Up to Main Index                           Up to Journal for January, 2020

                   JOURNAL FOR SATURDAY 18TH JANUARY, 2020
______________________________________________________________________________

SUBJECT: Everything except coding
   DATE: Sat 18 Jan 17:53:04 GMT 2020

I had such good hopes for my Christmas break with a lot of coding planned.
Instead I spent very little time actually coding. It’s now mid January and
I’ve only just found time to write the first journal entry for 2020. Not a
very auspicious start to the year :(

I now have a Raspberry Pi 4 Model B with 4Gb RAM. I have it in a Pimoroni
Rainbow Pibow Coupé 4 case with a fan SHIM and an official 3 amp USB-C power
supply. I also have the official micro HDMI to HDMI cable — after reading
about Wifi issues and different HDMI cables and adapters. Initially it will be
running headless but I plan on seeing how good it is as a desktop replacement.
A little worried about the fan SHIM as there have been issues reported with it
failing after a short while *sigh*.

While setting up the RPi4 SDCard I wanted to check some tweaks I had set up on
my RPi3 — which runs 24x7 and is regularly updated. Only it wasn’t on the
network which was strange. So I then had to mess about hooking up a monitor
and keyboard, as it runs headless, and booted it. It booted and the screen
went off. To save a little power I had ‘@reboot /usr/bin/tvservice -o’ in
root’s crontab *sigh*. So I then took the USB drive and hooked it up to my
desktop — which had no monitor attached now. So I then had a session on my
Chromebook to my desktop to sort out the drive for my RPi3. Confused? Well I
was — at one point I went to reboot the RPi3 only I used the wrong keyboard
and rebooted the desktop — damn :P

Normally for these types of recovery operations I would mount the drive and
chroot it so that /mnt is seen as the root of the file system. However the
desktop is Intel and the RPi software is for Arm. So I did what I shouldn’t
have done and directly edited /var/spool/cron/crontabs/root instead of using
crontab to edit it. Things seemed to go well, after hooking everything up
again the RPi3 booted and I could see what was going wrong. The ethernet
interface was not being brought up because udev had decided to rename eth0 to
something like enxbdeadb33f42ab. I’ve seen the problem before and absolutely
hate the new names.

I dropped a quick udev rule into /etc/udev/rules.d/70-persistent-net.rules:


  SUBSYSTEM=="net", ACTION=="add", ATTR{address}=="<mac>", NAME="eth0"


Maybe this will help someone else out. Just replace <mac> with the MAC address
for your interface — in this example I’d use “de:ad:b3:3f:42:ab”.

Once rebooted the RPi3 was back on the network. After disconnecting the
keyboard and screen from the RPi3 and hooking up my desktop again I had to
leave things there. I still haven’t powered on the RPi4 yet ;(

The story does not end there. The next day I noticed that the night jobs
hadn’t run on the desktop and that it was trying to run some jobs the RPi3
does. What!? It seems that editing /var/spool/cron/crontabs/root with Vim on
the RPi3 USB drive had updated and replaced the desktop’s copy. So I then had
to go to backups to recover the old crontab and reinstate it on the desktop.

It’s things like this that suck up my valuable coding time. Other recent
issues include Debian bug #948257 in initramfs[1] and an Exim update that now
causes it to spam the logs with “Warning: No server certificate defined; will
use a selfsigned one”. I also had NeoMutt fail on me and break in lots of
subtle ways. I use email a lot and always used Mutt. Then Debian switched Mutt
with NeoMutt behind the scenes and I was switched to NeoMutt — NeoMutt was
initially uploaded as the Mutt package replacing Mutt for all users with
NeoMutt[2]. Debian finally reinstated Mutt as Mutt and provided NeoMutt as a
separate package. By that time it was easier for me to stick with NeoMutt.
With all the recent NeoMutt breakage it was time to bite the bullet and go
back to vanilla Mutt. After the pain of sorting out my configuration file
that’s evolved over decades of use, due to NeoMutt/Mutt differences, my email
is tickety-boo again.

I still haven’t powered up the RPi4 for the first time yet :(

--
Diddymus

  [1] Debian bug #948257 - initramfs:
        https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=948257

  [2] Debian bug #870635 - mutt:
        https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=870635


  Up to Main Index                           Up to Journal for January, 2020