Raspios fscked again on upgrade

Feb 01, 2024 Last reply: 2 years ago 10 Replies

Bookworm. Pi 4B ..... Setting up mesa-vulkan-drivers:arm64 (23.2.1-1~bpo12+rpt3) ... Setting up mesa-vdpau-drivers:arm64 (23.2.1-1~bpo12+rpt3) ... Setting up firmware-atheros (1:20230210-5+rpt3) ... Setting up firmware-misc-nonfree (1:20230210-5+rpt3) ... Setting up linux-libc-dev:arm64 (1:6.1.73-1+rpt1) ... Setting up libkms++0:arm64 (0~git20231115~065257+9ae90ce-1) ... Setting up firmware-brcm80211 (1:20230210-5+rpt3) ... Setting up firmware-realtek (1:20230210-5+rpt3) ... Setting up raspi-firmware (1:1.20231024+ds-1+rpt5) ... Installing new version of config file /etc/default/raspi-firmware ... Installing new version of config file /etc/initramfs/post-update.d/z50-raspi-firmware ... Installing new version of config file /etc/kernel/postinst.d/z50-raspi-firmware ...


and here it sits. I cant ssh into it. It's bricked itself AGAIN



Last time I power cycled it, when this happened it wouldn't boot at all.


I've plugged an HDMI screen and key board into it and it shows the boot sequence and something about a disk drive aborted journal...detected by systemd and apt-get



Is there any hope or do I reinstall all over AGAIN.



And how to stop it happening again.


You boot from card or USB? I never faced that error. Maybe a hw problem with the card ...

It's come up with another message on the remote terminal

sync: error syncing '/boot/firmware/LICENCE.broadcom': Input/output error

The HDMI console is saying read error on swap device

read error reading directory iblock 0

On the other disk I think.

(it has two SSDs and no SD card)

Photo of HDMI console

formatting link
I will leave it on as I am scared to reboot it..last time it didn't

USB booting. Two SSDs - one to boot, one for data. Data one is never used normally - its a test setup for a new NAS and its not in any sort of use.

But errors on both drives are showing up

More errors

dpkg: error processing package raspi-firmware (--configure): installed raspi-firmware package post-installation script subprocess returned error exit status 1 Setting up libglapi-mesa:arm64 (23.2.1-1~bpo12+rpt3) ... dpkg: unrecoverable fatal error, aborting: unable to truncate for updated status of 'libglapi-mesa:arm64': Read-only file system E: Sub-process /usr/bin/dpkg returned an error code (2) W: Operation was interrupted before it could finish W: Problem unlinking the file /var/cache/apt/pkgcache.bin - pkgDPkgPM::Go (30: Read-only file system)

On BOTH USB drives? and ONLY when upgrading?

Hmm.

Well as per last time it eventually settled to bricked and totally unresponsive.

I power cycled it and now all it can say is 'Firmware not found'.

So the upgrade has once again royally f***ed it. Any ideas?

To cut a long story short, the upgrade totally wiped out the rootfs - its not even recognised as a *partition* any more - and the bootfs is looking a bit sick as well.

Thank you, bookworm.

Is it worth even continuing??

Something has scrambled the disk contents.

Or the board itself.

Upgrades involve much more CPU and IO (and therefore waste heat) than many other use cases - I could easily believe in a hardware fault that only surfaces during upgrades.

in both cases the crash happened at exactly the same place in the same upgrade

I discovered a few more points: That upgrade problem is not unique to me. In most cases people who did that upgrade and had a problem suffered a complete loss of the /boot/firmware directory.

I haven't got chapter and verse but the common factor seems to be upgrading from a previous version to the Dec 2023 version using apt-get upgrade..

I discovered why my December 2023 version didn't install, as well. It was corrupted somehow. It wouldn't mount. I re-downloaded it, re-installed it and that finally did boot, but there were strange file and read errors when I attached the *second* disk to the system. But a reboot cured *that*.

The system upgraded perfectly well once the new OS image was installed.

Perhaps the power supply is inadequate? I don't know

At the moment I have rebuilt to about the same point the development system was when it bricked

I've checked the disks with smartctl and there are no errors logged.

So the problem would seem to be with the pi itself or the OS. or some other part of the hardware.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required