You are not logged in.
*If* you added it back with the *same* UID and GID then everything should work OK. But if not you will have problems.
Did it's home directory change? If not then all the files you had beforehand should be in there. Look at them with ls -al and check they show user and group names in the output.
If not it's possible to chown them to your new id with something like find . -uid 1000 -exec chown chris:chris {} \; (replace 1000 with the old UID, chris:chris with your new userid and group and chown with echo for the first run).
*** WARNING *** I've not tested this, though I've had to do it many years in the past.
Should you have non-free-firmware in your sources.list? I don't think that would cause this problem but might cause problems with hardware support later.
I've recently found a new command most which is like less but automatically shows binary data in dump format. I've installed it on all my systems so I have some chance of making sense of binary data. But that's just a fallback although very worth having.
As far as Red Hat goes I've made a mental note to never use it again.
I suspect the PREEMPT_DYNAMIC kernel. I have one system I use to run CUDA (on a Nvidia card) where I have to boot it with the default kernel, not the preempt one, to get CUDA to work.
If you have an option to boot the 6.1 default kernel could you try it to see if that's the problem?
Could we have a few more details? What version of devuan are you running? What version of Firefox?
If not sure post output from cat /etc/os-release and copy whatever Help -> About Firefox shows (or post a screen shot).
nvidia-persistenced has no real purpose unless you want to use CUDA.
It's not essential for CUDA either. Both the CUDA apps I use (gmp-ecm and msieve polynomial selection) work without it running.
If I were to run an AV on Linux, I'd just go back to Windows. It would defeat the whole purpose (one of them, at least).
The obvious use is to scan files etc that have to be processed on a Windows system. It's a lot harder for a virus to infect the system doing the scanning if that's a different OS. And possibly a different CPU, eg a Raspberry PI.
I have got a system that runs CUDA on the GPU, but uses the onboard graphics to run the display. (It shares a monitor with 3 other systems via a KVM switch.) But for historical reasons it runs openSUSE 15.3 so I'm not sure how much I can help, beyond saying it is possible on some systems.
It has a GeForce GTX 970 with CUDA cuda-drivers-510 but a rather old motherboard (can't find the manual right now) in case that helps.
It's been there for a while:
chris@rigel:~/bin$ ls -l /boot/config-*
-rw-r--r-- 1 root root 186853 Jun 30 2022 /boot/config-4.9.0-19-amd64
-rw-r--r-- 1 root root 186567 May 7 2018 /boot/config-4.9.0-6-amd64
$ grep -i fanotify /boot/config-4.9.0-*
/boot/config-4.9.0-19-amd64:CONFIG_FANOTIFY=y
/boot/config-4.9.0-19-amd64:CONFIG_FANOTIFY_ACCESS_PERMISSIONS=y
/boot/config-4.9.0-6-amd64:CONFIG_FANOTIFY=y
/boot/config-4.9.0-6-amd64:CONFIG_FANOTIFY_ACCESS_PERMISSIONS=y
This is on a system that's still running ascii. The older config will be off the installation DVD, which is why it's from May 2018.
ziomario@Z390-AORUS-PRO-DEST:~$ ssh -Y marietto@192.168.1.6
where 192.168.1.6 is the IP assigned to the Chromebook (it is connected through the wi-fi network),but I get the following error :
Bad owner or permissions on /home/ziomario/.ssh/config
The message probably refers to /home/ziomario/.ssh/config on the system you are connecting from, not on the destination.
My advice to anyone with a similar problem would be to try ssh -v which would produce a lot more messages from ssh, which should tell you enough to solve the problem. If still stuck post *full* output here which should give us enough information to work out what the problem is.
I learnt that debuging ssh problems at work (although we had enough different OSes and ssh versions thatgetting them all to work properly took quite a while).
It might be worth using free and vmstat to see how much your swap file is being used. Eg leave vmstat 10 running in a terminal window while you try to start the slow version of firefox and see what it shows for si and so (swap-in and swap-out). If they sit at zero then firefox must have something else wrong with it.
I would suspect something in /etc/profile or ~/.bash_profile, ~/.bash_login or ~/.profile might be causing the wait. I'd start by putting echo "In .bash_profile" at the start of ~/.bash_profile (or ~/.bash_login or ~/.profile, whichever you have) and see if that comes out before or after the pause. If it's after the pause (or you don't have any of ~/.bash_profile, ~/.bash_login or ~/.profile) then I would make a similar temporary change to /etc/profile to see if that's causing the pause. Then trace through the relevant script to find out what's causing the pause. Adding set -x to the script near the start could be helpful.
Keep notes of what you change so you can change it back when you have stopped investigating.
@entropyagent - he is obviously trying to log onto the system that is running now via ssh. As the following output shows he managed to connect OK:
debug1: Connecting to devuan [127.0.1.1] port 22.
debug1: Connection established.
But I can't tell why it won't accept his password. The next place to look is in syslog, /var/log/messages should contain messages from sshd which might be revealing.
@devur - are there any special characters in your password on devuan that could get corrupted by different keyboard mappings? That's the only thing I can think of that could go wrong (I assume you are typing the *same* password as you use to log on to devuan)?
Setting up passwordless access should be possible. ls -al ~/.ssh/ should get something like:
$ ls -al ~/.ssh/
total 28
drwx------ 2 chris chris 4096 Jun 19 20:22 .
drwxr-x--- 30 chris chris 4096 Jul 16 06:13 ..
-rw-rw-r-- 1 chris chris 1959 Jun 19 20:22 authorized_keys
-rw-rw-r-- 1 chris chris 43 Jun 5 21:12 config
-rw------- 1 chris chris 2602 Oct 30 2021 id_rsa
-rw-r--r-- 1 chris chris 565 Oct 30 2021 id_rsa.pub
-rw-r--r-- 1 chris chris 4092 Jun 5 19:52 known_hosts
If one of the lines in authorized_keys is a copy of id_rsa.pub then you should be able to log on without needing a password. Although ssh is touchy about permissions, debug output should tell you if they are wrong.
the login offer me to login to my regular user account and dos nor accept server login
Try ssh -v devuan which should provide a lot of messages about the logon process. And look in /etc/ssh/sshd_config and /etc/ssh/sshd_config.d/ (grep -v '^#' /etc/ssh/sshd_config to exclude comments should make it easier to see the relevant entries).
Post output from the commands here if you are still stuck.
Although I have had to resort to:
ps -ef | grep sshd
(assume that shows sshd is /usr/sbin/sshd.)
strings /usr/sbin/sshd | grep -i config
Just to be sure I was looking at the right config file. But that was working as a UNIX system admin when we had a wide range of OSes to look after.
I'd try running a 'live' version of chimaera, & if it boots OK, install it.
It would probably boot up, but would not have CUDA installed. And installing CUDA needs a reboot, whihc loses any updates done to a live version. I'd have to try installing chimaera, install CUDA and see if it works with my GPU. Then reinstall ascii from backups if CUDA doesn't work on chimaera. Which is more work than I want to go to for what is probably a wild goose chase.
I now seem to be back at "Can't upgrade from ascii because Nvidia have dropped support for running CUDA on my GPU". So I'll probably keep this system on ascii until the GPU dies. But at least I'm not short of disk space any more since the new disk is 1Tb and the old disk was 332Gb.
I found that if you try to run apt ... once you have started synaptic you get an error message about not being able to get a lock. That would probably cut both ways so trying to start synaptic while an apt job or similar is running would probably fail. And if the lock didn't get cleared when someting ended you might well get this effect.
But I can't remember details and I'm doing something to recover lost data from a faulty disk that I dare not risk interrupting so I can't experiment now. Try apt update and see if it works (that updates the system, but it should be safe). If that fails it might give out an informative message.
df -i show you the number of inodes in a filesystem. df --total show how many 1k blocks of disk space there are (in both cases you get total, used and available). See man df for details.
It's possible to run out of inodes but not space. This makes it impossible to add any new files to the filesystem. Which can be very confusing if you don't realise what's happened (I fell into this trap, though the disk was faulty and bought a new disk before realising I only needed to tidy up junk files).
Also you don't usually need to run df as root, it should work just as well as a non-root account.
The hard disk on my Devuan system has failed, so I've had to replace it. I've recovered most of the contents of /home, but the OS has gone. So I've got to re-install Devuan on the new (and larger) disk.
I've got two DVDs from Linux Format with Devuan on them, an old one with ascii and a newer one with chimera. But for some reason the installer on the newer DVD won't work, it's probably not compatible with the rather old motherboard etc. So I've had to install ascii to get the system to boot up at all.
When I tried to update it I got errors because http://pkgmaster.devuan.org/merged doesn't contain ascii any more. I found the readme pointing me to http://archive.devuan.org/merged which gets a little further.
But that hasn't got ascii-updates:
root@rigel:~# apt update
Get:1 http://archive.devuan.org/merged ascii InRelease [29.4 kB]
Ign:2 http://archive.devuan.org/merged ascii-updates InRelease
Get:3 http://archive.devuan.org/merged ascii-security InRelease [29.8 kB]
Err:4 http://archive.devuan.org/merged ascii-updates Release
404 Not Found [IP: 88.198.100.113 80]
Get:5 http://archive.devuan.org/merged ascii/main amd64 Packages [7,193 kB]
Get:6 http://archive.devuan.org/merged ascii/main Translation-en [5,652 kB]
Get:7 http://archive.devuan.org/merged ascii/contrib amd64 Packages [50.8 kB]
Get:8 http://archive.devuan.org/merged ascii/contrib Translation-en [45.9 kB]
Get:9 http://archive.devuan.org/merged ascii/non-free amd64 Packages [78.8 kB]
Get:10 http://archive.devuan.org/merged ascii/non-free Translation-en [79.3 kB]
Get:11 http://archive.devuan.org/merged ascii-security/main amd64 Packages [741 kB]
Get:12 http://archive.devuan.org/merged ascii-security/main Translation-en [382 kB]
Get:13 http://archive.devuan.org/merged ascii-security/contrib amd64 Packages [1,776 B]
Get:14 http://archive.devuan.org/merged ascii-security/contrib Translation-en [1,759 B]
Get:15 http://archive.devuan.org/merged ascii-security/non-free amd64 Packages [14.1 kB]
Get:16 http://archive.devuan.org/merged ascii-security/non-free Translation-en [21.2 kB]
Reading package lists... Done
E: The repository 'http://archive.devuan.org/merged ascii-updates Release' does not have a Release file.
N: Updating from such a repository can't be done securely, and is therefore disabled by default.
N: See apt-secure(8) manpage for repository creation and user configuration details.
Is it worth trying to do a series of updates from ascii or should I try to do complete reinstall of a later release? If it's worth updating ascii should I just comment out ascii-updates in /etc/apt/sources.list ?
The program that re-creates it probably deletes it before re-writing the latest contents into it. If you own .local/share/ then a program, running as you can delete files from it, even if you don't own them.
You might be able to stop this by making .local/share/ owned by root. But this would break anything else trying to create a file in .local/share/
golinux wrote:@chris2be8 . . . Thanks for the great description. Found it . . . I think . . .
Which one is it? Just curious
The one I'm referring to was in issue 269, so it should be number 109 (number 099 was in issue 259). But that site won't show it.
I have a nasty suspicion you can't see the cartoons in the last year (you probably need a subscription). Sorry.
I've not used doas (I learnt sudo before doas was available), but you will need to think hard about what you want to do. If you let people run an arbitrary command as root they could just run bash as root to get the same effect as doas -s.
With sudo it's possible to let people run a list of commands as root, but not a shell, as long as none of the commands you allow can spawn a shell. But it's easy to let a way to a root shell slip through.
Forcing people to jump through a few hoops to get to a root shell *might* have some value, but it's more likely to give you a false sense of security.
I get the paper magazine, so I don't know if it's available online or where it would be. Sorry.
It's the Elvie cartoon if you want a search term.
Google's name comes from nonsensical babbling from a mathematician's child
You should look at the cartoon in the April 2023 issue of Linux Magazine (Linux Pro Magazine in North America). It has an alternative meaning for the name.
Compare what groups the users are in on the faulty system and both the working systems. And how sudo (or doas) is configured on them.