Problems with upgrade

23 respostas [Última entrada]
Korax
Desconectado
Joined: 02/24/2021

This is what I get when I try to run sudo apt upgrade

error: syntax error.
error: Incorrect command.
error: syntax error.
Syntax error at line 74
Syntax errors are detected in generated GRUB config file.
Ensure that there are no errors in /etc/default/grub
and /etc/grub.d/* files or please file a bug report with
/boot/grub/grub.cfg.new file attached.
run-parts: /etc/kernel/postrm.d/zz-update-grub exited with return code 1
dpkg: erreur de traitement du paquet linux-image-5.11.0-gnu (--remove) :
installed linux-image-5.11.0-gnu package post-removal script subprocess returned error exit status 1
Des erreurs ont été rencontrées pendant l'exécution :
linux-image-5.11.0-gnu
E: Sub-process /usr/bin/dpkg returned an error code (1)

amenex
Desconectado
Joined: 01/03/2015

The usual commands for routine maintenance are:
sudo apt-get update
followed by:
sudo apt-get upgrade

If you still get errors after these two commands, make sure that the
mirror settings in Software Updater are for a source in a country
that uses your preferred language. Also, be sure that the target
location for the upgrade is mounted to the file system beforehand,
and that it has the appropriate file structure. I doubt that msdos
will work very well. Ext4 is the preference of most contributors to
this forum.

Usage of apt vs. apt-get has been up to the user since time immemorial.
Just be consistent; pick one and stick with it.

If you are upgrading the operating system, such as Trisquel_8 to
Trisquel_9, searching this forum will get you a lot of opinions.
The suggestions made by Magic Banana have the greatest weight in my
opinion and experience.

George Langford

Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

Mirrors, well, mirror the main repository. They host the same packages. No language consideration applies. The filesystem(s) with the system is/are mounted when you run the system you upgrade. The installer would not let you choose an inappropriate type of filesystem, I believe. There is no problem with using apt and apt-get (and the Upgrade Manager and the Synaptic Package Manager and...).

Korax
Desconectado
Joined: 02/24/2021

I tried with apt-get, but it's exactly the same response : E: Sub-process /usr/bin/dpkg returned an error code (1)

Moreover, Trisquel crashes constantly for no reason.

When uploading code from the Arduino IDE to the Arduino, it also crashed every other time… It's extremely anoying.

I have D8 WorkStation from Vikings, running SeaBIOS. It's extremely buggy. Barely usable.

Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

I would install "memtests86+" (in Trisquel's repository) and let it test you RAM during one full night. Memtest86+ is to be launched from the menu of GRUB, the bootloader, that lists the installed operating systems right after the computer switches on. You may have to press [Esc] to have GRUB's menu appear. If memtest86+'s blue screen turns red, your RAM is defective. Instead of installing memtest86+, you can test your RAM with any live ISO that includes memtest86+.

Connochaetes

I am a member!

Desconectado
Joined: 12/13/2017

> I have D8 WorkStation from Vikings, running SeaBIOS. It's extremely buggy. Barely usable.

I'm writing this from an extremely buggy, barely usable D8 from Vikings. Let's keep in touch about this with each other (and with Vikings) while we suffer. I'm glad I'm not alone with this.

I can post more about my case later if you think it might be useful, but I haven't even finished reading this thread yet.

Avron

I am a translator!

Desconectado
Joined: 08/18/2020

I got my D8 from Vikings in November 2020, the only issue so far is that sometimes, coreboot does not complete its work and then the system does not start. In that case, I press reset and the next time it usually starts but it happened I had to press reset 2 or 3 times before successful boot.

At first, I misundersood what the reset button was so I could not boot. With a raspberry pi 2b, an USB-serial adapter and a null model cable (the Tx pin on one side of the cable is connected to the Rx pin on the other side of the cable, unlike a straight serial cable), I could see coreboot messages from the D8 and this is how I realized that what I thought was the reset button actually had no effect, then I tried pressing whatever I could and saw what was making coreboot restart, which I initally thought was a cache for an audio plug!

If you can't see coreboot messages, indeed it is a little frustrating. In that case try pressing other buttons (or even things that don't look like buttons) if the screen remains dark for more than 30s.

Besides, I have noticed that the graphic card I have is no more in the current offer, so we may have not exactly the same hardware.

In case that helps to compare with your system, I attach the ouput of 'sudo lshw' and coreboot logs with two resets until successful boot, visible with "soft_reset() called!".

AnexoTamaño
coreboot-log-with-2-resets.txt.gz 3.44 KB
hw.txt.gz 4.27 KB
Korax
Desconectado
Joined: 02/24/2021

I've already installed GRUB an have its menu when pressing ESC.

But I don't understand what to do with the iso file… where and how to install it so that it may be recognized by GRUB…

Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

Have you installed memtests86+?

Korax
Desconectado
Joined: 02/24/2021

I downloaded the iso file…

Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

Trisquel's ISO? As far as I remember, it does not include memtest86+. Just install the memtest86+ package from the system on your disk. The remark about using a live ISO was only in case you already had such an ISO with memtest86+ (what may be convenient in presence of a system that is so broken that you cannot log into it).

loldier
Desconectado
Joined: 02/17/2016

"ISO" -- that would be System Rescue CD.

https://www.system-rescue.org/

Korax
Desconectado
Joined: 02/24/2021

I downloaded memtests86+ from this link : https://www.memtest.org/#downiso
And I have a file named "mt531b.iso"
But I don't really know what to do with it

Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

As I have already written, memtests86+ is in Trisquel's repository. Just install it with the package manager.

Connochaetes

I am a member!

Desconectado
Joined: 12/13/2017

I don't know to what extent my problems resemble Korax's problems and whether I need memtest86+, but I downloaded it anyway and got the error messages below. But the installation worked, right? I'll try to let it test the RAM next time I switch on this computer.
~$ sudo apt install memtest86+
[sudo] password for usr:
Reading package lists... Done
Building dependency tree
Reading state information... Done
Suggested packages:
hwtools memtester kernel-patch-badram memtest86
The following NEW packages will be installed
memtest86+
0 to upgrade, 1 to newly install, 0 to remove and 3 not to upgrade.
Need to get 70,0 kB of archives.
After this operation, 2.548 kB of additional disk space will be used.
Get:1 https://archive.trisquel.info/trisquel etiona/main amd64 memtest86+ amd64 5.01-3ubuntu2 [70,0 kB]
Fetched 70,0 kB in 1s (102 kB/s)
Preconfiguring packages ...
Selecting previously unselected package memtest86+.
(Reading database ... 143173 files and directories currently installed.)
Preparing to unpack .../memtest86+_5.01-3ubuntu2_amd64.deb ...
Unpacking memtest86+ (5.01-3ubuntu2) ...
Setting up memtest86+ (5.01-3ubuntu2) ...
Sourcing file `/etc/default/grub'
Generating grub configuration file ...
File descriptor 3 (pipe:[52360]) leaked on vgs invocation. Parent PID 3320: /usr/sbin/grub-probe
File descriptor 3 (pipe:[52360]) leaked on vgs invocation. Parent PID 3320: /usr/sbin/grub-probe
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
File descriptor 3 (pipe:[52360]) leaked on vgs invocation. Parent PID 3350: /usr/sbin/grub-probe
File descriptor 3 (pipe:[52360]) leaked on vgs invocation. Parent PID 3350: /usr/sbin/grub-probe
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
Found linux image: /boot/vmlinuz-5.10.20-gnu
Found initrd image: /boot/initrd.img-5.10.20-gnu
Found linux image: /boot/vmlinuz-4.15.0-112-generic
Found initrd image: /boot/initrd.img-4.15.0-112-generic
File descriptor 3 (pipe:[52360]) leaked on vgs invocation. Parent PID 3715: /usr/sbin/grub-probe
File descriptor 3 (pipe:[52360]) leaked on vgs invocation. Parent PID 3715: /usr/sbin/grub-probe
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
Found memtest86+ image: /memtest86+.elf
Found memtest86+ image: /memtest86+.bin
File descriptor 3 (pipe:[52360]) leaked on lvs invocation. Parent PID 3906: /bin/sh
done

Connochaetes

I am a member!

Desconectado
Joined: 12/13/2017

Magic Banana wrote:
> Memtest86+ is to be launched from the menu of GRUB, the bootloader, that lists the installed operating systems right after the computer switches on. You may have to press [Esc] to have GRUB's menu appear.

That would be before unlocking the encrypted disk, right?
On my system, I don't notice any opportunity to do so, so I doubt that Korax will on theirs.

When I switch on my computer and a few seconds later hit ESC, I get:

SeaBIOS (version rel-1.12.0-0-ga698c89)
Press ESC for boot menu.

[The joke is, I need to press ESC even before the screen shows anything. If I wait until I see this, it's too late.]
Select boot device:
1. AHCI/3: CT2000MX500SSD1 ATA-10 Hard-Disk (1863 GiBytes)
2. Payload [nvramcui]
3. Payload [coreinfo]

And that's all.
Perhaps we do need a system rescue medium with memtest86+ on it to be able to run memtest86+ after all?

By the way, my system freezed again (including the panel clock's seconds counter, but usually excluding the cursor, which I can move fluently while unable to click or type anything) while I prepared to post this. I had to restart and rewrite my post (I had saved it as a TXT file before the freeze, but apparently it didn't survive the hard switch-off, the file was blank after rebooting) to post it here. I suspect Korax encounters similar freezes.

Connochaetes

I am a member!

Desconectado
Joined: 12/13/2017

I suspect that my problems are similar to Avron's and Korax's, so I tried some of the things mentioned in this thread and the one titled "Change boot device".

Thanks to Magic Banana's advice in the thread titled "Change boot device", I got my system to show me a GRUB screen:
GNU GRUB version 2.02
┌────────────────────────────────────────────────┐
│*Trisquel GNU/Linux                             │
│ Advanced options for Trisquel GNU/Linux        │
│ Memory test (memtest86+)                       │
│ Memory test (memtest86+, serial console 115200)│
│                                                │
└────────────────────────────────────────────────┘

When I choose "Memory test (memtest86+)", I get a nice blue-background screen titled "Memtest86+ 5.01" with options and information (see attached file if you read this on the forum instead of the mailing list), but after a few seconds, that screen is replaced with rainbow-coloured flashing mojibake on a black background (see attached file) and stays that way.

Among the blue screen's options are "Press F1 to enter Fail-Safe Mode", "Press F2 to force Multi-Threading (SMP)" and "(c)configuration".
So, on my next try, instead of doing nothing, I hit the "c" and "F1" keys while the blue menu was showing, but still got the rainbow mojibake flashing on a black background.

So I still don't know how to memtest, nor whether I should want to.

Oh, and $ sudo update-grub which Magic Banana advised in the "Change boot device" thread gave me this:
Sourcing file `/etc/default/grub'
Generating grub configuration file ...
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
Found linux image: /boot/vmlinuz-5.10.20-gnu
Found initrd image: /boot/initrd.img-5.10.20-gnu
Found linux image: /boot/vmlinuz-4.15.0-112-generic
Found initrd image: /boot/initrd.img-4.15.0-112-generic
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
error: cannot seek `/dev/mapper/sda5_crypt': Invalid argument.
Found memtest86+ image: /memtest86+.elf
Found memtest86+ image: /memtest86+.bin
done

memtest_blue.jpg memtest_party_mode.jpg
Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

rainbow mojibake flashing on a black background

First time I see that!

Since Korax, Avron and you all use a D8 from Vikings and face the same weird troubles, I believe you should complain together to Vikings.

Connochaetes

I am a member!

Desconectado
Joined: 12/13/2017

Thanks!
And thanks to Avron and Korax for sharing their experiences here.

The rainbow mojibake wouldn't look out of place as a screensaver :)

Avron

I am a translator!

Desconectado
Joined: 08/18/2020

By the way, my system freezed again (including the panel clock's seconds counter, but usually excluding the cursor, which I can move fluently while unable to click or type anything) while I prepared to post this. I had to restart and rewrite my post (I had saved it as a TXT file before the freeze, but apparently it didn't survive the hard switch-off, the file was blank after rebooting) to post it here. I suspect Korax encounters similar freezes.

After every boot, a little while after logging in, I have a screen freeze that lasts about 1 or 2 minutes but I have not noticed that it happens again. I suppose this is purely a software issue, screen freeze can be due to a process taking all the CPU for some time. Since it is a rather minor inconvenience, I did not further investigate (the way would be to write down the time and check system logs around that time).

It is only twice that I really decided to reboot, every time it is while using Libreoffice for quite some time, I had no time to wait so I rebooted. Other than that, no problem.

Select boot device:

1. AHCI/3: CT2000MX500SSD1 ATA-10 Hard-Disk (1863 GiBytes)

2. Payload [nvramcui]

3. Payload [coreinfo]

I have the same if I press ESC early enough (indeed, when the message shows up it may be too late). I guess if you have more disks, there should be additional entries there. I just added disks, I'll check at next reboot whether it is the case. I tried the other two options already, this nvramcui and coreinfo, it provides a set of messages among which error messages but don't really understand what they mean. At next reboot I will have a look at them and take pictures and I will turn on my log collection on the serial line in case there is something there.

Besides, I had an explanation from Vikings that this is a one-time selection only. Doing a permanent change of booting disk requires reflashing coreboot. I am not sure how easy and safe it is to do that.

I want to say that, while I would appreciate a bit more flexibility on the configuration allowed by coreboot, I don't really have any unsolved issue.

Avron

I am a translator!

Desconectado
Joined: 08/18/2020

Today, I had a screen freeze with unresponsive keyboard but the mouse pointer still moving and I was able to connect via SSH. I tried to kill the mate-session process but it ended up totally non-functional, I lost the SSH connection.

At the time corresponding to this, syslog shows a "general protection fault" for Xorg, then another one for gunicorn one second after the ssh connection, then a third one for Caja a little later. I don't know what could be the reason.

I don't know what a "general protection fault" is as compared to a "segmentation fault" that I am more familiar with.

As for the temporary freeze:
- the last things before the time of the freeze is a set of "WebKitNetworkProcess[xxxx]: Memory pressure relief" and "WebKitWebProcess[yyyy]", 24 in total, with 2 values for xxxx and 2 values for yyyy
- perhaps a the end of the freeze, about 2 minutes later, there is
dbus-daemon invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=-900
From the log, the process killed is indicator-datet. I understand the issue is lack of memory. For indicator-datet, the total_vm is less than for each of the Webkit processes but the rss is more than 1000 times the one of abrowser, which is second for rss.

So it seems indicator-datet is the guilty process. Anyone knows what it is and why it is started?

Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

If that relates to the package "indicator-datetime", indeed installed by default, the description of the package is "A simple clock appearing in the indicator bar". You can apparently disable it from the "Startup Application Preferences".

Avron

I am a translator!

Desconectado
Joined: 08/18/2020

Thanks, I just did that and also removed the time and date display in one of the two Mate panels I have, I'll see if any issue at next boot.

About the "general protection fault", is that a user process or a kernel problem?

In my last freeze, I had:

kernel: [599057.101161] general protection fault, probably for non-canonical address 0xc5f96587149dc4f1: 0000 [#1] SMP NOPTI
kernel: [599057.101173] CPU: 7 PID: 1139 Comm: Xorg Not tainted 5.10.19-gnu #1.0
kernel: [599057.101177] Hardware name: ASUS KCMA-D8/KCMA-D8, BIOS 4.11-ab8edda14a622ab46bdfd01b877d75c7bd385a4d 01/16/2042
kernel: [599057.101186] RIP: 0010:__kmalloc_node+0xda/0x2b0

and then a lot more debug info. I don't know whether it is an issue limited to Xorg or whether it affects the kernel and then possibly other processes. In this case, the graphic screen was frozen, mouse and keyboard unresponsive, but I could establish an SSH connection which was functional after that until two more "general protection fault" 3 minutes later for different processes that time. I used the SSH connection to look for running processes and CPU/memory load, I saw nothng remarkable and finally decided to kill mate-session which moved the screen back to text mode but frozen and the SSH connection was then lost.

Any suggestion on further investigation or way to avoid this is welcome.

Magic Banana

I am a member!

I am a translator!

Desconectado
Joined: 07/24/2010

The kernels sends the messages. As far as I have understood, you have already tested different kernel versions. It looks like a hardware issue to me. Maybe a defective RAM (often the reason for random crashes) that cannot even be tested with memtest86+.