Drivers – An AMDGPU-PRO tale about Debian (2019) – What should I do if I install AMDGPU-PRO and can not even get a startup screen?

There are a number of discussions on StackExchange on AMDGPU-PRO (such as (here) (1), (and here) (2)), but there seems to be a number of variables that can go wrong, and in my only case, none of the wires I found here or anywhere else could help me. In order to help other people who might follow my sad approach, I thought I should document this for posterity.

OS: Debian 10
GPU: AMD / Radeon rx580, also with Intel Integrated Graphics

I bought a GPU a long time ago with the hope of doing a bit of mining. I was a newcomer to programming and I did not really know how to make it work. She was there for a while. In light of recent material successes, I decided to try my old GPU. I've had the GPU running using (this) (3) wonderful tutorial that configures the GPU using Mesa. The test suite (and the noise) confirmed that the graphics processor was working. Yay! But I wanted to be able to exploit, and I quickly discovered that the extraction programs all used a different driver – AMDGPU-PRO. I thought it might not be too difficult and I downloaded the latest version (which might be important afterwards) from the AMD website.

The installation was not too difficult and I restarted.

I did not even have a splash screen.

I could tell that the computer was running because I had an external drive connected to a service system, and its relaxing flashing light meant that he was reading and writing as he was supposed to do it. Sometimes I got a loading screen that said things like:

( OK ) Started GNOME terminal...

But that would depend on the very example that I gave. I tried to connect to it blindly (in order to uninstall the drivers), but it did not work. The LED on my hard drive is not flashing at all, which makes me think that my keyboard controls were not working at all.

Starting in the BIOS menus will not show anything on the screen, although my external hard drive can not boot, I guess it would load the BIOS menu.

The classic wisdom is to keep SHIFT or ESC at startup to access a command line, but in my infinite wisdom, I use a wireless keyboard. I still have not gotten my hands on a PS2 keyboard, although I borrowed a USB keyboard and could not access GRUB with it.

I've thought of trying to boot from a live USB drive, but apparently, my BIOS does not give priority to USB boot. Good try.

So I was there, no screen, no keyboard. How to operate my computer? I tried IRC (from my other machine) and I was advised to erase everything and start over. Not that I'm sure how I would even get there.

linux – Can not install mongodb on debian 10 with the help of official instructions

Following the official documentation

sudo apt-get install -y mongodb-org

results

The following packages have unmet dependencies:
 mongodb-org : Depends: mongodb-org-shell but it is not going to be installed
               Depends: mongodb-org-server but it is not going to be installed
               Depends: mongodb-org-mongos but it is not going to be installed
E: Unable to correct problems, you have held broken packages.

I've deleted an earlier version of mongodb using purge as the service did not want to start, but I can not reinstall it anymore.
I'm trying to manually install the dependencies

The following packages have unmet dependencies:
 mongodb-org-shell : Depends: libcurl3 (>= 7.16.2) but it is not going to be installed
E: Unable to correct problems, you have held broken packages.

debian – How to install the linux version of VSCode on an ARM64 Chromebook?

Do I have the following error message because my Chromebook is equipped with an ARM64 chip instead of Intel?

Reading package lists... Done Building dependency tree
Reading state information... Done Note, selecting 'code:amd64' instead of './code_1.38.1-1568209190_amd64 (1).deb' Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. The following information may help to resolve the situation:

The following packages have unmet dependencies: code:amd64 : Depends: libnotify4:amd64 but it is not installable Depends: libnss3:amd64 (>= 2:3.26) but it is not installable Depends: apt:amd64 but it is not installable Depends: libxkbfile1:amd64 but it is not installable Depends: libsecret-1-0:amd64 but it is not installable Depends: libgtk-3-0:amd64 (>= 3.10.0) but it is not installable Depends: libxss1:amd64 but it is not installable E: Unable to correct problems, you have held broken packages.

I received the above message after downloading the .deb file to the address https://code.visualstudio.com for my Chromebook with linux (Debian) enabled and entering this command in Terminal:

sudo apt install ./code_1.38.1-1568209190_amd64 (1).deb

Microsoft provides instructions for "Debian and Ubuntu-based distributions" at the following address: https://code.visualstudio.com/docs/setup/linux.

Network – Debian Bridge 10 Disappears After Restart

I've created a bridge in my Debian router 10 like this one.
# brctl addbr br0 then I add a network interface on my bridge # brctl addif br0 eno1

brctl show
bridge name bridge id       STP enabled interfaces
bridge-lan      8000.0030bdb2810d   no      eno1

my /etc/network/interfaces looks like this

# The loopback network interface
auto lo
iface lo inet loopback

# Set up interfaces manually, avoiding conflicts with, e.g., network manager

iface eno1 inet manual

# Bridge setup
auto bro

iface br0 inet dhcp
      bridge_ports eno1

Everything works fine, but if I restart my router, brctl show returns an empty result.
In my syslog I have this strange line:
bridge: filtering via arp/ip/ip6tables is no longer available by default. Update your scripts to load br_netfilter if youd nedd this is this normal?
PS: I have disabled ipv6 in sysctl.conf with this line net.ipv6.conf.all.disable_ipv6 = 1

What should I do to have my bridge when my router reboots?

debian – 20 watts difference between C states enabled in BIOS and same states in Intel_idle?

I'm trying to reduce the standby consumption of a Dual-Haswell-EP server manufactured by Intel. If I activate the C states in UEFI-Bios, the minimum power consumption is about 80W. However, if I disable the C states in the BIOS and start the system, the minimum power consumption never drops below 100W. (everything else identical, same microcode, same frequencies, same version of Bios)

This is surprising, because in both cases, after starting Debian, the intel_idle The driver supports the control reported by / sys / devices / system / cpu / cpuidle / current_driver. I see no reason, why energy consumption should be different? powertop and turbostat reports the same C9 state at 99.9% for all cores.

Absorbed power is measured via sensors or the integrated BMC.

For reasons unrelated to this question, I would like to start with C states disabled, then leave the intel_idledriver takes over. Is there another implicit difference with the deactivation of C states in the Biuos and a way to achieve the same minimum energy consumption?

linux – Debian Buster only slows on one terminal

Please: my main terminal is an HP Compaq LA1905wg and my other terminal is (an old) HP 1702.

The applications work normally on the main terminal, but if I move them on the 1702, my whole system begins to explode. Gnome-shell consumes more than 100% of the CPU and when I hit a key, it repeats itself dramatically. "l" becomes "lllllll".

I use Gnome as delivered, nothing has been changed.

the public IP address of the Debian server no longer responds to anything after the remote upgrade

after an upgrade and an upgrade, my public IP address vps does not work at all (i can just login to ssh with which means the public IP address is correct)

I've checked the firewall connection with the ufw command and its ok, but there is no ping success and tcp response.

I've also checked the network interface file and the resol.conf file and everything seems to be going for the better and the server has the internet. I can install packages or ping other URLs.

before the update, everything worked but I can not even link a built-in php web service to a port on a public ip

I do not know if any changes were made to iptables when the server was being updated or not, and I do not know how to check if n works with iptables.
so what are you proposing to solve this problem?

(the debian version is 9)

the smallest jupyterhub on debian [on hold]

I have only one virtual machine under Debian. Jupyterhub must be installed so that almost 30 users have access to jupyternotebooks. The smaller Jupyterhub meets the needs, can I just install it? Because it is said that Ubuntu 18.04 is required.

Greet Patrick

encryption – Install Debian on a remote dedicated server without sending clear text identification information

As stated in the title, are there any best practices for installing Debian on a remote server, without sending clear text passwords on the link?

For example, many dedicated server vendors will give their customers direct access to the KVM of the supermicro card. This gives them a lot of choice – for example, I can mount any local ISO image as if it were in the CD drive.

However, the installation of Debian would not be easy, because I should type the root password during the installation, what an eavesdropper can also read.

My hypothesis would be to create a predefined installation image, with the SSH public key already installed. You would no longer depend on the contents of the KVM screen, but you would immediately connect to the SSH service. You would always open up to man-in-the-middle attacks because someone spying on the KVM connection could copy the public key and the host key and intercept the connection, but I imagine that it would be difficult to do in time.

Is there a better way to miss me?

debian – authby = secret serves as an alternative to AH in ipsec.conf

I used strongSwan U5.5.1 to install an IPsec tunnel between two Debian hosts with a firewall between them. my ipsec.conf precise authby=secret but no ah=. The firewall currently supports ESP (Protocol 50) and UDP / 500 (ISAKMP), UDP / 4500 (ISAKMP), but not yet AH (Protocol 51).

My question is: if I want to prevent middle-man attacks, is that authby=secret is it enough or do I need to turn AH on? In other words: do authby= and ah= to process the same type of authentication or different types of authentication in the IPsec protocol stack?