Stephen Ostermiller's questions

I just upgraded to Ubuntu 18.10 from Ubuntu 17.10.

I have my box configured to use bash as the default shell rather than the default dash. I've made this change with sudo update-alternatives --config sh which reports:

  Selection    Path            Priority   Status
------------------------------------------------------------
  0            /bin/dash        200       auto mode
* 1            /bin/bash        100       manual mode
  2            /bin/dash        200       manual mode

However, /bin/sh still points to dash rather than pointing to bash.

Actually, on previous versions of Ubuntu, I believe that used to point /bin/sh -> /etc/alternatives/sh.

It looks like the latest Ubuntu no longer allows the alternatives mechanism for sh or at least broke the alternatives during the upgrade. I see from How Can I Make /bin/sh point to /bin/bash? that it is now recommended to reconfigure dash instead.

Why doesn't alternatives work anymore? Where is this change documented? I've been using alternatives with sh for at least three years in Ubuntu and I've never had it broken during an upgrade before. There are many guides that suggest that such as: https://jwaghetti.blogspot.com/2015/09/changing-dash-to-bash.html

To get a wildcard SSL certificate from LetsEncrypt you need to run certbot version 0.22 or later according to https://community.letsencrypt.org/t/acme-v2-production-environment-wildcards/55578

I've installed certbot from apt using

sudo apt-get install certbot

However, that reports:

certbot is already the newest version (0.17.0-2).

How can I upgrade certbot to a version that supports wildcard certificates?

I'm used to installing ubuntu core updates from apt, and then rebooting to get to the latest kernel. However, this process doesn't seem to be working right on one of my servers. Kernel 4.10.0-21 is installed but the machine only seems to use 4.10.0-20 when it reboots.

During the apt update process to grub2, a dialog asked me if I want to chain my legacy grub configuration. I chose "no". I think that may be part of the problem.

Updating grub seems to do the right thing, it puts the newest kernel first:

$ sudo update-grub
Generating grub configuration file ...
Found linux image: /boot/vmlinuz-4.10.0-21-generic
Found initrd image: /boot/initrd.img-4.10.0-21-generic
Found linux image: /boot/vmlinuz-4.10.0-20-generic
Found initrd image: /boot/initrd.img-4.10.0-20-generic
Found memtest86+ image: /boot/memtest86+.elf
Found memtest86+ image: /boot/memtest86+.bin
done

My /etc/grub/default file says that kernel 0 should be the default:

GRUB_DEFAULT=0
GRUB_HIDDEN_TIMEOUT_QUIET=true
GRUB_TIMEOUT=3
GRUB_DISTRIBUTOR=`lsb_release -i -s 2> /dev/null || echo Debian`
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"
GRUB_CMDLINE_LINUX=""

I can verify that grub.cfg gets the kernels added:

$ grep menuentry /boot/grub/grub.cfg
...
    menuentry 'Ubuntu, with Linux 4.10.0-21-generic' --class ubuntu --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-4.10.0-21-generic-advanced-6af3f559-58c5-4c56-b1c8-ffe02c700c97' {
    menuentry 'Ubuntu, with Linux 4.10.0-21-generic (recovery mode)' --class ubuntu --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-4.10.0-21-generic-recovery-6af3f559-58c5-4c56-b1c8-ffe02c700c97' {
    menuentry 'Ubuntu, with Linux 4.10.0-20-generic' --class ubuntu --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-4.10.0-20-generic-advanced-6af3f559-58c5-4c56-b1c8-ffe02c700c97' {
    menuentry 'Ubuntu, with Linux 4.10.0-20-generic (recovery mode)' --class ubuntu --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-4.10.0-20-generic-recovery-6af3f559-58c5-4c56-b1c8-ffe02c700c97' {
...

$ grep -Poz "menuentry 'Ubuntu' (.|\n)*?\}" /boot/grub/grub.cfg
menuentry 'Ubuntu' --class ubuntu --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-simple-6af3f559-58c5-4c56-b1c8-ffe02c700c97' {
    recordfail
    load_video
    gfxmode $linux_gfx_mode
    insmod gzio
    if [ x$grub_platform = xxen ]; then insmod xzio; insmod lzopio; fi
    insmod part_msdos
    insmod ext2
    set root='hd0,msdos1'
    if [ x$feature_platform_search_hint = xy ]; then
      search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos1 --hint-efi=hd0,msdos1 --hint-baremetal=ahci0,msdos1 --hint='hd0,msdos1'  6af3f559-58c5-4c56-b1c8-ffe02c700c97
    else
      search --no-floppy --fs-uuid --set=root 6af3f559-58c5-4c56-b1c8-ffe02c700c97
    fi
        linux   /boot/vmlinuz-4.10.0-21-generic root=UUID=6af3f559-58c5-4c56-b1c8-ffe02c700c97 ro  quiet splash $vt_handoff
    initrd  /boot/initrd.img-4.10.0-21-generic
}

The kernel file actually exists on disk:

$ ls -l /boot/vmlinuz-4.10.0-21-generic
-rw------- 1 root root 7575312 Apr 28 06:41 /boot/vmlinuz-4.10.0-21-generic

On the chance that it was corrupted somehow, I tried re-installing it:

$ sudo apt-get install --reinstall linux-image-4.10.0-21-generic

dpkg shows that both kernels are installed:

$ dpkg -l | grep ii | grep linux-image-[0-9].*-generic
ii  linux-image-4.10.0-20-generic               4.10.0-20.22                              amd64        Linux kernel image for version 4.10.0 on 64 bit x86 SMP
ii  linux-image-4.10.0-21-generic               4.10.0-21.23                              amd64        Linux kernel image for version 4.10.0 on 64 bit x86 SMP

However after rebooting:

$ uname -r
4.10.0-20-generic

$ cat /proc/version 
Linux version 4.10.0-20-generic (buildd@lcy01-05) (gcc version 6.3.0 20170406 (Ubuntu 6.3.0-12ubuntu2) ) #22-Ubuntu SMP Thu Apr 20 09:22:42 UTC 2017

$ cat /proc/cmdline
root=UUID=6af3f559-58c5-4c56-b1c8-ffe02c700c97 ro quiet splash 

I really think that grub legacy is somehow still being used on this machine. /boot/grub/menu.1st has references to 20 but not to 21:

title           Ubuntu 17.04, kernel 4.10.0-20-generic
root            (hd0,0)
kernel          /boot/vmlinuz-4.10.0-20-generic root=UUID=6af3f559-58c5-4c56-b1c8-ffe02c700c97 ro quiet splash
initrd          /boot/initrd.img-4.10.0-20-generic
quiet

title           Ubuntu 17.04, kernel 4.10.0-20-generic (recovery mode)
root            (hd0,0)
kernel          /boot/vmlinuz-4.10.0-20-generic root=UUID=6af3f559-58c5-4c56-b1c8-ffe02c700c97 ro  single
initrd          /boot/initrd.img-4.10.0-20-generic

title           Ubuntu 17.04, kernel 4.8.0-49-generic
root            (hd0,0)
kernel          /boot/vmlinuz-4.8.0-49-generic root=UUID=6af3f559-58c5-4c56-b1c8-ffe02c700c97 ro quiet splash
initrd          /boot/initrd.img-4.8.0-49-generic
quiet

title           Ubuntu 17.04, kernel 4.8.0-49-generic (recovery mode)
root            (hd0,0)
kernel          /boot/vmlinuz-4.8.0-49-generic root=UUID=6af3f559-58c5-4c56-b1c8-ffe02c700c97 ro  single
initrd          /boot/initrd.img-4.8.0-49-generic

title           Ubuntu 17.04, memtest86+
root            (hd0,0)
kernel          /boot/memtest86+.bin
quiet

This machine is in a data center, and I don't have console access during the reboot process. I'm unable to see what grub says during the reboot, I can only ssh in after it is up and running.

What could be causing the newest kernel not to be used and how can I fix it?

I have my desktop computer configured with a static IP address:

However, /etc/resolve.conf keeps getting overwritten with:

nameserver 127.0.0.1
search localdomain example.com

This seems to happen periodically or when I run sudo resolvconf -u.

I would really like to get example.com removed from the dns search domain list. I used to have it in there but its causing problems now. I just can't seem to get it out. I don't know where it is coming from or why it keeps getting put back. As far as I can grep, there is nothing referring to example.com anywhere in my /etc directory anymore.

What is putting this search domain into my resolv.conf file and how can I prevent it from doing so?

When I try to print from a Gnome application, the print dialog has non-functional printer listed. It is called "duerqxesz5090 -- Basement Utility Room -- Rejecting Jobs". Here is a screenshot of the Gnome print dialog box:

• If I try to select that printer, the application will freeze for minutes or even crash.
• I only get this with Gnome programs (such as gedit or glabels), but not with firefox or libreoffice
• It is possible that it is the remnants of an old printer I did have installed at some point. I did have a printer installed in the basement utility room.
• I'm not using Gnome (I'm using xfce), but I have used Gnome in the past.
• My Ubuntu version is Ubuntu 15.04 (Vivid)
• CUPS doesn't know anything about this ghost printer:
• system-config-printer doesn't know anything about this printer:

How can I find out where this non-existent printer is listed on my system and remove it?

I've tried searching for it in various places, but none of these return any results:

• dconf dump / | grep duerqx
• sudo grep -r duerqx /etc/
• gsettings list-recursively | grep duerqx

Vivid switched from upstart to systemd. I used to have xfvb configured to run as a service under updstart. That has now broken.

Under upstart, I had /etc/init/xvfb.conf:

description     "Xvfb X Server"
start on (net-device-up
    and local-filesystems
    and runlevel [2345])
stop on runlevel [016]
exec /usr/bin/Xvfb :99 -screen 0 1024x768x24

and I could start it with sudo service xvfb start

Under 15.04, I now get this message when I try to use service to start xvfb:

Failed to start xvfb.service: Unit xvfb.service failed to load: No such file or directory.

What is the new way to configure xvfb to run as a service with systemd?

I have a cron job that checks for new releases. I want to make sure it emails me when a new release is out of beta and ready to install. I want to upgrade to both long term support (LTS) releases and non-LTS releases. Is there a command that will tell me that such a release is available?

I've tried using do-release-upgrade -c -q and it tells me about the new release unless I'm currently on LTS and the next release is not LTS, (as documented here).

I've tried using do-release-upgrade -d -c -q. It tells me about non-LTS releases, but it also tells me about beta releases. For example, if I use it right now it tells me that 15.04 is available, but it hasn't actually been fully released yet, (as documented here).

What is the command to tell me about all releases that have been fully released, whether or not they are LTS?

I'm trying to install the Fuze Meeting client which I'm planning to use for Webinars at work. They have a Linux client and make it available through an apt Repository.

Their installation instructions are:

sudo bash -c 'echo "deb http://apt.fuzebox.com/apt $(lsb_release -sc) main" > /etc/apt/sources.list.d/fuzebox.list'
sudo wget -O /etc/apt/trusted.gpg.d/fuzebox.gpg http://apt.fuzebox.com/apt/keyring.gpg
sudo apt-get update
sudo apt-get install fuzelinuxclient

However I'm on Utopic 14.10 and their repository is only current through Trusty. Is there a way to install this software on Utopic even though it is only officially available for older versions of Ubuntu?

I tried replacing $(lsb_release -sc) with trusty in the above steps but I get the following warning when I try to apt-get update.

W: Failed to fetch http://apt.fuzebox.com/apt/dists/trusty/Release Unable to find expected entry 'main/binary-amd64/Packages' in Release file (Wrong sources.list entry or malformed file)

And then this error when I try to install:

E: Unable to locate package fuzelinuxclient

I'm using XFCE in Ubuntu 14.04. When I open gEdit, the tabs at the top are unreadable:

Having done some research, it appears that this is because gEdit is a GTK 2.0 application and this problem manifests itself if the theme you are using does not have specific workarounds in it for GTK 2.0 applications.

Here is what I have tried so far:

• This forum thread seems to indicate that using the Greybird theme would solve the problem. The above screenshot was taken with Greybird, and it did not solve the problem.
• This appears to be the same problem asked on StackOverflow. None of the answers are very helpful.

• This is a similar question asked here but is about tabs that are very dark but not black:

• This forum thread suggest creating a configuration file for a fallback theme for GTK 2.0 applications. Creating this configuration file did not solve my problem.

I'd like my cronjobs to be able to send mail from my home computer. My servers configure exim4 like this:

dc_eximconfig_configtype='internet'
dc_other_hostnames=''
dc_local_interfaces='127.0.0.1'
dc_readhost=''
dc_relay_domains=''
dc_minimaldns='false'
dc_relay_nets=''
dc_smarthost=''
CFILEMODE='644'
dc_use_split_config='false'
dc_hide_mailname=''
dc_mailname_in_oh='true'
dc_localdelivery='maildir_home'

However, at home, my ISP blocks mail sent directly via a firewall. I have a password proteced ssl smtp server that I could route the mail through, but I have been unsuccessful in figuring out how to configure exim4 to make this work. I'm hoping somebody can give me a configuration example where I can plug in the host name, user name, password, and "use ssl" to get it to send mail through a single smtp server.