You are not logged in.
Hendrick, I don't get any delays when I ssh into my Asus netbook. If you want to try the same system I'm running, here's a live-iso I made for my EEE. Passwords are root:root and user:user and you'll need to use startx to get a desktop.
https://get.refracta.org/files/experime … 5_1348.iso
On my daedalus system, aptitude says it will install ceph and cephadm. How did you determine that it requires systemd?
If apt or apt-get won't let you install it, please post the terminal output so we can see what's wrong. Also, if you try it with aptitiude, you might be given some alternate solutions.
apt show ceph*Will give you information about every package whose name starts with ceph. None of them depend on systemd. One thing I'm not clear on is whether docker containers work in devuan.
There were no changes in the tests for efi partition in refractainstaller between chimaera and daedalus. (see code below)
Were you able to determine which device has no space left? Can you please post the output of df -h and fdisk -l?
If you have a uefi-bootable install, you can install all the others without a bootloader and then run update-grub (with os-prober enabled) in the bootable install to get all the others into the boot menu.
Here's the test for efi partition(s). Maybe someone will look at it and get a good idea.
# Check for UEFI boot and EFI partition
if [[ -d /sys/firmware/efi ]]; then
uefi_boot="yes"
esp_count=$(env LC_ALL=C fdisk -l | awk '/EFI System/ { print $0 }' | wc -l)
if [ -z "$gpt_list" ] ; then
gpt_message=$"There is no disk with a gpt partition table.
You should exit this script and run gdisk to create one for uefi boot."
fi
if [ "$esp_count" -eq 1 ] ; then
esp_dev=$(env LC_ALL=C fdisk -l | awk '/EFI System/ { print $1 }')
esp_dev_message=$"EFI partition found at ${esp_dev}
If this is not on the first hard disk, something may be wrong,
and you should investigate the situation."
if ! blkid -c /dev/null -s TYPE "$esp_dev" | grep -q "vfat" ; then
must_choose_esp="yes"
esp_dev_message=$"EFI partition found at ${esp_dev}
will need to be formatted FAT32"
fi
else
must_choose_esp="yes"
if [ "$esp_count" -eq 0 ] ; then
esp_dev_message=$" There is no EFI partition. You will need to create one."
elif [ "$esp_count" -gt 1 ] ; then
esp_dev_message=$"More than one EFI partition was detected.
You will need to select one. Normally, it's on the first hard disk."
fi
fiIgnore the missing final "fi". The rest of this code block just checks to see which grub is installed and generates some messages.
It's pretty much the same as before. The iInstructions for upgading devuan chimaera to daedalus are relevant. Note that firmware is now in a new section called non-free-firmware and add that to sources.list if needed.
https://www.devuan.org/os/documentation … o-daedalus
From the linked post with minor changes (version, sysctl.conf and lvm.conf). (General rule: take the new ones if they are different. Add or edit as needed.
Config files. Take defaults yes/no as indicated below.
/etc/issue No, keep old version and edit to change 11 to 12
/etc/pam.d/login Yes, install new version
/etc/sysctl.conf Install new version and add kernel.sysrq=1
/etc/initramfs-tools/initramfs.conf Yes, install new version. (edit to set COMPRESSION=xz)
/etc/ssh/sshd_config take new
/etc/cryptsetup-initramfs/conf-hook take new
/etc/lvm/lvm.conf take newoptional:
/etc/initramfs-tools/conf.d/resume (RESUME=none)
@aluma - are you using a live-iso or one of the installer isos? In either case, If you booted in uefi, then you don't tell grub where to put the bootloader. It knows to put it in the esp partition. If you're doing it on command-line, the commands are just
grub-install
update-grubTo be sure whether you booted bios mode or uefi, check to see if /sys/firmware/efi exists. It will be there in uefi mode but not in bios mode.
I ran into the same problem a few days ago. The upgrade to 6.6.9 kernel and adding the usrmerge package (I forget which order I did them) broke it. I could still boot the 6.5 kernel, but I couldn't get to the desktop. Some files for lvm2 and dmsetup were missing. In several stages, I reinstalled lvm2 dmsetup, xorg, xserver-xorg* (whatever was installed) and xinit, but I still couldn't use startx or lightdm. I reinstalled elogind libpam-elogind and policykit-1-gnome and that fixed it. (at least I think the last thing I did fixed it.)
Right. Firefox-esr does not require pulseaudio or pipiewire and works fine with just alsa. Getting it work with jack is a neat trick and is something I've wanted for years, so I was happy to see the solution here.
Refracta does not vary from Devuan very much. Most things are installed without recommends and without metapackages, to make it easier to remove anything you don't want. The only reason apulse is installed is for tor-browser, and sometimes ff-esr will need apulse if it runs after tor-browser.
Any suggestions for a small footprint front end like xarchiver?
I was about to ask you the same thing. So I asked the Duck instead, and top choices seem to be PeaZip (not in debian/devuan) or Engrampa.
On my full-desktop xfce system, engrampa wants to add 15mb and file-roller wants to add 21. Ark wants to add 200mb, but that's because it needs a lot of qt stuff that isn't installed. On my minimal lxqt system, ark wants 90mb, lxqt-archiver wants 7mb, xarchiver wants 1.7mb.
Have you seen any evidence that xarchiver can do anything with encryption? I'm not finding it with google and manpages. I know you can use encryption with zip/unzip. I usually use gpg if I want to encrypt a file.
If you feel like testing, here's a modified refractasnapshot script that handles editing the initramfs better. I've tested it in a VM with encrypted lvm and then again in the same system after adding amd64-microcode and intel-microcode. I did this in daedalus, which is using zstd compression for the initramfs. Seems to work fine.
Copy the file, make it executable and run it from your current directory:
sudo ./refractasnapshot -d (or as root without the sudo)
https://git.devuan.org/devuan/refractas … tasnapshot
I think your problem is related to the fact that after Beowulf, the su command got moved to another package. (to or from util-linux. I forget which) and they changed the way the PATH is handled. If you use 'su -' you can no longer open graphical programs on the user's desktop. To revert to the old behavior, do one of the following:
- Edit or create /etc/default/su and add the following line
ALWAYS_SET_PATH yesOR
- Edit the ENV_SUPATH line in /etc/login.defs to include /usr/local/sbin, /usr/sbin and /sbin
Then use just 'su' to get root. You might also notice that you don't change directory when you become root.
Just because someone posts something on the devuan forum does not make it the official word of devuan. Users get to speak for themselves.
I'm confused about what the problem is with your computer. You said you installed the latest Refracta, but then you're talking about kernels 6.5 and 6.6. The latest Refracta release is based on Devuan Daedalus, and it has a 6.1 kernel. I see that daedalus-backports has a 6.5 kernel but not a 6.6 kernel, so maybe the reason the 6.6 kernel doesn't work is because it wasn't built for the release that you're running.
There's a fix for refractasnapshot coming soon. Until then, another workaround is to remove the microcode package, make the snapshot iso, then install the microcode after you install a system on hardware. You could include the microcode packages in the iso to install them without a network connection if desired.
The Refracta iso not booting in VM is weird. Check the sha256sum to make sure the download was good. But that doesn't explain the other failures.
The script output suggests that you have a resume file in your initramfs. (remove_resume=yes) but I don't see a corresponding 'rm -f ...resume' in the section under "edit_initrd". Run lsinitramfs | grep resume $initrd where $initrd is the full path to the initrd in the excalibur iso. If you mount the iso, it will be in $mountpoint/live/initrd.img. Or you could retrieve it from the build host (assuming you set save_work=yes in refractasnapshot.conf). It would be in /home/work/iso/live/initrd.img.
You could also look in the build system for /etc/initramfs-tools/conf.d/resume and see what's in it. If it exists, make sure it says RESUME=none inside or delete it, and then run update-initramfs -u You only need the resume file if you want to resume from disk hibernation, and you usually don't want that in your snapshot.
Oh, that log doesn't have enough in it. Please run refractasnapshot in debug mode. For the cli script, run refractasnapshot -d and if you used the graphical version, start it from a root terminal (or sudo) as refractasnapshot-gui -d
If the snapshots are being made on a system with a swap partition or an encrypted root filesystem, that might be causing a problem. I'd need to see the debug log to be sure.
Looking at the varied selection of isos that are not working, I'm thinking maybe it's a problem with virtualization. But that's a guess.
Are you using a custom kernel? If so, compare your kernel config file to the one for the stock kernel to see if something is missing.
It may be something in the way the isos were made. I boot isos in qemu and vbox all the time. The only thing that doesn't work well is if the VM is set to do a uefi boot. In that case, I have to go into the efi setup and find the actual file to boot.
Take a look at /var/log/refractasnapshot.log or let me look at it.
@bobsmith
To become root, use su - and you will have root's PATH. Otherwise, you need to use the full path to the command: /usr/sbin/grub-install
If you post or paste or send me refractainstaller.log I might be able to see what the problem is. (If you use the cli installer, please start it as 'refractainstaller -d' to get a more verbose error log. Thanks.)
We removed the fork on live-build. Debian versions are/will be available in devuan repo. Ceres now has 1:20230502
If this was not the right thing to do, please let me know.
The extra zcat was needed at the time it was added because the microcode put an extra layer on the initrd.img. I guess that got fixed and snapshot needs to be changed again. That section of code also needs to support the new compression, zstd.
Thanks for posting this.
Hi and welcome to Devuan!
We only have around a dozen people maintaining forked packages, and that load keeps getting bigger. If you ever get the urge to start developing again, it would probably be easy for you to find something to do around here.
From what I've seen, migrations from debian to devuan are a crap-shoot. I think it depends on what all you have installed. Upgrades and fresh installs are a lot more reliable.
In general, I think devuan is more like debian than debian is. I think you'll feel at home here.
I'm not seeing any weird behavior with claws mail. I was going to post this a few hours ago and include the warning that my results might not match yours because I'm using lxqt with openbox. Now I see you're using the same.
I have the sapphire theme selected in the openbox settings (and in openbox configuration manager) and silver theme selected in lxqt preferences.
This is daedalus on ASUS EEE netbook, 1366x768 display.
I don't think it's normal behavior, but I'm not sure. My lxqt is minimal. I can't tell you off the top of my head which parts I left out, but I do have lxqt-powermanagement and lxqt-config-monitor. I do not have xscreensaver installed. I'm using openbox as the window manager.
lxqt-powermanagement is running. Watching top in an ssh session from another machine, I did not see lxqt-config-monitor run when the screen blanked. Screen blanking is set in the power manager settings. Nothing shows up in ~/.xsession-errors when the screen goes blank or when it comes back on.
Here's what I have installed with lxqt in the package name. Let me know if you think of something else you want me to check.
ii liblxqt-globalkeys-ui1:i386 1.2.0-1 i386 daemon used to register global keyboard shortcuts (ui files)
ii liblxqt-globalkeys1:i386 1.2.0-1 i386 daemon used to register global keyboard shortcuts (shared libs)
ii liblxqt1:i386 1.2.0-8 i386 Shared libraries for LXQt desktop environment (libs)
ii lxqt-config 1.2.0-1+b1 i386 LXQt system settings center
ii lxqt-globalkeys 1.2.0-1 i386 daemon used to register global keyboard shortcuts (appl.)
ii lxqt-notificationd 1.2.0-1 i386 LXQt notification daemon
ii lxqt-panel 1.2.1-1 i386 LXQt desktop panel
ii lxqt-policykit 1.2.0-1 i386 LXQt authentication agent for PolicyKit
ii lxqt-powermanagement 1.2.0-1 i386 power management module for LXQt
ii lxqt-qtplugin:i386 1.2.0-3+b2 i386 LXQt system integration plugin for Qt
ii lxqt-session 1.2.0-3 i386 session manager component for LXQt
ii lxqt-system-theme 1.2.0-1 all System theme for LXQt
ii lxqt-themes 1.2.0-1 all Themes for LXQtLooks like it's working correctly. You get automatic login in the live-isos. You picked a nox iso, so there's No X. You can run commands. In the live session you can use sudo for any commands or you can use su to become root. The root password is root and the user password is user in case you need it.
Next you should run mocp and listen to half a minute of Dark Star.
If you want to install it, sudo refractainstaller will work.
Yes, it's isohybrid. I installed grub-efi-amd64-signed and shim-signed, which pulled in a couple other things. I assume the kernel is signed because there is no kernel package linux-image-*-signed, but there is an -unsigned kernel package. I did not install the -unsigned.
Edit:
When I get to fast internet, I'll download debian-live to compare.
OK, that's nice that you don't have an overheating problem, but the point I was trying to make is that if you want to reduce cpu load, you can stop letting a bunch of strangers use your cpu. Javascript is executable code that gets run on your computer and is provided by people you don't know.