XavierStuvw's questions

I have upgraded Ubuntu 16.04 to Ubuntu 18.04. I do not intend to make a further upgrade to Ubuntu 20.xx for the moment. All looks to have gone well except the switching of users from within the desktop environment. My desktop environment was and is Unity.

There are two users in this computer. The expected behaviour is that within the desktop environment I can switch to any other profile without logging out of the current one, like in a multi-user system. In fact, I always have to log out from one profile to log into the other.

What's below happens always regardless of which profile is first.

1 Straight after the upgrade

1.1 Issue

The situation straight after the upgrade was:

1. I log in into both desktop profiles separately: PASS.
2. Once I am in either profile and click on the menu to access the other, nothing happens: FAIL.

So I stay where I am. I need to log out completely in order to access the other profile.

1.2 Partial fix

I followed the suggestion to reinstall the desktop environment with

sudo apt remove lightdm
sudo apt install lightdm
sudo reboot

This fixed the issue only partially.

1. I can log in into either profile separately: PASS, OLD.
2. If I click on the menu inside a profile to access any other profile, something happens: I am presented with the login screen with the password input: PASS, NEW.
3. The resolution of the login screen where to input the password is rather coarse: FAIL, NEW.
4. After I type in the password of the other profile, it hangs for several seconds before it loops back into the log-in screen.
   1. Therefore, I cannot access the other user profile: FAIL, NEW.
   2. However, I can log in back into the previous profile.

Other actions like

sudo apt install ubuntu-unity-desktop

or

sudo dpkg-reconfigure lightdm

have no effect or are redundant. Indeed, the same problem occurs regardless of whether the desktop environment is Unity or Ubuntu.

2 Current situation

2.1 Tests from the terminal

I can log in into both profiles from the terminal. After I am in the desktop environment of profile A, I move to a tty terminal and log in into profile B.

1. At the first login, before the prompt the system message is

        Could not find key with description: [SIGNATURE]
        Could not find process_request_key_err in user session keyring for sig specified
        Error parsing option; rc = [-2]
   

   Both tests keyctl show and ls pass: so I CAN see the session keyring (the not-found message above is nonsense) and the directory content (the message above is immaterial).

2. I exit and log in again. Before the prompt there is another message

        Signature not found in user keyring
        Perhaps try the interactive 'ecryptfs-mount-private'
   

   The test keyctl show fails (the not-found message above is consistent) and ls passes (the message above remains immaterial). Even if I try ecryptfs-mount-private, I load the key, the user switch problem persists, and these message occur again at the next reboot.

3. Sometimes these messages do not show up at all (so, plain sailing within the terminal), but the problem of switching accounts in the desktop environment persists.

In the terminal I can see and browse the contents of the home directories. In all cases there is no blockage due to encryption.

2.2 From the desktop environment

1. If I log out of the desktop environment of profile A and start afresh, the log-in screen with the password input is rendered neatly (no coarse resolution). I can log into profile B.
2. When I try to switch from profile A to B and have left behind the desktop environment of profile A without logging out, I can always come back to profile A by typing the user password. That is: it goes in the way I would expect to work for activating profile B as well.
3. If I create a new user C with administrator rights (a suggestion of @matigo), the following happens:
   1. I cannot switch from A and B to C precisely in the same way as I could not between A and B;
   2. I cannot log in the newly created C from the login screen either: the greeter screen returns to itself. The latter happens whether or not I have set a password for user C, so it's no human error.

2.3 Other posts

I have looked into these other posts, which refer to older releases of Ubuntu

• Unable to switch users in 13.10
• Ubuntu cannot switch user
• Switch user after upgrade to 14.04 not working
• Could not find valid key in user session keyring for sig specified in mount option after upgrade from 16.04 to 18.04

3. Analysis of errors in journalctl

This follows up a suggestion of lemrm. This is my situation:

wmctrl -m
Name: Compiz
Class: N/A
PID: N/A
Window manager's "showing the desktop" mode: OFF

echo $DESKTOP_SESSION 
unity

echo $GDMSESSION
unity

echo $XDG_CURRENT_DESKTOP
Unity:Unity7:ubuntu

3.1 Without libpam-kwallet*

The command journalctl -p err -b returns among others the following interesting lines on lightdm:

dec 10 21:33:59 computer lightdm[3992]: PAM unable to dlopen(pam_kwallet.so): /lib/security/pam_kwalle
dec 10 21:33:59 computer lightdm[3992]: PAM adding faulty module: pam_kwallet.so
dec 10 21:33:59 computer lightdm[3992]: PAM unable to dlopen(pam_kwallet5.so): /lib/security/pam_kwall
dec 10 21:33:59 computer lightdm[3992]: PAM adding faulty module: pam_kwallet5.so
dec 10 21:33:59 computer lightdm[3992]: pam_unix(lightdm:auth): conversation failed
dec 10 21:33:59 computer lightdm[3992]: pam_unix(lightdm:auth): auth could not identify password for ...

I found these resources

• 2017: Slow boot on Ubuntu 16.04 ... issue with pam_kwallet5.so.
• 2018: Is libpam-kwallet required/necessary in ubuntu 18.04?
• https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/1309535

The libraries pam_kwallet.so and pam_kwallet5.so belong to the packages

• libpam-kwallet4
• libpam-kwallet5

which are missing in my system. I do install them.

The interesting difference is that, when I move out to another profile via the menu, I do not splash again into the greeter with a choice of profiles, but in the "re-entry" page of the current profile (as though I had pressed Ctrl + L). I still need to log out to find a way into another profile. I can only come back to the profile I was in.

3.2 With libpam-kwallet*

The lines of journalctl -p err -b regarding lightdm become:

dec 10 22:16:23 computer lightdm[1766]: pam_kwallet(lightdm:session): pam_kwallet: Impossible to write walletKey to walletPipe
dec 10 22:16:23 computer lightdm[1767]: pam_kwallet5(lightdm:session): pam_kwallet5: Impossible to write walletKey to walletPipe

I have not found much on this Impossible to write walletKey to walletPipe that I could make use of:

1. the source code: https://github.com/KDE/kwallet-pam/blob/master/pam_kwallet.c
2. a complaint: https://forums.gentoo.org/viewtopic-p-8220820.html
3. a bug report: https://bugs.launchpad.net/lightdm/+bug/1781418

3.3 pam_ecryptfs

After I try to change user from the menu, as per 3.1, the journal has appends one more line to the messages in 3.2

dec 10 23:28:11 computer compiz[2490]: pam_ecryptfs: seteuid error

after which I check that:

dpkg -S pam_ecryptfs
ecryptfs-utils: /usr/share/man/man8/pam_ecryptfs.8.gz
ecryptfs-utils: /lib/security/pam_ecryptfs.so

This shifts the focus from lightdm and kwallet to compiz and ecryptfs. These resources apply:

• https://bugs.launchpad.net/ecryptfs/+bug/1085706 (confirmed and unassigned bug, first reported in 2012 and last reported in 2021)

4. Question

Therefore, the problem seems to be purely one of managing the switch from any first current profile into any other profile inside the desktop environment.

Basically, I always have to log out from one profile to log into the other. So no multi-user behaviour.

Any suggestions to fix this?

Purely as the result of the upgrade from Ubuntu 16 to 20.04LTS, my external webcam is not found by some software any longer.

Situation

1. uname -r says 5.4.0-54-generic (so this answer should not apply)

2. lsusb says: ` Bus 006 Device 007: ID 046d:08d9 Logitech, Inc. QuickCam IM/Connect

3. dmesg says

   [ 2440.365843] usb 6-1: new full-speed USB device number 7 using uhci_hcd  
   [ 2440.577885] usb 6-1: New USB device found, idVendor=046d, idProduct=08d9, bcdDevice= 1.00
   [ 2440.577888] usb 6-1: New USB device strings: Mfr=0, Product=0, SerialNumber=0
   [ 2440.580922] gspca_main: gspca_zc3xx-2.14.0 probing 046d:08d9
   [ 2441.795952] input: gspca_zc3xx as /devices/pci0000:00/0000:00:1d.0/usb6/6-1/input/input21
   

4. lsmod says

   ... 
   videobuf2_vmalloc      20480  1 gspca_main
   videobuf2_memops       20480  1 videobuf2_vmalloc
   videobuf2_v4l2         24576  1 gspca_main
   videobuf2_common       49152  2 gspca_main,videobuf2_v4l2
   videodev              225280  4 
   gspca_main,gspca_zc3xx,videobuf2_v4l2,videobuf2_common  
   ...
   

5. file /dev/video0 says /dev/video0: character special (81/0)

6. Cheese 3.34.0 responds from the start with

   

   and Skype even more plainly with

   

However, the web-conferencing system Big Blue Button does work (via Firefox 83.0).

The similar questions suggested seem date back to 2012-2014 and are not current any longer. I found this one closer to my case and used it as a blueprint to collect info, but its solution does not apply here.

Any tip for restoring full functionalities much appreciated.

I need to ask again this old-fashioned question because the internet is flooded by posts of people struggling with problems contrary to what I have.

UEFI is totally irrelevant here. Dual boots are totally irrelevant here.

The ingredients are:

• I have an old computer that only works in BIOS mode; it does not have a clue of what UEFI is and will not recognize anything about it (at least as far as I could understand);
• I want to install Ubuntu 16.04 (perhaps a later version) on it;
• legacy mode is what the computer wants.
• the system is 64bit

The directions I am asking for are about:

• which image of the Ubuntu distribution should I use?
• where do I find it?
• how should I partition and set up my USB?
• how should I copy the Ubuntu image in the USB?

The desired result is that, at boot time, plugging such an USB drive starts the Ubuntu installer.

Again, doing this with from the UEFI perspective works fine for me; I lack information on this corner 'retro' situation.

Situation

Using Ubuntu 16.04

I managed to activate the connection between a desktop computer and a Nokia 206 (with Bluetooth standard 2.1 EDR) mobile via a Bluetooth dongle.

Please take note:

• The hardware side looks fine.
• Sending individual files to the mobile works as expected.
• Sending individual files from the mobile works as expected.
• It is the capability of file browsing from the computer into the mobile that misses out.

Issue

I then make steps towards browsing the files on the mobile. It starts off well with a progress bar completing its swipe, but at the next step I get the message

Failed to launch default browser
The specified location is not supported
You can enter an alternate browser in service settings

Research

I found an earlier mention of this issue in Ubuntu 16.04 iPhone iOs 10 Bluetooth Devices not allowing Browse function (2016)

From the Blueman project https://github.com/blueman-project/blueman/wiki/Integrations#obex-file-transfer I found

blueman can use a file browser for browsing files on a remote device. The command for this can be customized in the local services settings.
Most file browsers provide this OBEX FTP support using gvfs which in terms makes use of obex-data-server. This works only with BlueZ 4 and was thus removed from gvfs completely in version 1.23.92.
blueman 2.1 will drop support for browsing files.
[last edited 2016]

Now: I have Bluez 5.37, gvfs 1.28.2, Blueman 2.0.4. Even though I cannot rely on gvfs, support for browsing should be still there -- see Blueman project changelog.

I found out the window Local Services down Bluetooth Devices:

Questions

Question 1 What could I set in the advanced options to activate the alternate browser?

Question 2 If none, are there alternatives to Blueman/Bluez for receiving files in the computer instead of sending them from the mobile one by one?

======================================

Appendix on libraries

This is the snapshot of the Synaptic search on the theme bluez, the tools to manage Bluetooth devices for linux:

Regarding obex that is OBject EXchange protocol

None of the obex-related tips in

• Ubuntu 17.04 Sending/Receiving files by Bluetooth not working
• Error : Could not display "obex://" when paired with a Nokia 3110c

helped

Situation

Ubuntu 16.04, skypeforlinux 8.53.0.85 and 8.58.0.93 (latest to date from the Skype website)

• The external webcam works with Cheese

• The external webcam is recognized by Skype as per screenshot below

  • the vendor and product ID given by lsusb show up in the Skype dialogue window (in hexadecimals)

  • the webcam is on but the screen is not activated: black screen

    

Research

This is not on the official list of known issues of Skype for Linux: https://support.skype.com/en/faq/FA34713/faq-and-known-issues-with-skype

A tack of old posts suggests checking whether the libraries v4l and gstreamer are installed. They are:

ii  libv4l-0:amd64  1.10.0-1  amd64  Collection of video4linux support libraries  
ii  libv4l-dev:amd64  1.10.0-1  amd64  Collection of video4linux support libraries (development files)
ii  libv4l2rds0:amd64  1.10.0-1  amd64  Video4Linux Radio Data System (RDS) decoding library
ii  libv4lconvert0:amd64  1.10.0-1  amd64 Video4linux frame format conversion library
ii  gir1.2-gst-plugins-base-1.0  1.8.3-1ubuntu0.3  amd64  GObject introspection data for the GStreamer Plugins Base library
ii  gir1.2-gstreamer-1.0  1.8.3-1~ubuntu0.1  amd64  GObject introspection data for the GStreamer library
ii  gnome-video-effects  0.4.1-3ubuntu1  all  Collection of GStreamer effects
ii  gstreamer1.0-alsa:amd64  1.8.3-1ubuntu0.3  amd64  GStreamer plugin for ALSA
ii  gstreamer1.0-clutter-3.0  3.0.18-1  amd64  Clutter PLugin for GStreamer 1.0
ii  gstreamer1.0-libav:amd64  1.8.3-1ubuntu0.2  amd64  libav plugin for GStreamer
ii  gstreamer1.0-plugins-base:amd64  1.8.3-1ubuntu0.3  amd64  GStreamer plugins from the "base" set
ii  gstreamer1.0-plugins-base-apps  1.8.3-1ubuntu0.3  amd64  GStreamer helper programs from the "base" set
ii  gstreamer1.0-plugins-good:amd64  1.8.3-1ubuntu0.4  amd64  GStreamer plugins from the "good" set
ii  gstreamer1.0-plugins-ugly:amd64  1.8.3-1ubuntu0.1  amd64  GStreamer plugins from the "ugly" set
ii  gstreamer1.0-plugins-ugly-amr:amd64  1.8.3-1ubuntu0.1  amd64  GStreamer plugins from the "ugly" set
ii  gstreamer1.0-pulseaudio:amd64  1.8.3-1ubuntu0.4  amd64  GStreamer plugin for PulseAudio
ii  gstreamer1.0-tools  1.8.3-1~ubuntu0.1  amd64  Tools for use with GStreamer
ii  gstreamer1.0-x:amd64  1.8.3-1ubuntu0.3  amd64  GStreamer plugins for X11 and Pango
ii  libclutter-gst-3.0-0:amd64  3.0.18-1  amd64  Open GL based interactive canvas library GStreamer elements
ii  libgstreamer-plugins-base1.0-0:amd64  1.8.3-1ubuntu0.3  amd64  GStreamer libraries from the "base" set
ii  libgstreamer-plugins-good1.0-0:amd64  1.8.3-1ubuntu0.4  amd64  GStreamer development files for libraries from the "good" set
ii  libgstreamer1.0-0:amd64  1.8.3-1~ubuntu0.1  amd64  Core GStreamer libraries and elements
ii  libreoffice-avmedia-backend-gstreamer  1:5.1.6~rc2-0ubuntu1~xenial10  amd64  GStreamer backend for LibreOffice
ii  phonon-backend-gstreamer:amd64  4:4.8.2-0ubuntu2  amd64  Phonon GStreamer 1.0 backend
ii  phonon-backend-gstreamer-common:amd64  4:4.8.2-0ubuntu2  amd64  Phonon GStreamer 1.0.x backend icons
ii  totem  3.18.1-1ubuntu4  amd64  Simple media player for the GNOME desktop based on GStreamer

Directing the launch towards the compatibility or conversion libraries either with

LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libv4l/v4l2convert.so /usr/share/skypeforlinux/skypeforlinux

or

LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libv4l/v4l1compat.so /usr/share/skypeforlinux/skypeforlinux

launches Skype but doesn't fix the issue. Note that /usr/bin/skype is a bash script launching /usr/share/skypeforlinux/skypeforlinux, which I have used above as test.

Question

Any suggestion on how to have Skype actually send signals through the webcam?

Situation

I have installed the nvidia driver in a computer run by Ubuntu 18.04 LTS with the GDM3 display. I followed the recommendation of the command ubuntu-driver devices, and the driver in point nvidia-graphics-driver-435 is available from the repository.

The installation sudo apt install nvidia-driver-435 goes well. If I type sudo lshw -c video the line about configuration shows that nvidia is the driver on duty.

Forewarned is forearmed

I had at reach the exhaustive contribution of @LuisAlvarado at https://askubuntu.com/a/61433/446253. In case the installation fails, he suggests the following two-stage troubleshooting strategy (boldface is mine).

If your desktop does not load after installing the corresponding driver, then do the following:

sudo nano /etc/gdm3/custom.conf

then remove the comment (# symbol) from the line that says

# WaylandEnable=false

and save. Then reboot. If this still does not work, then please disable Secure Boot since you might actually be using UEFI.

Troubleshooting step 1

If I reboot, I can log in but the desktop does not load up. I then comment in WaylandEnable=false (so the flag is active) as suggested and reboot. However, the bootstrap hangs with a purple screen; if I type F12 to leave the graphics environment, the console informs me "A start job is running for hold until boot process finished" and this proceeds 'endlessly'.

Troubleshooting step 2

Then I reboot and enter the boot loader. Contrary to expectations, SecureBoot is already disabled. It is enabling it that makes loading boot and desktop pass.

Undoing troubleshooting step 1

Once in my desktop environment, however, the system warns me of problems occurring. The logs in /var/crash regard gnome shell and XWayland. Hence, I comment out WaylandEnable=false again, undoing the first troubleshoot step. I reboot and all proceeds seamlessly.

The only strange thing is that, unlike earlier, sudo lshw -c video lacks the driver entry in the configuration line; silence; so no guarantee that nvidia is the driver on duty. I take reassurance that it is so from the tab Additional Drivers in the Software & Updates, and from

grep -iq nouveau /var/log/Xorg.0.log which fails

grep -iq nvidia /var/log/Xorg.0.log which passes.

Question

In my case there was no need to tweak the custom configuration of GDM3. Apparently, enabling Secure Boot took the pain away. I feel there is much folklore about disabling Secure Boot as the great healer for many installation troubles. However, I lack directions to orientate myself in how it should be and how it is. Would anyone please clarify how this works?

Situation

• Ubuntu 18.04.4 LTS
• ubuntu-desktop 1.417.4
• ubuntu-unity-desktop 0.1
• lightdm 1.26.0-0ubuntu1

Action

Starting off from a GNOME enviroment, I have run sudo dpkg-reconfigure lightdm, chosen lightdm as default, exited with no error message, and rebooted. Indeed, cat /etc/X11/default-display-manager gives /usr/sbin/lightdm.

Issue

The new greeter is lightdm/Unity-style, but the eventual desktop environment is GMD again (all works in it). This happens all the time. The switch to lightdm has been stable, though, in another laptop with the same versions as above. So, this is not impossible.

Questions

• Which settings make it so that lightdm/Unity does not proceed beyond the greeter?
• What should I look into and edit?

Similar cases

• Ubuntu 18.04 lightdm does not work properly: gdm pops again after the screen goes into locked mode.

Research

• Looking into /var/log/lightdm/lightdm.log the part where things go gdm is

  [+143.56s] DEBUG: Greeter closed communication channel

  [+143.56s] DEBUG: Session pid=1246: Exited with return value 0 ...

  [+143.56s] DEBUG: Seat seat0: Greeter stopped, running session ...

  [+143.56s] DEBUG: Session pid=1531: Running command /usr/sbin/lightdm-session env GNOME_SHELL_SESSION_MODE=ubuntu gnome-session --session=ubuntu ...

  [+143.56s] DEBUG: Session pid=1531: Logging to .xsession-error

• the line in boldface is contained in /usr/share/xsessions/ubuntu.desktop

  [Desktop Entry]

  Name=Ubuntu

  Comment=This session logs you into Ubuntu

  Exec=env GNOME_SHELL_SESSION_MODE=ubuntu gnome-session --session=ubuntu

  TryExec=gnome-shell

  Type=Application

  DesktopNames=ubuntu:GNOME

  X-Ubuntu-Gettext-Domain=gnome-session-3.0

Using Ubuntu 16.04, deja-dup 34.2, and duplicity 0.7.06.

To start a weekly backup I connected the usual external drive with the usual computer, gave the password and received the halt message:

The existing backup is of a computer named correctname, but the current computer name is correctname.lan. If this is unexpected, you should back up to a different location.

Cancel/Continue

This is unexpected. Not because I plugged the external drive into the wrong computer, but because the name of the computer is indeed correct. Checked with uname -n.

I am not quite inclined to back up in a different location, because I would not want to have this issue happen again with other names/computers. Fresh backups are time-consuming, among other considerations.

The questions arise: how comes that deja-dup/duplicity is making up this name? how to fix this?

Added. By clicking on Continue you do perform an incremental backup. Still the halt message is an unnecessary stop in the procedure to avoid if possible.

1 Starting points and intents

I wish to have the Windows/MSDOS fonts installed in my Ubuntu 18.04, installing the package ttf-mscorefonts-installer (3.6ubuntu2).

2 Ubuntu 18.04

2.1 The Synaptic way

After launching the installation of ttf-mscorefonts-installer, I am asked for acceptance of the EULA license

The following message makes me think that all is fine:

Instead, shortly afterwards, I am presented with this one which betrays my expectations:

If I click on 'Run this action now' there is another windows asking for authentication for /usr/lib/update-notifier/package-data-downloader. If I proceed, the screen is greyed out but the application does not seem to proceed within any reasonable time. I can cancel the operation, but I am unable to fix the problem

2.2 The terminal-based way

2.2.1 sudo apt install ttf-mscorefonts-installer

Repeating the same from the terminal, the interesting part of the input is

/usr/lib/msttcorefonts/update-ms-fonts: 87: /usr/lib/msttcorefonts/update-ms-fonts: cabextract: not found

repeated 11 times, and

mv: cannot stat 'andalemo.ttf': No such file or directory

repeated 30 times for the fonts andalemo, ariblk, arial, arialbd, arialbi, ariali, comic, comicbd, cour, courbd, couri, courbi, georgia, georgiab, georgiai, georgiaz, impact, times, timesbd, timesbi, timesi, trebuc, trebucbd, trebucit, trebucbi, verdana, verdanab, verdanai, verdanaz, webdings.

The concluding remark is

One or more fonts could not be extracted. The fonts are NOT installed. Please run 'dpkg-reconfigure ttf-mscorefonts-installer' to perform the installation again

2.2.1 dpkg-reconfigure ttf-mscorefonts-installer

Launching the above command, first I get this window about font files location:

I leave the line blank since the fonts have presumably not been downloaded (not extracted, not found).

Second, this other window on mirror repository

where I again leave the line blank trusting that the system is informed about where to fetch the material.

The command is completed with exit status 0, but there is no messaging about what happened.

2.2.3 sudo apt install ttf-mscorefonts-installer take 2

Reading package lists... Done Building dependency tree
Reading state information... Done ttf-mscorefonts-installer is already the newest version (3.6ubuntu2). 0 to upgrade, 0 to newly install, 0 to remove and 0 not to upgrade.

At this point I don't have a benchmark to know whether the installation has been complete and successful. In /usr/share/fonts/ I cannot find the ttf files mentioned above in 2.2.1: find -name verdana.ttf produces no match in both /usr/share/fonts and $HOME

Questions

• How can I be sure that the installation has been terminated successfully?

• Where are the fonts supposed to be fetched from (remotely) and saved into (locally)?

Knowing this is important for triaging other issues down the line; more later in a different post.

Tip

I have seen in another computer with Ubuntu 14.04 and ttf-mscorefonts-installer 3.4+nmu1ubuntu1 that the fonts in points have actually been stored in /usr/share/fonts/truetype/msttcorefonts/. So I imagine that now there must be something wrong with sourcing and saving these files.

I am using Ubuntu 16.04 and listening to music using Banshee. At a certain point the graphics interface greys out and the application is unresponsive.

If I launch ps, the process list is monopolized by an application 'Main Thread' which is taking in excess of 100% of the CPU time and has been running for 20+ minutes as of now.

If I launch pstree the relevant excerpt is

    |-lightdm-+-Xorg-+-{InputThread}
    |         |      `-{disk_cache:0}
    |         |-lightdm-+-upstart-+-at-spi-bus-laun-+-dbus-daemon
    |         |         |         |                 |-{dconf worker}
    |         |         |         |                 |-{gdbus}
    |         |         |         |                 `-{gmain}
    |         |         |         |-at-spi2-registr-+-{gdbus}
    |         |         |         |                 `-{gmain}
    |         |         |         |-bamfdaemon-+-{dconf worker}
    |         |         |         |            |-{gdbus}
    |         |         |         |            `-{gmain}
    |         |         |         |-bash---Main Thread-+-{Finalizer}
    |         |         |         |                    |-7*[{Main Thread}]
    |         |         |         |                    |-{Timer-Scheduler}
    |         |         |         |                    |-{aqueue:src}
    |         |         |         |                    |-{audiosinkqueue:}
    |         |         |         |                    |-{dconf worker}
    |         |         |         |                    |-{flacparse8:sink}
    |         |         |         |                    |-{gdbus}
    |         |         |         |                    |-{gmain}
    |         |         |         |                    |-{mono}
    |         |         |         |                    |-{threaded-ml}
    |         |         |         |                    |-{typefind:sink}
    |         |         |         |                    `-{vis-queue:src}    

Waht does this application 'Main Thread' do? And is there a way to avoid pre-emptively this way of things going stuck?

The hardware in point is an oldish desktop with a msdos partition table and booting in BIOS mode. The objective is to change the operating system arrangement from

• dual boot Ubuntu 14 + Windows

to

• Ubuntu 16.04 only

by using a bootable USB pen drive with a Live 16.04 (gpt partition table) and by redrawing the partition table with the install option "Something else".

Early in the installation procedure I receive the following message

The machine's firmware has started the installer in UEFI mode but it looks like there may be operating systems already installed using "BIOS compatibility mode".

If you continue to install Debian [sic] in UEFI mode, it might be difficult to reboot the machine into any BIOS-mode operating system later.

If you wish to install in UEFI mode and don't care about keeping the ability to boot one of the existing operating systems, you have the option to force that here.

If you wish to keep the option to boot an existing operating system, you should choose NOT to force UEFI installation here

Go back/ Continue in UEFI mode

I am slightly confused by the diagnose above because the machine has no efi directory inside /sys/firmware to start with, so the first diagnose should not apply. Also, any operating system should be in BIOS mode in this machine.

Not knowing the inner workings of this request, I would go for 'go back', which in passing means 'go forward along the other way'. However, I am unsure about what really applies to my case, and better safe than sorry.

Would an expert/experienced fellow explain which option is safe?

I would like to have the current system (16.04) manage the upgrade (to 18.04.2) as suggested by the Software Updater. However, I would also like to preserve the content of /home, /opt and /usr/local that, in the current filesystem, are mounted to dedicated partitions.

Keeping /home untouched sounds like a trivial requirement. On the other hand, the other two directories also contain custom installations and links that a user would not want to delete, but I am not so sure whether the Software Updater will format them.

The question here is: will the automatic upgrade procedure only overwrite those filesystem directories strictly needed by the system (/, /boot, /var, /tmp, /etc, etc), and leave the others untouched?

The answer will help me realize the amount of work needed down the line if I take this avenue. I am aware that, alternatively, I could steer the installation manually and preserve those directories.

I have installed Ubuntu 16.04 Desktop on a laptop. I wanted to create separate partitions for each main directory in the default filesystem of Ubuntu. For these I have my ballbark estimate of how much space they need, but I am new to the `/srv' directory.

Which purposes does it serve in the first place? The description in man hier (thanks to https://askubuntu.com/a/944/446253)

/srv   This directory contains site-specific data that is served by this system.

is pretty generic. Any examples? What are the typical volumes of the data stored there? How quickly should I expect that it fills up?

Any tip that helps guide intuition much appreciated.

Working with Ubuntu 14.04 LTS.

I have a tracklist composed with Banshee 2.6.2 that consists of 20 tracks for a total of 485 MB. A CD with a capacity of 700 MB is in the tray. I then send this list to Brasero for burning (via the write CD command in Banshee). Brasero 3.10.0 then tells me that the blank CD-R has not enough free space.

I have checked that the file sizes reported by Banshee correspond to those shown by a vanilla ll -h command. I have also tried with different CD's. Also, this is not a matter of MiB and MB, since there should be enough space anyway.

Any clues on how to have Brasero use the full capacity of the CD? Or, if Brasero is not reliable in that respect, on how to export a playlist from Banshee to a functioning CD-burner?

Situation

I am working with Ubuntu 14.04 LTS. My computer was (and is) running a several-day long task launched from a terminal in the desktop environment. I then sent it to the background and disowned from the host shell.

In this way I could log out of the desktop environment and have it going all the same. Prospected advantage: the memory usage will be kept at a minimum, for the applications in the graphical environment will not possibly leak memory and cause the system to freeze in the long run.

By using the tty consoles I could check the state of the run, and possibly log in back into the desktop environment if so wished. I tested this way of operating, and it went very well.

There was plenty of memory available for all processes running.

Issue

When I was logged out of any console, I played with the key combinations Alt + Ctrl + F1 through F7, just to see once again how the system switched between the login prompts/greeter for each tty.

I then went a few keystrokes too far, Alt + Ctrl + F8 and onwards, and landed on a blank page with a blinking cursor; this is expected. The unexpected is that I cannot go move away from this blank page any longer. I have checked on another machine that such a move back is possible.

The computer seems to keep on working as it should. The keyboard is properly wired onto the machine (the screen's sleep responds to it).

Questions

What could be a reason for this behaviour?

Is there a way to get the computer responding to the console selection other than kill all processes and starting from scratch?

Note

No avail from the following suggested answers:

ctrl+alt+f* is not working

How do I get Ctrl+Alt+F* virtual terminals to work in Ubuntu 12.04?

Ctrl+Alt+F[1-12] don't switch to TTY

There are several threads concerning deactivating encryption by ecryptfs on a home directory. A selection is listed below. All (seem to) suggest that the current home directory will be wiped out by the process of removing encryption.

The output of ecryptfs-setup-private --undo says

In the event that you want to remove your eCryptfs Private Directory setup, you will need to very carefully perform the following actions manually:

1. Obtain your Private directory mountpoint $ PRIVATE=cat ~/.ecryptfs/Private.mnt 2>/dev/null || echo $HOME/Private
2. Ensure that you have moved all relevant data out of your $PRIVATE directory
3. Unmount your encrypted private directory $ ecryptfs-umount-private
4. Make your Private directory writable again $ chmod 700 $PRIVATE
5. Remove $PRIVATE, ~/.Private, ~/.ecryptfs Note: THIS IS VERY PERMANENT, BE VERY CAREFUL $ rm -rf $PRIVATE ~/.Private ~/.ecryptfs
6. Uninstall the utilities (this is specific to your Linux distribution) $ sudo apt-get remove ecryptfs-utils libecryptfs0

Since $PRIVATE is actually /home/user (the directory protected by encryption, please check), the current home directory gets emptied and deleted because of one's own doing through rm -rf $PRIVATE

I noticed one fellow's interesting remark on line 5. Reportedly https://askubuntu.com/a/10049/446253 did not find deleting home necessary, which is a very interesting hint. Making back-ups of data is good in its own right, but it's no immaterial thing if you could remove encryption without manipulating your user home content!

I am not certainly asking to put your own data at risk in place of mine, but could more experienced people judge whether a step 5 like rm -rf ~/.Private ~/.ecryptfs would be sufficient instead?

This could help out people who struggle with full disks to be emptied and cannot resort to network file systems because of encryption, for example.

Sources

How to stop using built-in home directory encryption?

Can you unencrypt / remove encryption from a user home folder?

How to disable encrypted home directory?

Can I disable full-disk encryption?

Removing Encryption on Ubuntu 14.04

Antecedents

I am working with both Ubuntu 14.04 and 16.04 in separate desktops. All disks are encrypted. I have changed the root and user passwords in several computers with sudo passwd and sudo passwd <username>.

Issues

When I log in, the new password is accepted by the greeter but the systems loops and I get the greeter again and again.

This has occurred to me

• in one of two desktops running on Ubuntu 14.04
• in one of two desktops running on Ubuntu 16.04

where I changed the password in the same way. No clue of what causes this difference between either computer that runs with the same version.

As I go into the console (CTRL+ALT+F1), I successfully log in as well. The console prompts into the appropriate /home/user.

However, in 14.04, the home directory contains only two files: Access-Your-Private-Data.desktop and README.txt instead of the full directory tree. On suggestion of the latter I do the following

1. run ecrypts-mount-private,
2. input my old password
3. launch cd again,

and the whole directory content show up.

Since in 16.04 the booter explicitly asks for the ecryptfs password prior to moving on to the greeter, the user can actually type in the old password in due course and avoid falling in the loop. This is less inconvenient than in 14.04 but still an undesired behaviour.

Therefore, the encryption password is lagging behind the root and user passwords when these are updated.

This topic is similar to the questions

• Ubuntu 14.10 does not log in despite correct password
• https://askubuntu.com/questions/683677/
• https://askubuntu.com/questions/733086/

The answers given there, if any at all, are inconclusive for me. I have gathered the following information

• x-session error
• tail -50 /var/log/Xorg.0.log
• ls -l ~/.Xauthority gives -rw------- 1 user user 227 Sep 2 16:04 .Xauthority hence no permission issue, I believe
• ls -l / | grep tmp gives drwxrwxrwt 4 root root 28672 Sep 2 19:13 tmp hence no sticky bit issue, I believe

Question

• How do I align the ecryptfs password with the user password, so that they are the same at all times?

This is my hardware/software configuration:

• HP Color LaserJet MFP M277dw
• HPLIP 3.15.11
  • Note of 6 Oct 2016: the problem persists with HPLIP 3.16.2
• HPP0053.PPD version 19.12, freshly updated last week.
  • Note of 6 Oct 2016: the driver is still the same

I want to print some pages from a pdf file with images in colour. I visualize this file with evince and want to print it in greyscale rather than in colour. Therefore, I activate the instruction to "print colour as gray" in the printer dialogue window.

In disregard of my settings, the preview of evince itself is in colour and so is the print. (Moreover, this occurs regardless of whether I set gray printing as default in the printer settings upstream of evince.)

This issue is recent is my computer. Previously there was a perfect correspondence between evince, its printer dialogue settings and the printer response. Beside the regular Ubuntu updates, the only change that I can obviously connect this new situation to is the driver upgrade from HPP00036.PPD v 19.10 to HPP0053.PPD v 19.12.

This issue is not new to HP printers, though. This post suggests to add the same printer once again, but this does not work for me. I get a responsive colour/greyscale toggle only if I install my printer as a M577 model instead of M277 (thereby giving in the duplex printing capabilities). This indicates that the glitch has to do with the driver devised for this specific model.

In the internet it's possible to find similar complaints on HP printers in Windows environments as well.

I would like to restore a normal state of affairs.

• Has anyone had a similar experience?
• Is the HP driver the actual culprit?
• Is there a proven solution strategy to fix this?

Addendum 11/12/15

The screenshot below shows the job attributes of one print that I intended to print in grayscale and was printed in colour instead. The line `HPPJLColorAsGray yes' is apparent. Even worse, the document is printed in high resolution, ignoring the EconoMode setting too.

• Does anyone spot a fault that could invalidate these settings?

I run my computer on Ubuntu 14.05 LTS and have recently connected a new HP Color LaserJet MFP-M277dw. The connection is local via USB.

At the point of first usage, the system did recognize the device somehow so that I could print, but I cannot scan documents yet. Full scanning functionalities are the aim.

From https://help.ubuntu.com/community/HpAllInOne I learn that

You have selected Ubuntu 14.04 using the HP Color LaserJet MFP m277dw. Ubuntu 14.04 supplies HPLIP 3.11.5 by default, which does not support your printer. You must ensure latest HPLIP version (recommemded), or at least HPLIP 3.15.4 in order to use your printer with Ubuntu 14.04.

Checking on http://packages.ubuntu.com/search?keywords=HPLIP&searchon=names&suite=trusty&section=all shows that, to date, Ubuntu 14.04 covers up to HPLIP 3.14.3 only.

The question is: what is the fail-safe strategy (i.e. the sequence of commands) that I should follow to get this all-in-one printer to scan?

I have run hp-check -r whose output is here.

Thanks for dealing with this.

I was thinking to have solved one problem with Paraview but have come across a bigger one with Brasero.

1. A few days ago I wanted to install Paraview (amd64) in my computer (x68_64, Ubuntu 12.04 LTS) where, among others, Google-Earth and Skype used i386 libraries. This generational conflict caused apt-get to ask the removal of those older packages and attending libraries. (Kudos to Sneetsher for pointing me to this.) In the end, I managed to get Paraview, Skype and Google Earth properly installed and functioning after having run apt-get -f install, that is a general overhaul of broken dependencies. I did so on suggestion of apt-get itself.
2. However, while apt-get -f install was tending to its task, I noticed that some applications were disappearing from the desktop menu. I considered this natural, hence I moved on to reinstalling those. The installation of Banshee as first go went well (I can't recall whether using apt-get or dpkg). With Brasero came the fun.
3. After commanding apt-get install brasero, apt-get asks me to remove another large number of libraries, in the order of hundreds. Mistakenly I consider those applications to add and give my go ahead. At the end of this spree of autoimmune exuberance, the first evidence is that the browser cannot connect to URLs any longer -- thus the internet connectivity is knocked off. I then give the reboot a chance. The system goes through the POST seamlessly, then it hangs on the Ubuntu splash screen.
4. Cutting the long story short, I managed to reboot in recovery mode and get to a responsive command line terminal. As the POST rolls through its checks, I can see that it fails on starting LightDM Display Manager (but maybe somewhere else that I miss).

Now the state of the play from the recovery-mode command line:

• ping yields a 'user unknown'
• xterm yields a 'display is not set'
• both lightdm and gdm yield a 'currently not installed'
• dpkg-reconfigure lightdm yields a 'broken or not fully installed'
• apt-get install lightdm gives a streak of 'failed to fetch' arguably because the internet has been cut out (checked out that the cable is in place)
• I can see the external drives though!

Hence I was thinking to circumvent the internet cut-out by saving a bridgehead of deb files on one of those drives and then running dpkg or apt-get on them from the command line. Clearly, I am not so much in favour of reinstalling the whole OS.

Does it make sense to you? If it does, the questions are

1. Which packages and dependencies would be able to restore the internet connection?
2. Since the dependencies could easily be in the order of tens if not hundreds, is there a place where these are conveniently zipped/tarred in a single bundle? Of course I can't rely on remote fetching.
3. Say I manage to restore the internet connectivity, then I will be able to fetch the remote repositories. Which commands should I run to make sure that all 'broken pieces' are rescued and restored?
4. Why on Earth did Brasero needed to remove 100+ library packages of essential nature? This seems truly excessive to me!

Thanks for having patiently read this long post. I look forward to your answers and remarks.

Update: requested outputs

• /var/lib/apt/extended_states
• /var/log/dpkg.log