You are not logged in.
Updated iso:
snapshot_chimaera_01-20200702_1212.iso
https://get.refracta.org/files/experimental/
This one is bigger because it has both 5.6 and 5.7 kernels (and headers).
It boots 5.7 by default. To boot 5.6 you need to press TAB or e to edit the boot command. Change vmlinuz and initrd.img to vmlinuz.1 and initrd.img.1
Also added crda, wireless-regdb, ca-certificates, firefox-esr.
To get rid of the network delay on boot, edit /etc/network/interfaces and change 'allow hotplug' to either 'auto' or 'manual', or else you could apply the changes to /etc/init.d/networking as described here:
https://dev1galaxy.org/viewtopic.php?pid=15493#p15493
I don't know about that old bug report for 4.15 kernel. Someone else was having problems with the 5.6 backports kernel, maybe related:
https://dev1galaxy.org/viewtopic.php?id=3621
I added crda and wireless-regdb to that system and previously added ca-certificates. I'll upload a newer iso soon. (I almost deleted it yesterday and then changed my mind!)
I installed that iso in a qemu VM in uefi mode. I can't get it to boot at all without booting from the iso and using grub commands to boot the virtual hard disk. That's a qemu issue, I'm pretty sure. Anyway, it boots like that, and when I install gnome, it still boots that way, and it's pretty fast.
If you try installing from that iso, user the cli installer, not the graphical one in the apps menu. I ran into gtk problems with that. Just run 'refractainstaller' as root.
https://get.refracta.org/files/experime … 0_1456.iso
sha256sum:
dc97d884f860b12791f40fbe4d1a4b04a1a2e2cbe4e937dd4a64d6ed598998be refracta-test-oblx_5.6bpo_openrc-20200630_1456.isoOne way to get more information is to edit /etc/udev/udev.conf and set udev_log="debug" and make sure it's uncommented.
If you want to try with the system that I'm using for testing, I can upload a live-iso later today.
Install haveged.
Use apt, apt-get or aptitude remove Package.
You probably won't need to name all of them, as some will come out automatically with others.
Aptitude will most likely remove more than the others.
You can do 'apt autoremove' afterward to get rid of unneeded packages that weren't already removed.
'apt-get autoclean' will get rid of the obsolete packages in /var/cache/apt/archives
'apt-get clean' will get rid of all the packages in /var/cache/apt/archives
After the packages have been uninstalled/removed, can I, then, copy the previously downloaded non-dmo .deb files into "/var/cache/apt/archives/" before executing apt-get dist-upgrade with the Laptop not connected to the Internet from the chroot environment?
I have no idea, but it sounds like you know what you're doing. I've never done it that way.
I guess the plan would be to remove whatever is blocking the upgrade and whatever else you don't need. You can probably remove all or most of the dmo packages. And anything with "deb9" in the version is for stretch/ascii.
Cool! I'm glad you like it.
I neglected to make an announcement, but the beowulf version of refracta is no longer in beta. Refracta10 is here! Some bugs mentioned in this thread have been fixed or workarounds are in place.
That was discussed here: https://dev1galaxy.org/viewtopic.php?id=3520
Isos are here:
https://get.refracta.org/files/stable/
Add 'sleep 1' to /etc/init.d/eudev as shown here:
https://bugs.devuan.org/cgi/bugreport.cgi?bug=483#10
That should take care of 1 and 2. If 1 is still a problem after that, install acpi-fakekey.
I don't have any ideas for 3, but maybe that will get fixed, too.
If you want to test this before you do it, you can run (as root):
udevadm trigger --action=add
and see if that gives you sound. This won't persist beyond a reboot. Editing the init script will.
The question mark and square bracket are the prompt. I got the same. And it did what I said!
And more...
This required some minor hacking to get around dependency issues. When sysvinit gets removed, live-config-sysvinit and refractasnapshot go with it. I copied all the files into place and it works. There needs to be a live-config-runit package.
user@refracta:~$ cat /proc/1/comm
runit
user@refracta:~$ df -h
Filesystem Size Used Avail Use% Mounted on
udev 480M 0 480M 0% /dev
tmpfs 99M 392K 99M 1% /run
/dev/sr0 965M 965M 0 100% /run/live/medium
/dev/loop0 927M 927M 0 100% /run/live/rootfs/filesystem.squashfs
tmpfs 494M 440K 493M 1% /run/live/overlay
overlay 494M 440K 493M 1% /System was installed in a VM, and the live-iso was made with refractasnapshot.
The iso boots in qemu and also isohybrid usb boots on hardware. Stuff works.
This is in beowulf. I think aitor may have done this already. I'm surprised at how easy it was.
Maybe...
https://files.devuan.org/devuan_beowulf … _notes.txt
Power buttons are disabled on the lightdm login screen with elogind.
(See Debian bug [#932047](https://bugs.debian.org/932047))Add the following line to /etc/pam.d/lightdm-greeter
session optional pam_elogind.so
Here's the test in veracrypt:
FILE* pipe = popen("sudo -n uptime 2>&1 | grep 'load average' | wc -l", "r"); I've seen this test before, thanks to dzz (the other refracta dev):
refractasnapshot-wrapper.sh:15:sudo_allowed=$(sudo -n uptime 2>&1 | grep load | wc -l)If the output is '1' then the user has sudo privs.
I'm seeing the same configs for crda and regdb in 4.19.0-9-amd64, 5.6.0-0.bpo.2 and 5.6.0-1 in ceres.
CONFIG_CFG80211_CRDA_SUPPORT=y
CONFIG_CFG80211_REQUIRE_SIGNED_REGDB=y
CONFIG_CFG80211_USE_KERNEL_REGDB_KEYS=yThis is just a wild guess, but maybe you are being affected by this bug:
https://bugs.devuan.org/cgi/bugreport.cgi?bug=483
Simple test is to add sleep 1 to /etc/init.d/eudev as shown in the diff.
@fsmithred: The link to your modified version stopped working, I assume this one is supposed to be the current one:
Yeah, that's it. 'devuan-packages' on git changed to 'devuan' on the new server.
git.devuan.org is now the same place as gitea.devuan.dev. The old git is at gitlab.devuan.org in case anyone wants to retrieve something before it goes away completely.
Thanks. I fixed the link in my earlier post.
Yes, you can use an installed system.
Wow. I think (hope) that 120 is the correct number of dmo packages installed and not 633.
You can get rid of sources.list.d/devuan.list. I don't know about the other lists. You might need to disable them temporarily and/or you might need to change the suite they're using. (stretch to buster?)
It looks like your dmo packages are older versions than what beowulf has. If you remove dmo from sources, you should be able to upgrade most of them to stock beowulf versions.
If there's something in dmo you need that isn't in the devuan repo, you can add it later. I highly recommend pinning dmo to a lower priority or disabling it altogether after you install what you want. That will prevent pulling in other stuff from dmo that you don't need or want. Just avoid dmo if you can.
And yes, you can do all this in a chroot from another system. I like to use a live-cd (or usb) for that, but rescue mode in the installer isos will also work.
Thanks. I'll have to play with this when I have time. Extra modules were added when I was testing on my uefi toshiba laptop and much of the boot messages were not being displayed.
The empty isolinux.cfg is a know characteristic. If you use refracta2usb to make a multiboot live-usb that uefi capable, it will actually put an empty isolinux.cfg on the usb so grub can find it. All the other isolinux files get renamed as syslinux in that case.
I have a couple of ideas, but nothing certain.
Maybe you could remove libgstreamer-plugins-bad1.0-0 and eliminate the conflict. Two packages should not provide the same file. Maybe the newer version doesn't have the conflicting file.
You really should remove deb-multimedia from your sources. Look in /etc/apt/sources.list, look for any files in /etc/apt/sources.list.d/ and also wherever synaptic stores its sources. That alone might fix the upgrade problem, or maybe you have a bunch of other packages from dmo that might cause problems.
dpkg -l |grep dmowill give you a list of files with dmo in the name or the version. The ones with dmo in the version are from deb-multimedia. See how many you have.
How old is this system? Was it upgraded from jessie? From a jessie beta?
Indeed!
You can delete everything in /isolinux but isolinux.cfg (and even everything in isolinux.cfg) in order to get grub's menu.
Yeah, grub looks for isolinux.cfg just to find the right location. It doesn't care what's inside it. I forgot about that.
This one boots uefi:
https://get.refracta.org/files/experime … l-live.iso
sha256sum:
8655c89c68a454458e3d9ea1c2e0b266315f72a2f81f6e043309a4d334d0a893 devuan_beowulf_3.0.0_amd64_uefi_minimal-live.isoEdit: I didn't modify the grub menu, so it does not have the same choices as the normal minimal-live. The first choice is the same.
Sounds like you didn't make a grub.cfg.
Look at the xorriso command around line 1078 - https://gitea.devuan.dev/devuan/refract … tasnapshot
Mount the iso, rsync copy it to a different directory, add the efi files, run xorriso. I don't think it's less pain in the ass to do it that way. Quite the opposite.
Also easier than explaining how to make an iso is to make one, which I am doing right now. I'll let you know if it boots and where you can download it.
The minimal-live is not set up to boot uefi. If you have another linux set up on the machine, you can boot the minimal-live on usb from grub command-line. It's also possible to install it as is and use the first linux to boot this one.
If you want this to be the main or only linux, you would need to install grub-efi-amd64 before you run refractainstaller.
If you don't have another linux to boot this, it might be possible to do it with two usb sticks - one with desktop-live to get to a grub command line and one with the minimal live. I didn't try that.
Press c at grub boot menu to get to command prompt and enter the following commands.
set root=(hd0) # Your motherboard might call it hd1. Tab-complete will help you.
linux /live/vmlinuz boot=live username=devuan
initrd /live/initrd.img
bootI tried to reproduce your problem and failed. Maybe it's because I'm testing in qemu instead of hardware, maybe it's specific to your hardware, and maybe it's specific to the version of openrc that you're using. I can't find that one. I'm using 0.40.3-1 in beowulf with the current 5.6 backports kernel.
Where did you get your version of openrc? (0.42-1~bpo10+1)