rwp | Is there an equiv to https://deb.debian.org/debian/dists/buster/main/installer-i386/ that can be used for virt-install --location to netboot an install ?? | 00:28 |
---|---|---|
gnarface | rwp: http://deb.devuan.org/devuan/dists/beowulf/main/installer-i386/ | 00:31 |
rwp | gnarface, Awesome! I failed to figure it out on my own. Thanks for the second brain. | 00:33 |
gnarface | no problem | 00:33 |
gnarface | i dunno what you're talking about so i don't know if it'll actually work but the files seem to be there... | 00:33 |
rwp | virt-install is part of libkvm, a rather heavy KVM virtual machine management framework. --location allows a fully netboot installation. | 00:36 |
rwp | Alternatively I can download an ISO image and boot the image and install that way. | 00:36 |
gnarface | interesting | 00:36 |
rwp | What's really very cool is that if you have either a local mirror or a local caching proxy and everything local then you can spin up VMs in literally just a few minutes from start to login. | 00:38 |
rwp | That path you gave me seems to be working. I am able to netboot from it. And I am currently working through an install right now. | 00:40 |
rwp | Which means you have enabled me to spend some endless many hours, days, weeks, fixing my automation for Beowulf. :-) | 00:41 |
gnarface | cool | 00:42 |
gnarface | i think? | 00:44 |
rwp | I know, right!? It's a sucking hole of time. But automation really is worth the time investment. preseed files plus system configuration scripts and all of that to automate everything. | 00:45 |
rwp | A few hours of trial and error can save you several minutes of looking at the README file. :-) | 00:47 |
rwp | And just casually working through a serial console installation I just finished it and am now logged into a pristine new Beowulf VM. | 00:55 |
rwp | And if one does not want the serial console (I would call that server style) then virt-manager is the fully graphical GUI. | 00:55 |
mexon | Hi Folks, I know the general answer to this is 'when its ready' which is completely understandable but I'm looking for a general approximation of when Devuan's current testing branch will move to stable, given that Debian 11 is pretty close at this point | 02:38 |
ShorTie | i'd say give it a month or 2 after debian stabilizes | 02:39 |
mexon | Fair enough, excellent, appreciate it. | 02:40 |
systemdlete | error during dist-upgrade from ascii to beowulf in eudev: https://paste.debian.net/1191923 | 05:45 |
systemdlete | I've run --fix-missing but after applying additional upgrades (ones that got missed due to the breakage), the same error comes back. | 05:45 |
lts- | systemdlete: I've used https://www.devuan.org/os/documentation/dev1fanboy/en/buster-to-beowulf successfully, and that states "[apt-get install evdev] The last command is known to cause package breaks but we will fix this as part of the migration process" and then instructs to run "apt-get -f install". I haven't had issues with that. | 06:08 |
lts- | i.e. --fix-broken | 06:09 |
gnarface | systemdlete: do you have a custom kernel or a very old one or one from a different distro or something? | 06:20 |
gnarface | systemdlete: i haven't seen this error but it's the type of thing i'd expect to see if you had compiled the kernel without inotify | 06:20 |
gnarface | or like, maybe if you forgot to upgrade the kernel to the current release | 06:21 |
pandakekok9 | Hi, can anyone look as to why my build of iwd for experimental is stalling? https://jenkins.devuan.dev/job/devuan-package-builder/499/console | 06:35 |
rrq | pandakekok9: seems to be something lost in the final handshake w i386-builder | 07:47 |
* rrq looking ... | 07:59 | |
gnu_srs | systemdlete: The comment about buster-to-beowulf is wrong. eudev is installed by a package udev depending on eudev. You should not need to force the install. | 09:44 |
gnu_srs | Anybody know hos to update that page? | 09:46 |
gnu_srs | how** | 09:46 |
Xenguy | gnu_srs, Yes, there has been discussion of that page recently... | 11:20 |
Xenguy | But what to correct exactly is always the question : -) | 11:21 |
gnu_srs | Xenguy: Regarding eudev this should be sufficient: Upgrade to eudev is made by the transitional package udev. No special measures are needed. | 12:39 |
gnu_srs | So all text surrounding apt-get install eudev and apt-get -f install can be removed. Replaced with the text above. | 12:40 |
Xenguy | Thanks for that gnu_srs . Will capture that idea and run it by a few people. | 19:47 |
infobot | Xenguy: pas de quoi | 19:47 |
fsmithred | wtf? the bot now initiates conversation? | 20:39 |
sixwheeledbeast | ~botsnack | 20:40 |
crashoverride | in french, too | 20:40 |
rwp | I for one welcome our new info bot overlords. | 20:48 |
ibanja | I have upgraded several systems to beowulf. Most kept my apt-mark holds, but one did not. | 20:52 |
ibanja | Is there a way to downgrade and keep those old versions? | 20:53 |
ibanja | It is important for compatibility reasons. | 20:53 |
plasma41 | ibanja: Downgrade an individual package or the whole system? | 20:54 |
ibanja | an individual package | 20:54 |
plasma41 | ibanja: That's often doable. Which package? | 20:55 |
ibanja | I am thinking a better option would be a user installed version of the package. | 20:55 |
ibanja | the two important packages are keepassx and lyx | 20:55 |
plasma41 | ibanja: You're wanting to continue using the ascii versions of those packages, correct? | 20:56 |
ibanja | yes | 20:56 |
plasma41 | ibanja: Ok, let me check the dependencies for those packages. One moment | 20:58 |
crashoverride | :q | 20:58 |
crashoverride | oops | 20:58 |
ibanja | plasma41: thanks... I am also wondering if using NIXos packaging or other universal packaging system such as flatpack would be the best option, but I have no experience and don't want to trash my system. | 21:00 |
plasma41 | ibanja: The changelog for keepassx indicates that the ascii version (2.0.3-1) and beowulf version (2.0.3-2) differ only in terms of minor tweak to the debian packaging, but are otherwise the same upstream version (2.0.3). (Source: https://tracker.debian.org/media/packages/k/keepassx/changelog-2.0.3-2) Is there a change affecting you that is not mentioned in the changelog? | 21:05 |
ibanja | WOW... I must be using pre-ascii... sorry for the missinfo... (Installed: 0.4.3+dfsg-0.1+deb8u1 ---- Candidate: 2.0.3-2) | 21:06 |
ibanja | the problem is some old keepassx databases... and they are shared via nextcloud.... | 21:07 |
ibanja | I think it's time to let the straglers upgrade or have a second install to deal with the problem.... OR user install an older version if that is even possible. | 21:09 |
plasma41 | ibanja: Ok, that version of keepassx is from jessie. | 21:09 |
ibanja | yes | 21:09 |
ibanja | I could try to cull together the dependencies and put them into my local repo, but I wonder if this is getting a bit silly... | 21:11 |
ibanja | ??? | 21:11 |
plasma41 | ibanja: Looking at the build dependencies for the jessie version of keepassx (https://snapshot.debian.org/archive/debian/20151208T035513Z/pool/main/k/keepassx/keepassx_0.4.3%2Bdfsg-0.1%2Bdeb8u1.dsc) I see that Qt4 is required. Unfortunately Qt4 is not included in Beowulf so you would need to manually pull in all the Qt4 dependencies. Simply re-installing the old keepassx deb file wouldn't be enough. | 21:15 |
ibanja | plasma41: I see... | 21:16 |
ibanja | plasma41: Thanks... I have a solution now... the computers that can't upgrade are stuck in version because of legacy software... so let's put them into a VM and let the host os handle the upgraded programs | 21:16 |
ibanja | I think that would work. | 21:17 |
ibanja | plasma41: appreciate the helpful input... | 21:19 |
plasma41 | ibanja: As for lyx, while the ascii binary depends on older ascii versions of some dependencies, namely libboost, I believe a simple recompile of the older version atop beowulf should provide a usable binary. | 21:20 |
ibanja | plasma41: lyx... that would be very helpful. I will look into that. Thanks | 21:21 |
plasma41 | ibanja: Happy to help | 21:22 |
systemdlete | gnarface: I was following the official instructions at debian.org to upgrade to beowulf. | 21:41 |
systemdlete | Also, mine is ascii-to-beowulf, not buster-to-beowulf. I suspect that buster already has a "base" load of software that is more compatible with beowulf than ascii. | 21:42 |
systemdlete | So if upgrading the kernel is a necessary step, then that should be a dependency that the upgrade finds automatically. If a reboot is necessary, that should be clear from the procedure. | 21:43 |
systemdlete | I have whatever the latest stock kernel is from the ascii repo. I think it is 4.19.0 | 21:44 |
fsmithred | upgrade kernel is not required | 21:48 |
systemdlete | I am unable to use my laptop in its current condition. | 21:48 |
systemdlete | It fails, the same way, every time I try to finish the upgrade. There is plenty of disk space. | 21:48 |
fsmithred | I've never seen that error before | 21:49 |
systemdlete | I am even using ethernet rather than wireless to speed things up. | 21:49 |
fsmithred | did you verify that those things are in your kernel? | 21:49 |
systemdlete | Again, I am using whatever stock kernel is in the repo | 21:49 |
fsmithred | grep INOTIFY /boot/config-whatever | 21:49 |
systemdlete | 4.19.0 | 21:49 |
fsmithred | that's nice | 21:50 |
systemdlete | is that no good? | 21:50 |
fsmithred | did you verify that those things are really in it? | 21:50 |
fsmithred | it should be good | 21:50 |
fsmithred | prove it | 21:50 |
fsmithred | not necessarily to me, but to yourself | 21:50 |
systemdlete | A stock kernel that is not compatible with stock packages? Odd. | 21:50 |
fsmithred | would be very odd | 21:50 |
fsmithred | you can take it on faith or you can verify | 21:51 |
systemdlete | It says CONFIG_INOTIFY_USER=y | 21:52 |
fsmithred | yeah, so does mine | 21:52 |
* systemdlete scratches his head | 21:55 | |
fsmithred | that error message is coming from the preinst script | 21:55 |
fsmithred | egrep -q "^[a-fA-F0-9]+ T \.?sys_${symbol}$" /proc/kallsyms | 21:56 |
fsmithred | that's in the failing test | 21:56 |
* systemdlete listens as fsmithred steps through debugging the failing package... | 21:57 | |
fsmithred | can you boot a live-usb and chroot the system? | 21:58 |
fsmithred | or boot to single-user? | 21:58 |
systemdlete | If you know of a way to work around this, maybe we can run apt-get with some magic options | 21:58 |
fsmithred | not magic options. Just comment out the test. | 21:58 |
systemdlete | How does one "comment out the test" without building a custom package? | 21:59 |
systemdlete | Or will building a custom package be relatively straightforward? | 22:00 |
fsmithred | nano /var/lib/dpkg/info/eudev.preinst | 22:01 |
systemdlete | won't that get overridden when I attempt the install again? | 22:01 |
systemdlete | or does it pick up where it left off? | 22:02 |
systemdlete | (I'm not very familiar with this, if you can't tell) | 22:02 |
fsmithred | you might get away with dpkg --configure -a | 22:02 |
fsmithred | or even upgrade might not download it again | 22:03 |
fsmithred | I'm not certain | 22:03 |
fsmithred | anyway | 22:03 |
fsmithred | nano +33 <that file> | 22:03 |
fsmithred | and you'll be on a line that says 'exit 1' | 22:03 |
fsmithred | preced that with # | 22:03 |
fsmithred | precede | 22:03 |
systemdlete | Yes, I know how to comment it out. | 22:03 |
systemdlete | And how to use vi. | 22:04 |
systemdlete | (I don't use nano) | 22:04 |
fsmithred | you'll still see the error messge go by, but it won't dump you out | 22:04 |
fsmithred | fine | 22:04 |
systemdlete | ok | 22:04 |
fsmithred | the exit is right after the list of kernel features | 22:05 |
systemdlete | I don't see that line in the preinst file | 22:06 |
systemdlete | You are saying line 33? | 22:06 |
systemdlete | (I'm guessing that is what you meant with +33) | 22:06 |
fsmithred | yeah, you should see the error message you got | 22:06 |
systemdlete | nope | 22:06 |
fsmithred | near the top of the script | 22:06 |
systemdlete | Are we really talking about the same package | 22:06 |
fsmithred | you must be looking at the older one | 22:07 |
systemdlete | and script? | 22:07 |
fsmithred | eudev | 22:07 |
fsmithred | second function is called check_kernel_features | 22:07 |
systemdlete | soooo... the upgrade somehow downloaded the wrong package? | 22:07 |
fsmithred | no, you're probably looking at the ascii version | 22:08 |
fsmithred | # Hack to get the dpkg process' PID despite using debconf | 22:08 |
fsmithred | ^^^ see that at the top? | 22:08 |
systemdlete | why would it be doing that -- this is an upgrade to beowulf, and the packages should be coming from beowulf since the sources file was set -- by me -- to the beowulf repos. | 22:08 |
fsmithred | I thought it was an upgrade FROM ascii | 22:09 |
fsmithred | see the line I just posted from the ascii version? Is that what you have? | 22:09 |
systemdlete | Yes | 22:09 |
systemdlete | It IS an upgrade from ascii to beowulf | 22:09 |
fsmithred | so anything that didn't get replaced by the beowulf version should be the ascii version | 22:10 |
systemdlete | right. so why is it upgrading eudev if it is already at the ascii version? If beowulf has a newer eudev package, it should have downloaded it. | 22:14 |
systemdlete | Sorry, this is still not making sense. | 22:14 |
systemdlete | I don't see any "exit 1" in the eudev.preinst file | 22:14 |
fsmithred | because we are not looking at the same file | 22:15 |
systemdlete | I am looking at the precise path you gave above | 22:15 |
fsmithred | omg, you have ascii, I have beowulf | 22:15 |
systemdlete | Maybe I misunderstnad | 22:15 |
systemdlete | no kidding | 22:15 |
systemdlete | LOL | 22:15 |
fsmithred | the two different versions do not have the same preinst file | 22:16 |
systemdlete | no surprise there... | 22:16 |
systemdlete | so? | 22:16 |
fsmithred | I don't know why the upgrade failed for you | 22:16 |
systemdlete | Right. I think that part I get. Completely. | 22:16 |
systemdlete | :D | 22:16 |
systemdlete | were you being sarcastic when you said "omg, ..." ? | 22:17 |
systemdlete | or did you really just realize that we were on different releases? | 22:18 |
fsmithred | no, but you keep asking why our files are not the same | 22:18 |
systemdlete | (peoples' tones are hard to detect on chat) | 22:18 |
systemdlete | I did not ask, actually. | 22:18 |
systemdlete | I expect that they would be different, by default. | 22:18 |
systemdlete | esp. for a package like eudev, being the pandoras box it is | 22:19 |
systemdlete | fsmithred: I'm still trying to find the webpage that I am using to do the upgrade | 22:20 |
fsmithred | should be on devuan.org | 22:20 |
fsmithred | devuan.org/os/install | 22:21 |
systemdlete | it is. All I recall is that it said to modify the sources file, then run update and then release-upgrade (something like that) | 22:21 |
systemdlete | not that page, no | 22:21 |
fsmithred | change sources, apt update, apt dist-upgrade is what I usually do | 22:22 |
systemdlete | right | 22:22 |
systemdlete | that was it | 22:22 |
systemdlete | and it broke, exactly as we have been sorting this out | 22:22 |
fsmithred | oh there are a couple of extra instructions: kill xscreensaver, force-downgrade upower | 22:24 |
fsmithred | repeat dist-upgrade, apt -f install | 22:24 |
fsmithred | not necessarily in that order | 22:24 |
fsmithred | and I don't think any of those things would affect your error | 22:24 |
fsmithred | anyway, I have to go | 22:24 |
systemdlete | (they don't) | 22:24 |
systemdlete | thanks for your help | 22:24 |
* systemdlete is left with a broken system just by following official debian instructions for performing a release upgrade. Go figure, right? | 22:25 | |
sadsnork | ibanja, I am successfully using keepass 0.4.3 on Beowulf by (1) "apt-get install --no-install-recommends libqt4-xml libqtcore4 libqtgui4" (2) dpkg -i /path/keepass.deb (3) apt-mark hold keepassx | 22:35 |
gnu_srs | systemdlete: Did you follow the Devuan instructions for an upgrade? I assume you did not mean Debian! | 22:39 |
systemdlete | I went to debian because I thought that the upgrades would be the same | 22:51 |
systemdlete | uh-oh? | 22:51 |
systemdlete | But it sounds, from what fsmithred was saying, that the same instructions apply | 22:52 |
xrogaan | Is this ... an April fool? https://security-tracker.debian.org/tracker/underscore | 23:06 |
xrogaan | oh, no, it's real. It's a node package. | 23:07 |
DPA | I've now for the first time installed chimaera on a normal laptop (using debootstrap). The keyboard-configuration package doesn't seam to work anymore. And OMG, xfce now looks awful, the CSD, the icons, etc... | 23:55 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!