Panic Averted (for now)

-=> Dennis Lee Bieber wrote to All Now the question: What have I been doing in my backups since early 2018 >that finally came back and bit me in the butt? I'd like to do an in situ >upgrade to Buster pretty soon, but I won't proceed with that until I >learn what I've been doing wrong.

DLB> Haven't I read recommendations to do a clean install of DLB> Buster rather than trying to upgrade by changing the repository DLB> list...

DLB> It's just the cost of a SD card and a few hours (well, DLB> it is if installing Raspbian via NOOBS , which I did a few DLB> days ago onto two SD cards) -- and leaves you with your current DLB> SD card for use if something doesn't upgrade properly.

If you're running an "important" function with a RPi, it's only smart to set it up to *boot* from the SD but actually *RUN* from a connected USB/SSD hard drive. The SD becomes read-only and is not used other than to read from to boot. All writing is done to the hard drive.

Doing anything else is just...................... silly.

... Backup not found: (A)bort (R)etry (P)anic === MultiMail/Linux v0.52

Reply to
Dan Clough
Loading thread data ...

I have been running WeeWX on a Pi 3B with Raspbian Stretch since early

2018. I do an automated Duplicity backup once-a-week, and a separate backup of my WeeWX files just before I update my software and reboot, which I also perform weekly. My software update script contains these lines:

sudo /etc/init.d/weewx stop sudo apt-get update sudo apt-get upgrade -y sudo apt-get dist-upgrade -y sudo apt-get install raspberrypi-ui-mods sudo apt-get autoremove sudo apt-get autoclean sudo apt-get clean

Like I said, I've been doing this weekly since early 2018 with no headaches. This week however, I lost my entire visual interface: no command line, no GUI, as if my HDMI port was completely dead. I could still ssh in and verify that the Pi was running more or less properly.

A few web searches suggested that HDMI problems could be alleviated by editing /boot/config.txt, which currently looks like this:

disable_overscan=1 hdmi_force_hotplug=1 hdmi_group=2 hdmi_mode=47 dtparam=i2c_arm=on dtparam=audio=on program_usb_boot_mode=1 dtoverlay=i2c-rtc,ds3231

I made a few tweaks suggested by a few sites. Nothing worked. Still no HDMI.

One site suggested updated the Pi's firmware, so I did a sudo rpi-update.

BTW, verified on another Pi that my monitor was still working.

I was getting nowhere, so I finally threw in the towel and replaced the contents of /boot with the latest backup of that directory. Rebooted. HDMI came back! Joy!

Now the question: What have I been doing in my backups since early 2018 that finally came back and bit me in the butt? I'd like to do an in situ upgrade to Buster pretty soon, but I won't proceed with that until I learn what I've been doing wrong.

[For those who wonder, WeeWX interfaces with a Davis Vantage Vue weather station, and aside from some initial setup issues, thanks to the WeeWX community, has given me no trouble.]
--
chrispam1@me.com is a filtered spam magnet. Email replies may be lost. 
You're better off replying to this newsgroup.
Reply to
Chris Schram

On Sun, 14 Jul 2019 23:27:45 +0000 (UTC), Chris Schram declaimed the following:

Haven't I read recommendations to do a clean install of Buster rather than trying to upgrade by changing the repository list...

It's just the cost of a SD card and a few hours (well, it is if installing Raspbian via NOOBS , which I did a few days ago onto two SD cards) -- and leaves you with your current SD card for use if something doesn't upgrade properly.

--
	Wulfraed                 Dennis Lee Bieber         AF6VN 
	wlfraed@ix.netcom.com    http://wlfraed.microdiversity.freeddns.org/
Reply to
Dennis Lee Bieber

My main justification for wanting to attempt a "dirty" upgrade is that over time I have made many poorly documented changes. Yeah, my bad, but I have not heard of any upgrades of this type that have gone wrong.

BTW, SD card is not in my current vocabulary. My current system started out when I used Etcher to copy the image directly to a WD Pi Drive.

--
chrispam1@me.com is a filtered spam magnet. Email replies may be lost. 
You're better off replying to this newsgroup.
Reply to
Chris Schram

I read those too, but merely did my weekly backup as usual and followed this up with an upgrade in situ by following the instructions given by the RPI foundation.

See the "Buster out of testing, update requires action" thread for my comments and niggles about the process, but the bottom line is that my old 512MB RPi B system is running Buster as I write.

Make a full backup (use dd to copy both partitions on your current SD card to a new SD card) before you do anything else.

Also note that I run my RPi headless via SSH sessions from a laptop running Fedora Linux and that I found it very useful to run the upgrade in one console window and have another open for editing non-standard RPi config files, looking at RPi man pages, etc.

I found the upgrade pretty straight forward even though apt stopped (twice IIRC) before completing the upgrade. In all cases simply rerunning the same command made the upgrade continue, so the end result was a successful upgrade from Stretch to Buster. So, some warnings and suggestions:

- you should definitely do an upgrade on Stretch and then reboot before starting the upgrade in situ.

- I've been doing this sort of Fedora in-situ upgrade for several years now and its evident that Fedora does this much better than Raspbian.

Fedora does not stop if it finds a user-modified config, but automatically applies the changes needed to bring it into line with the new Linux version while Raspbian pauses the upgrade, presents options and waits for you to deal with the issue.

The options are sensible and include diffing the file versions and letting you edit the new version. This is OK, but it does need you to keep an eye on the upgrade and take these actions as they arise.

- I found that having a second console window open was useful because I could see the diff report in the first window while fixing the config file in the second console, read man pages etc., but YMMV.

- I ran "apt autoremove" after the upgrade completed and this removed around 1000 redundant packages, saving a large amount of RAM.

- after the upgrade I replaced Exim (installed as part of Buster) with Postfix and also installed the logwatch package and both of these actions reported lists of redundant packages, causing me to run "apt autoremove" again after each of these installs.

- I would not try an upgrade in situ on an SD card thats less than 8GB - I think its quite likely to run out of space on a smaller card.

--
Martin    | martin at 
Gregorie  | gregorie dot org
Reply to
Martin Gregorie

I just did the upgrade, and all went well. I checked the old release, changed the repositories.

I just did 'sudo apt-get update' and 'sudo apt-get dist-upgrade'. The operation took nearly 3 hours on a Pi 2 and a 100 Mb/s net.

--

-TV
Reply to
Tauno Voipio

[Snip]

[Snip]

I don't think it was the backups that went wrong, but an upgrade that failed part way through. Quite often part of the upgrade is to put in redirects for the overlays in /boot, and restore the contents afterwards.

---druck

Reply to
druck

I often wonder why Debian-based distros do that while RedHat Fedora-based distros don't.

--
Martin    | martin at 
Gregorie  | gregorie dot org
Reply to
Martin Gregorie

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.