh file is being executed, but it appears that something is going wrong with the first one. What steps can I take to debug and fix this?

I have a folder named Meteor. In Unity, when I right-click and display the Properties for the folder, it tells me "2,734 items, totalling 6.1 MB". However, inside this is a hidden folder named .meteor. The Property dialogue for this folder claims "673 items, totalling 28.0 MB".

How can a child folder be bigger than its parent? Does the Properties dialogue not display the contents of hidden folders, even when Unity is set to display them?

Is there a special key combination that I can press while choosing Properties, to make it show the real total?

In Ubuntu 16.04.3, I have a very simple bash script:

test.sh

[[ 0 == 0 ]] && result="true" || result="false"
echo $result
echo $USER $SHELL $0

When I call it as the non-root user me or as root, it works as expected. If I use sudo ./test.sh, it complains about a syntax error:

$ ./test.sh
true
me /bin/bash ./test.sh

$ sudo su
# ./test.sh 
true
root /bin/bash ./test.sh

# exit
$ sudo ./test.sh
./test.sh: 1: ./test.sh: [[: not found
false
root /bin/bash ./test.sh

What could be causing this? How can I fix it so that me can use this script both normally and with sudo?

I have one bash script that is designed to output another bash script. I am using cat. I want to evaluate some of the expressions in output script, but not others.

PROJECT=myproject

cat << EOF > create_dir.sh
#!/usr/bin/env bash

DATE=`date '+%y%m%d-%H%M'`
mkdir $PROJECT/\$DATE
EOF

The resulting create_dir.sh file looks like this:

#!/usr/bin/env bash

DATE=171123-1834

mkdir myproject/$DATE

The result that I want is this:

#!/usr/bin/env bash

DATE=`date '+%y%m%d-%H%M'`

mkdir myproject/$DATE

How can I modify this script so that the expression following DATE= is not evaluated, while at the same time ensuring that $PROJECT is evaluated?

I am writing a bash script to install a number of packages. Do I need to call apt-get -y update each time before I call apt-get install <package_name>? Or is it enough to call apt-get -y update once at the beginning of the script?

If it needs to be called multiple times, could you explain why?

The problem
In a bash script, I am using the command...

sudo -u node bash

.. to switch from root to a non-sudo user, and this is failing.

The context
I am writing a provision.sh script for Vagrant, to set up a server running Ubuntu 16.04.3 with all the packages required to deliver an app with Meteor 1.6.

One required step is to install nvm as a non-root user. After nvm is installed, you need to log out and log back in again in order to activate nvm. I therefore create a non-sudo user named meteor, and want to switch to that when I download and install nvm.

Subsequently, I want to switch back to being root and log in gain immediately as meteor, in order to start using nvm to install Node.js.

You'll find a heavily commented script below. Vagrant runs this script each time I call vagrant reload --provision.

What command should I be using instead of sudo -u node bash?

echo "# whoami" && whoami && echo "^^^^ root expected"
echo "As root, create non-sudo user meteor:"
pass=$(perl -e 'print crypt($ARGV[0], "password")' $password)
useradd -m -p $pass meteor
echo "User meteor created. ls -al /home/meteor:"
ls -al /home/meteor
echo "Install curl as root:"
apt-get install -y curl
echo "Trying sudo -u meteor bash"
sudo -u meteor bash #### THIS IS THE LINE THAT FAILS ###

echo "$ whoami" && whoami && echo "^^^^^^ meteor expected"
curl -o- https://raw.githubusercontent.com/creationix/nvm/v0.33.6/install.sh | bash
echo "ls -al /home/meteor/.nvm # should be populated"
ls -al /home/meteor/.nvm

echo "ls -al /root/.nvm # should not exist"
ls -al /root/.nvm

echo "command -v nvm will fail silently until we log out and back in again"
#command -v nvm
exit

#### Because the script is still running as root, it halts here ####

echo "# whoami" && whoami && echo "^^^^ should be root"
sudo -u meteor bash
echo "$ whoami" && whoami && echo "^^^^^^ should be meteor"
echo "command -v nvm should work now"
command -v nvm

I have a 128 GB USB key on which I would like to install Ubuntu 16.04.3 as a Persistent LiveUSB.

I have tried three different solutions, and run into difficulties each time:

• LiveLinux on Windows 10 limits the size of the persistent sector to 4 GB, and, worse, the USB key is not recognized as a bootable drive
• mkusb decided that Zenity would not work in the environment provided by my (non-persistent) LiveUSB key, and when I tried to use version 11 instead, it failed to mount the USB key after installation.

Are there any other solutions available? Or is there a step-by-step Idiot's Guide that will help me troubleshoot one of the above methods, so that I can coax it into working for me?

Or is it simply that 128 GB is too big?

Ideally, I would like to have a partition on the USB key which I can use for copying files to and from Windows machines, but right now, I will settle for anything that allows me to keep my preferred key board layout and other personal settings from one startup to the next.

UPDATE: I have successfully installed a Persistent LiveUSB version of Ubuntu on an 8 GB USB key using mkusb. First, I had to create a non-persistent version using Ubuntu 16.04.3, where mkusb was able to launch Zenity. * The built-in Startup Disk Creator did create a working LiveUSB key, but it simply has a button for Make Startup Disk which seems to provide no persistence.

A year ago, I bought a Dell XPS with Windows 10 installed in the factory. When it arrived, I added new partitions to the internal hard disk — boot, home and swap — and installed Ubuntu 16.04 LTS to create a dual-boot system. At that time, I followed the official instructions from Dell on how to install Ubuntu, with two changes:

1. The Dell instructions say to choose SHIMx64.EFI, but this is not available. The choice I have is between \EFI\BOOT\BOOTx64.EFI and \EFI\BOOT\grubx64.efi. I believe that I chose grub64.efi
2. I set the SATA Operation from RAID On to AHCI.

After this, I was able to install Ubuntu 16.04.1, and everything worked until today.

Today, I was offered a firmware update, which I accepted. Presumably this update was designed for a Windows-only machine. Now, when the laptop boots, it no longer recognizes the Ubuntu operating system on the internal SSD. After running a system scan (F12 during start-up), I have cleared up a partition error, and after that the Boot Sequence menu in the BIOS appeared like this:

[✓] UEFI: THNSN5256GPU7 NVMe TOSHIBA 256 GB, Par(tition 1)
[✓] UEFI: SanDisk Partition 1 
[✓] Windows Boot Manager  

SanDisk is the LiveUSB key that I am using.

To this, I have added a new Boot Option [✓] Ubuntu, which uses \EFI\BOOT\grubx64.efi, as before. I put this as the first in the list.

I have disabled Secure Boot, but I do not think that this was an issue before.

In the SATA Operation Panel, if I select AHCI, the laptop boots into the SanDisk LiveUSB. If I set it to RAID On, and disable the SanDisk in the Boot Sequence, then the laptop boots into Windows.

In no case does it boot into the Ubuntu OS that is installed on the 256 GB internal SSD.

What steps do I need to take to make the laptop capable of booting into the installed version of Ubuntu?

I want to find all PHP scripts that are inside a directory named web. In particular, I'm interested in any files named index.php in any of these directories.

I know that I can use sudo find / -type d -name "web" to list all the directories, and then I can use sudo find <directory path> -type f -name "*.php" for each of the output paths, to see if it contains a PHP script.

But is there a way to do this in one line and two nested find commands: one find command used to get the paths to the web directories, and the second to use each of those directories in turn as the root for finding a file whose name matches the pattern *.php?

My goal is to provide a secure email service from a new Ubuntu 16.04 server. In order to do this, I want to set up Apache2 with nginx running as its reverse proxy.

My immediate problem is that Apache fails to start, claiming that there is no SSL certificate assigned (see output from Terminal at the end of this question), even though I am not (yet) attempting to use HTTPS.

My questions:

• If Apache is running on a local port, is it not enough that nginx knows about any SSL certificates?
• What do I need to do to get Apache to start listening on at a local address?

Here is my understanding of what I need to do. I would be grateful if you could point out where my understanding is mistaken.

• Install nginx and Apache2
• Create a nginx server block to:
  • Listen on ports 80 (and 443)
  • (Redirect traffic from port 80 to port 443)
  • (Deal with an SSL certification)
  • Forward any requests for a PHP file to Apache at 127.0.0.1:8080
• Set up Apache to:
  • Listen only on port 8080, not on port 443
  • Remain ignorant of any SSL certification

I like to deal with one problem at a time, so I am currently leaving aside the question of SSL certificates and attempting to get Apache running with nginx as its reverse proxy on port 80 only.

Here are the configuration files that I have in place:

nginx

$ sudo nano /etc/nginx/sites-available/webmail

server {
  listen   80;

  root /var/www/webmail/web;
  index index.php index.html index.htm;

  server_name webmail.mydomain.com;

  # Look for...
  # * the exact path
  # * a default (index.*) file, considering the path to be a directory
  # ... and if that fails:
  # * get the index.php script at the root to deal with the request

  location / {
    try_files $uri $uri/ /index.php;
  }

  # If the chosen path leads to a PHP fie:
  # * forward the request to Apache running at 127.0.0.1:8080
  # * after having modified certain headers

  location ~ \.php$ {
    proxy_set_header X-Real-IP  $remote_addr;
    proxy_set_header X-Forwarded-For $remote_addr;
    proxy_set_header Host $host;
    proxy_pass http://127.0.0.1:8080;
  }

  # If there are any Apache HT files in the chosen directory
  # ignore any direct requests for them

  location ~ /\.ht {
    deny all;
  }
}

Apache2

$ sudo nano /etc/apache2/sites-available/webmail.conf 

<VirtualHost 127.0.0.1:8080>
        ServerName webmail.mydomain.com
        ServerAdmin [email protected]
        DocumentRoot /var/www/webmail/web
        <Directory "/var/www/webmail/web">
            Options FollowSymLinks
            AllowOverride All
        </Directory>

        ErrorLog ${APACHE_LOG_DIR}/error.log
        LogLevel warn
        CustomLog ${APACHE_LOG_DIR}/access.log combined
</VirtualHost>

Apache2 ports

$ sudo nano /etc/apache2/ports.conf

Listen 127.0.0.1:8080

#<IfModule ssl_module>
#       Listen 443
#</IfModule>

#<IfModule mod_gnutls.c>
#       Listen 443
#</IfModule>

Note that I am explicitly not asking Apache to listen on port 443. Is there another place in some other config file that is telling Apache to listen on port 443?

Here are my placeholder PHP and HTML files, in their expected places:

index.php

$ sudo nano /var/www/webmail/web/index.php

<?php
  echo "PHP from /var/www/webmail/web/index.php\n";
?>

index.html

$ sudo nano /var/www/webmail/web/index.html

<html>
  HTML from /var/www/webmail/web/index.html
</html>

In the browser, a request to http://webmail.mydomain.com/index.html succeeds as expected, but a request to http://webmail.mydomain.com/index.php results in the file being downloaded, not executed.

Here's what happens when I try to start Apache2:

$ sudo /etc/init.d/apache2 restart
[ ok ] Restarting apache2 (via systemctl): apache2.service.

$ sudo service apache2 status
* apache2.service - LSB: Apache2 web server
   Loaded: loaded (/etc/init.d/apache2; bad; vendor preset: enabled)
  Drop-In: /lib/systemd/system/apache2.service.d
           `-apache2-systemd.conf
   Active: inactive (dead) since <Time>; 3s ago
     Docs: man:systemd-sysv-generator(8)
  Process: 16059 ExecStop=/etc/init.d/apache2 stop (code=exited, status=0/SUCCESS)
  Process: 16043 ExecStart=/etc/init.d/apache2 start (code=exited, status=0/SUCCESS)

systemd[1]: Starting LSB: Apache2 web server...
apache2[16043]:  * Starting Apache httpd web server apache2
apache2[16043]: Action 'start' failed.
apache2[16043]: The Apache error log may have more information.
apache2[16043]:  *
apache2[16059]:  * Stopping Apache httpd web server apache2
apache2[16059]:  *
systemd[1]: Started LSB: Apache2 web server.

Here's the contents of the error.log:

$ sudo nano /var/log/apache2/error.log

[ssl:emerg] [pid 15943] AH02572: Failed to configure at least one certificate and key for myhostname.hostingservice.com:443
[ssl:emerg] [pid 15943] SSL Library Error: error:140A80B1:SSL routines:SSL_CTX_check_private_key:no certificate assigned
[ssl:emerg] [pid 15943] AH02311: Fatal error initialising mod_ssl, exiting. See /var/log/apache2/error.log for more information

(This is in var/log/apache2/error.log, and there is no more information.)

What have I been missing?

NOTE: The primary purpose of this server is to deliver a Meteor app. I have chosen to do this using Phusion Passenger and Nginx. In addition to the email service, Apache will also be used to deliver a WordPress site.

On my Ubuntu 16.04 LTS server, I would like to do the following things:

• Enable a non-root admin user with sudo privileges to SSH in to the server using an RSA key (no password)
• Enable selected non-admin users to upload files by SFTP to their own home directory, using a password to log in
• Prevent the non-admin users from gaining access to the rest of the file system

I'm working with a freshly-installed version of Ubuntu 16.04.3 LTS, so everything is in its default factory condition.

I have read this question and the answers carefully, but I have not been able to find a solution.

I have created a nonrootadmin user with sudo privileges.

I have created a nonadminsftp user who is a member of the sftpaccess group. The /home/nonadminsftp/ directory looks like this:

$ ls -al ~nonadminsftp
total 24
drwxr-xr-x 3 root         root       4096 Oct 25 00:52 .
drwxr-xr-x 5 root         root       4096 Oct 24 22:29 ..
-rw-r--r-- 1 nonadminsftp sftpaccess  220 Sep  1  2015 .bash_logout
-rw-r--r-- 1 nonadminsftp sftpaccess 3771 Sep  1  2015 .bashrc
drwxr-xr-x 3 nonadminsftp sftpaccess 4096 Oct 25 00:50 ftp
-rw-r--r-- 1 nonadminsftp sftpaccess  655 May 16 13:49 .profile

Their respective entries in /etc/passwd are as follows:

nonrootadmin:x:1000:1000::/home/nonrootadmin:/bin/bash
nonadminsftp:x:1002:1002::/home/nonadminsftp:/usr/sbin/nologin

The changes I have made to the /etc/sshd/sshd_config file are as follows:

PermitRootLogin no
#PasswordAuthentication no

AllowUsers nonrootadmin nonadminsftp
Subsystem sftp internal-sftp
Match group sftpaccess
#ChrootDirectory %h
X11Forwarding no
AllowTcpForwarding no
#ForceCommand internal-sftp

The solution that I have seen proposes uncommenting 3 of these lines, but I discover that:

• PasswordAuthentication no prevents the nonadminsftp user from connecting with a password:

  $ sftp [email protected]
  Permission denied (publickey).
  Couldn't read packet: Connection reset by peer
  

• ChrootDirectory %h prevents the nonrootadmin user from connecting at all:

  $ ssh [email protected]
  packet_write_wait: Connection to 12.34.56.78 port 22: Broken pipe`
  

• ForceCommand internal-sftp prevents the nonrootadmin from gaining SSH access:

  $ ssh [email protected]
  This service allows sftp connections only.
  Connection to mydomain.com closed.`
  

With these lines commented out:

• nonrootadmin does have SSH access using an RSA key
• nonadminsftp can connect using a FTP client such as FileZilla

BUT:

• nonadminsftp is not chrooted to the /home/nonadminsftp directory
• nonrootadmin can log in with a password

What is it that I am missing?

Thanks in advance

This question follows on from this one about natural scrolling.

Before I made changes, I could simply tap on the trackpad of my Dell XPS 13 to generate a click. Now I have to explicitly click, which requires extra effort. I would like to restore tap-to-click if that is possible.

I have a Dell XPS 13 which, by default, has a very sensitive trackpad. As you are typing, if you brush the trackpad by accident, the trackpad considers it as a click and the text insertion point jumps to wherever the mouse happens to be. I found and installed this fix for this issue:

sudo apt-get install xserver-xorg-input-libinput

This inverted the scrolling direction, so I also edited the entry for MatchIsTouchpad in /usr/share/X11/xorg.conf.d/90-libinput.conf:

Section "InputClass"
        Identifier "libinput touchpad catchall"
        MatchIsTouchpad "on"
        MatchDevicePath "/dev/input/event*"
        Driver "libinput"
        Option "NaturalScrolling" "true"
EndSection

With this new setup, tap-to-click does not happen. Is this inevitable? Is this the only way to prevent accidentally brushing the trackpad from generating a click? Or is it possible to regulate the sensitivity so that a deliberate tap is treated differently from a fleeting touch?

I have a Dell XPS 13 which, by default, has a very sensitive trackpad. As you are typing, if you brush the trackpad by accident, the trackpad considers it as a click and the text insertion point jumps to wherever the mouse happens to be. I found and installed this fix for this issue:

sudo apt-get install xserver-xorg-input-libinput

However, now the scrolling is unnatural. When I use the trackpad to scroll a long document, it is the viewport that moves over the content. I want the content to move in the same direction as my fingers, just as it did before, just like on a touchscreen.

Can you explain to me what changes I should make to the files at /usr/share/X11/xorg.conf.d/ that will make this happen?

The solution proposed here did not help.

I have set up a Host-only network in VirtualBox, and created a brand new VM with Ubuntu 14.04.5 Server installed. My goal is to give the Ubuntu server a static IP address which will not change if I move to a different location with a different physical network.

In VirtualBox, using File > Preferences > Network > Host-only Networks, I created a new network: vboxnet0

Adapter

IPv4 Address: 10.1.1.1
IPv4 Network Mask: 255.255.255.0
IPv6 Address: fe80:0000:0000:0000:0800:27ff:fe00:0000
IPv6 Network Mask Length: 6


DHCP Server
[x] Enable Server

Server Address: 10.1.1.100
Server Mask: 255.255.255.0
Lower Address Bound: 10.1.1.101
Upper Address Bound: 10.1.1.254

(I don't need to use the DHCP server, so I have tried with it both enabled and disabled, with the same lack of success.)

I set the Network for the Ubuntu server VM to:

Adapter 1: Bridged Adapter, wlan 0
Adapter 2: (enabled) Host-only Adapter, vboxnet0 (default settings)

When I launch the server and run ifconfig, I see only entries for eth0 with an inet addr of 192.168.1.18 (provided by the Bridged Adapter) and lo with an inet addr of 127.0.0.1.

When using a Desktop version of Ubuntu 14.04, I also see an entry for eth1, with an inet addr in the range 10.1.1.x, and I can configure this to be the static address that I want. With Ubuntu 14.04 Server, I do not see the eth1 network.

In the Ubuntu server, I have edited /etc/network/interfaces as follows:

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet dhcp

auto eth1
iface eth1 inet static
    address 10.1.1.2
    netmask 255.255.255.0

I have also tried eth1 with various combinations of:

    network 10.1.1.0
    gateway 10.1.1.1
    broadcast 10.1.1.25

When I run sudo ifup eth1, the output is:

RTNETLINK answers: File exists
Failed to bring up eth1

Does this mean that the Ubuntu virtual machine cannot see the 10.1.1.x network? Is there a command that allows me to check what networks are available?

Is there something else I need to set somewhere to make the Host-only Network visible for the VM?

I have used Ubuntu from a live USB drive before, but now I want to install it permanently. I am looking for step-by-baby-step instructions, so that I can understand exactly where I have gone wrong.

I have just bought a Dell XPS13 running Windows 10, and I want to install Ubuntu to get a dual-boot laptop. I used Windows Disk Manager to shrink the Windows partition, and leave unallocated space. I follow the instructions from the Dell Knowledge Base until I reach Figure 6. At this point, I see the options ...

BOOTx64.EFI
grubx64.efi

... in the Files panel, but there is no sign of SHIMx64.EFI. I have been able to boot into Ubuntu using grubx64.efi, but when I get to the Install Type screen, I do not see the options shown in this screenshot.

Instead, the installer jumps directly to a screen similar to the partition table, but the table is completely blank, only /dev/sda is visible in the Device dropdown, and none of the buttons react ... except for Install Now, which complains (understandably) that no root is defined. Just before the Installation Type screen appears, the WiFi connection dies and "No network devices available" appears in the WiFi menu.

I have tried using both 14.04 and 16.04, and both Rufus and Universal-USB-Installer to create my bootable USB drive.

Is there a step that I have missed that will display SHIMx64.EFI as an option at the EFI Boot Selection step? If grubx64.efi works to boot into Ubuntu, then is there a step I have missed that will allow me to select the unallocated partition?

Thanks in advance for all insights that you can share.

EDIT: the log from Rufus 2.11.995 shows only two lines containing the string "shim" when used with ubuntu-16.04.1-desktop-amd64.iso:

Extracting: F:\pool\main\s\shim\shim_0.8-0ubuntu2_amd64.deb (432.6 KB)
Extracting: F:\pool\main\s\shim-signed\shim-signed_1.17~16.04.1+0.8-0ubuntu2_amd64.deb (307.6 KB)

The same is true for ubuntu-14.04.5-desktop-amd64.iso:

Extracting: F:\pool\main\s\shim\shim_0.8-0ubuntu2_amd64.deb (432.6 KB)
Extracting: F:\pool\main\s\shim-signed\shim-signed_1.18~14.04.1+0.8-0ubuntu2_amd64.deb (307.2 KB)

I have Sony VAIO Tap 20 with Windows 8.1 installed. I want to install Ubuntu 14.04 or 15.04 on it, but I am failing right from the beginning. I cannot get Ubuntu to launch.

One particularity is that it is a French version of Windows 8.1, but I doubt that this should be an issue.

On a USB key with 7.7 GB of free space, I have created a Live USB of Ubuntu 14.04, using the instructions I find here.

• If I run wubi.exe before rebooting the computer, then I get an error saying that /ubuntu/winboot/wubildr.mbr is missing or contains errors.
• If I uninstall using Control Panel > Uninstall a program > Ubuntu, then I get no option to boot from the Live USB drive. The computer boots directly into Windows 8.1
• If I try to run the wubi.exe on the 15.04 iso, then it tells me "Wubi does not currently support EFI"

I understand that Windows 8.1 runs in UEFI mode, and that I should install Ubuntu in UEFI mode too.

I have ventured into the BIOS and disabled secure boot. I have switched off fast startup. I have shrunk the Windows partition, and created a 3 partitions for /, /home, and /swap... but I since I cannot get the computer to boot into Ubuntu from a Live USB, there is no way for me to install it onto these partitions.

I have already read this article (among many others), but I cannot get past "Boot a live ... USB flash drive in EFI mode".

On this page, I see a suggestion that I should disable Intel Smart Response Technology, but I did not find any clear steps on how to do this, while I did find reports that Ubuntu runs happily on systems with SRT enabled.

Is there a vital step that I have missed?

EDIT: Following the recommendations of @Rod Smith, I have used Rufus to create a bootable USB drive, using the iso for Ubuntu 14.04, with a GPT partition for UEFI.

I had imagined that it would be enough to reboot the computer with the USB drive in place, and that I would be offered a choice of Ubuntu or Windows. However, the computer simply boots straight into Windows without giving me a choice, even when I tell the BIOS to give first priority to the external drive.

What do I need to do to tell the computer to look at the alternative operating system on the USB stick?