blakep | Hi! I'm part of the team that runs mirror.aarnet.edu.au. We've been having some trouble since 2023-01-17 with getting updates from files.devuan.org via rsync over IPv6 (IPv4 works fine). Connection over IPv6 works ok at first, but a few kb into the sync we get a disconnect from remote. Is this the correct place to get help with that, or shoudl I ask eslewhere? | 00:01 |
---|---|---|
rrq | I'll have a look.... | 00:04 |
rrq | mmm there's rsync limiting in place, but afaics the same for ipv4 and ipv6 | 00:08 |
rrq | bb|hcb would know better but might be rejuvenating right now | 00:11 |
blakep | Here are some additional details that may help: https://pastebin.com/EiDQPEzx | 00:12 |
blakep | timestams in UTC | 00:12 |
rrq | right; this end gets: deflate on token returned 0 (18966 bytes left) | 00:17 |
rrq | consistently, and only on that ipv6 | 00:19 |
rrq | that ip is not rsync-limited | 00:21 |
rrq | no, wrong, ... let me add an escape for it | 00:22 |
blakep | oh nice! thankyou | 00:23 |
rrq | I think it's automated but maybe some table is wrong | 00:24 |
rrq | when bb|hcb is young enough he'll look into it | 00:25 |
bb|hcb | Give me some time and I will check. In the meantime, blake, can you for it to use ipv4 (like -4) to see what happens? | 00:28 |
bb|hcb | s/for/force/ | 00:28 |
blakep | I ran using -4 just before asking here and everyting worked perfectly | 00:31 |
blakep | src ip 138.44.164.198, started Jan 22 22:27:47, completed at Jan 22 22:52:35 | 00:33 |
systemdlete | I have two computers running. One is my "workbox" and the other is my "testbox." I have virtualbox installed on both. On the workbox, there are no problems launching VMs after boot. On the testbox, I cannot launch a VM until I either (1) open the GUI and launch it from there, or (2) copy the exact string the icon is trying to launch. | 00:35 |
systemdlete | Now I know there are vbox haters here. Thank you so much for that. But let's put that aside, OK? This might be some desktop config problem. | 00:36 |
systemdlete | Both boxes are running the latest xfce4. | 00:36 |
systemdlete | Workbox has a FX8350 with 32GB, testbox has an Athlon II Thuban X6 with 16GB. Plenty of space available on both. | 00:37 |
rrq | probably due to PATH or some other environment setting | 00:37 |
systemdlete | rrq: OK, good thought. I'll check that. | 00:38 |
rrq | (I assume it's headless) | 00:38 |
systemdlete | The hosts are both connected to the same keyboard, video, and mouse via an electronic switch. | 00:38 |
rrq | ok; but the vboxes are headless? | 00:40 |
systemdlete | no, not really. | 00:43 |
systemdlete | I mean, by default, when you launch them, there is a virtual "head" if you will. | 00:43 |
systemdlete | They CAN be made headless though. | 00:43 |
systemdlete | But for our purposes, the two machines are configured similarly for vbox. However, your suggestion shows that there are some differences. It will take me a few minutes to sort this out. | 00:44 |
systemdlete | (with sdiff, etc) | 00:44 |
rrq | if they need console display window then the x auth might play a role as well; though should be same issue on the hosts I suppose | 00:46 |
systemdlete | XDG_SESSION_TYPE=tty on the workbox; =x11 on the testbox | 00:51 |
systemdlete | Also, there is a RUNLEVEL=2 on testbox, but no such variable on workbox | 00:51 |
systemdlete | iirc, the testbox was installed from refracta. I think workbox was installed from devuan ISO | 00:52 |
systemdlete | DESKTOP_SESSION is differnt also. xfce on workbox, default on testbox. | 00:53 |
systemdlete | ah, INIT_VERSION is set on testbox, but not on workbox | 00:54 |
rrq | idle browsing suggests vbox does look at XDG_SESSION_TYPE | 00:55 |
systemdlete | aha. So I can do an experiment. See if I can "force" the testbox to launch the first VM by setting that var locally on the cmd line | 00:55 |
systemdlete | hold on (have to switch boxes) | 00:56 |
systemdlete | I'm going to do this differently. I'm going to run a script from the desktop(s) to see the acctual environments those are getting. | 00:59 |
systemdlete | Apparently, cmd line and desktop envs on each box are not identical. Then see what the diff is between the two boxes. | 00:59 |
rwp | By what method are the VMs launched? From the command line? Or maybe it sounds like from an init script? | 01:03 |
rwp | And basically the same question for X on each system. Do you log in through the interactive "xdm" (gdm/gdm3/lightdm/slim)? Or does it auto-start X? | 01:04 |
rwp | Also... How is networking started? NM won't connect until the Desktop Environment is started and the user logged in. Networking is probably needed for the VM to start. | 01:06 |
systemdlete | ok. Let me give you the overview. | 01:08 |
systemdlete | There are several ways to launch vbox VMs. One is from their GUI. And that works faithfully; no problems there. | 01:08 |
systemdlete | on either box. | 01:08 |
systemdlete | I can also launch them from the command line, faithfully, on either box. | 01:08 |
systemdlete | what I cannot seem to do is launch the very first VM (after reboot, or some amount of time goes by that one of their system tools stops running automatically) | 01:09 |
systemdlete | I can do this without issue on the workbox, but not on the testbox | 01:09 |
systemdlete | host-side networking is not needed for the VM to start, no. | 01:10 |
systemdlete | at least, I've never heard of that. | 01:10 |
systemdlete | If host-side networking is unavailable, the VM will still run, but its internal notion of the networking world outside of it will be limited. | 01:10 |
systemdlete | (not sure if that is what you are asking) | 01:10 |
systemdlete | now, rwp and rrq, here's "the thing" (as our prez would say): The cmd line invocation I use--successully!-- on the testbox is exactly copy-and-pasted from the icon! | 01:11 |
systemdlete | IOW, that calling sequence works successfully when performed from the cmd line, but not from the icon. | 01:12 |
systemdlete | The icon, btw, is created by their GUI, not by me. | 01:12 |
systemdlete | Btw, also, I tried to get help from them yesterday, but I supposed it was a slow day at virtualbox and the only "help" there was this incredibly nasty and arrogant soul who kept badgering me rather than trying to really help. | 01:13 |
systemdlete | (This is why I love devuan's irc; so much more pleasant to work with! thanks again) | 01:13 |
systemdlete | I thought they had booted that about a year ago, but apparently they just chose a new nick. | 01:14 |
systemdlete | The temperment seems to be about the same... but anyway | 01:14 |
systemdlete | I may post this on their forum. | 01:14 |
rwp | I am not sure what copied from icon means but I presume that means the command line associated with the icon which is the icon.freedesktop file Exec line. | 01:15 |
systemdlete | I just thought maybe someone here might have a clue since our OS does not depend on systemd. | 01:15 |
systemdlete | yes, rwp, that's what I meant. | 01:15 |
systemdlete | which gives me an idea rwp | 01:16 |
rwp | I don't use vbox but I do use libvirt and the command line there is ten miles long at the least. | 01:16 |
systemdlete | maybe I should compare the icon files generated on each system. Even though both run xfce4 (same version too) | 01:16 |
rwp | The vbox.freedesktop icon file does not come from XFCE but would come from VirtualBox so that depends upon VB rather than XFCE. | 01:17 |
rrq | afair I used "VBoxManage startvm blaha" or "VBoxHeadless startvm blaha" | 01:17 |
systemdlete | Well, the command line for vbox is just calling a utility with the title bar name for the VM and a long hash represening the VM. | 01:17 |
rrq | it all did require the vboxnet service to have been started | 01:19 |
systemdlete | yeah, it is like vboxmanage --comment titlebarnameIwantForVM "{abcdef-99175-blah...}" | 01:19 |
rwp | libvirt example command line for ONE vm start: https://paste.debian.net/plain/1268176 | 01:19 |
systemdlete | Yes, the problem is that the VBoxSVC daemon is not started | 01:20 |
systemdlete | (when the problem occurs) | 01:20 |
systemdlete | the rest of the time it starts, faithfully | 01:20 |
systemdlete | but let me finish the env comparison... | 01:20 |
rwp | I would also A-B compare the two systems for presence and use of eudev and elogind too. | 01:21 |
rwp | I could imagine that cgroups (via {,e}logind) might come into play too. | 01:22 |
rwp | Also I have lost sync with the details above on how they are started that causes problems. Copy-pasting the icon properties Exec line always works. So it is the clicking on the icon that fails? | 01:24 |
rwp | Or is it failing from an init script? If from an init script then I would also look to see exactly when during the init boot they are run. Dependencies might be insufficiently declared on one and starting one before an undeclared dependency is ready. | 01:25 |
systemdlete | well, vboxmanage handles the whole launch processing | 01:26 |
systemdlete | it's not like qemu | 01:26 |
systemdlete | or libvirt | 01:26 |
rwp | If it is a difference between clicking and copy-pasting then the terminal shell's environment for bash is loaded through the .bashrc file but that's not going to be seen by clicking the icon. | 01:26 |
systemdlete | good point | 01:26 |
systemdlete | I shall look at my .bashrc file | 01:26 |
systemdlete | on each system that is | 01:26 |
rwp | Debugging these things I often hack the Exec line to call a script that I create that dumps the environment and other debug to a file in /var/tmp/rwp.debug.out and then see what's different. | 01:27 |
rwp | I dump to a file in /var/tmp because /tmp is (usually) cleared on reboot and I don't want to lose the data if I reboot. | 01:28 |
rrq | the failing system should also have some complaint, eg in syslog | 01:29 |
rrq | I think starting a vbox VM does depend indirectly on networking being started | 01:33 |
xrogaan | popularity contest isn't working here: popularity-contest: unable to submit report to http://popcon.devuan.org/cgi-bin/popcon.cgi. | 01:34 |
xrogaan | though, exim does show emails sent to survey@popcon.devuan.org | 01:35 |
xrogaan | using the wrong From: header. | 01:36 |
rrq | do you have an example to show? | 01:38 |
xrogaan | MAILFROM wasn't populated by default. Does popcon expects an return email? | 01:41 |
rrq | (idk) it looks at "From" header | 01:46 |
rwp | Data sent by email must be properly email formatted which means that From lines and such must be acceptable to be mailed. | 01:47 |
rwp | But doesn't it try to POST the data first to the http server? If that isn't working... Why not? I would look there. | 01:47 |
systemdlete | Ok, thanks for all the feedback. I'll keep looking into this. It's probably something fairly simple. Maybe something I changed on the testbox to actually test somthing and I forgot to put it back later on... as will happen in test environments. | 01:50 |
systemdlete | I think I might even re-set the entire testbox. If I do, I will install the snapshotting software for the system so I don't have to continually go through the install process. | 01:50 |
rwp | systemdlete, If you figure it out please post an update because I would like to know. Thanks! | 01:52 |
rwp | xrogaan, I see that /usr/share/popularity-contest/popcon-upload has a -d debug option. But running it myself I have yet to see any output from it. | 01:53 |
systemdlete | rwp: np. will do. thanks again to all of you for your patience as I struggle through these issues. | 01:53 |
rrq | right .. the popcon host had a dist-upgrade last week | 01:53 |
xrogaan | rrq: do you receive anything on the host? | 01:54 |
xrogaan | as in, broken for everybody or just me? | 01:54 |
rwp | I see that popcon logs to /var/log/popularity-contest and may contain more debug information there. | 01:54 |
rrq | well, there were nginx errors .. (missing DOCUMENT_ROOT) | 01:55 |
rrq | and nginx got touched by the upgrade | 01:55 |
rwp | rrq, So... Problem on the Devuan popcon server side? | 02:00 |
rrq | yes, likely | 02:01 |
rrq | how do I test it? | 02:02 |
xrogaan | run popcon? | 02:02 |
xrogaan | it's normally a cronjob | 02:03 |
rwp | I think it is called like this (adding -d for interactive debug): /usr/share/popularity-contest/popcon-upload -u http://popcon.devuan.org/cgi-bin/popcon.cgi -f /var/log/popularity-contest.new -d | 02:03 |
rwp | Oh, I see that's *after* it has generated the data. Initially on a fresh install there won't be any data to upload. | 02:04 |
xrogaan | it's in /etc/cron.daily/popularity-contest | 02:05 |
rwp | Looks lie (as root) generate the data: /usr/sbin/popularity-contest --su-nobody > /var/log/popularity-contest.new | 02:06 |
rwp | Then the popcon-upload will upload it. I just did that. No output using -d for it on my system. | 02:07 |
rrq | ta | 02:07 |
xrogaan | my SUBMITTED variable in that script never changes from "no" | 02:07 |
xrogaan | so, sendmail is failing (reportedly) but exim does show an email going out in the logs | 02:08 |
rwp | Have you configured TOR for example which changes the upload protocol? | 02:08 |
rwp | Also if you just run /etc/cron.daily/popularity-contest it won't be the right DAY="0" for it and it will exit without doing anything. | 02:08 |
xrogaan | I have TOR installed, but not running. I'll change the variable to "no" | 02:08 |
rwp | See /etc/popularity-contest.conf for the DAY setting as well. | 02:09 |
rwp | The cron.daily script looks to dynamically decide to use TOR if /usr/bin/torsocks and /usr/bin/torify are both installed. | 02:10 |
xrogaan | but it never worked though, and I don't get the "unable to submit report to $URL" error | 02:11 |
xrogaan | anyhow, made changes, we'll see at next cronjob | 02:11 |
rwp | if USETOR=no then that overrides the dynamic logic and avoids using tor. | 02:11 |
xrogaan | yes | 02:12 |
rrq | sorry. got sidetracked.. so your email submission seems to report fail locally although an email gets delivered? | 02:33 |
xrogaan | something like that | 02:41 |
xrogaan | which is weird | 02:42 |
xrogaan | nevermind. I read the errorlog wrong. | 02:44 |
xrogaan | was just the http issue (probably 'cause of TOR) | 02:44 |
gnstaxo | How do I search for a package that contains X file like this website but from command line | 03:20 |
gnstaxo | https://packages.debian.org/search?mode=path&suite=bullseye§ion=all&arch=any&searchon=contents&keywords=gcrypt.h | 03:20 |
xrogaan | gnstaxo: `apt-file search' | 03:31 |
golinux | Also https://pkginfo.devuan.org/cgi-bin/policy-query.html | 03:38 |
golinux | Oops that's not command line but will find if a pkg is available . . . | 03:38 |
gnstaxo | Thanks guys | 03:51 |
gnstaxo | I have another question | 04:01 |
gnstaxo | The other day I installed a i386 package and ever since whenever I apt update I get the list from the i386 repo | 04:02 |
gnstaxo | But I am no longer interested in i386 | 04:02 |
gnstaxo | Do I have to uninstall all of my i386 packages? | 04:03 |
gnstaxo | So that I stop getting those updates | 04:03 |
rrq | nope.. did you set up multiarch? | 04:04 |
gnstaxo | I don't know | 04:04 |
gnstaxo | I put in some commands wine told me | 04:04 |
rrq | ok. that's a clear aybe, then :) | 04:05 |
rrq | add the setting "[arch=amd64]" after "deb" in /etc/apt/sources.list | 04:05 |
rrq | (no double-quotes) | 04:05 |
rrq | like: deb [arch=amd64] http://deb.devuan.org/merged .... | 04:06 |
rrq | the same for anything in /etc/apt/sources.list/* | 04:07 |
rrq | the same for anything in /etc/apt/sources.list.d/* | 04:07 |
gnstaxo | Let me try | 04:07 |
gnstaxo | rrq: It worked! Thank you! | 04:10 |
rrq | note that it might end up uninstalling i386 packages when corresponding newer amd64 packages get installed | 04:12 |
rrq | apt wants to keep arch versions of pacakges in sync | 04:12 |
gnstaxo | I'm getting "E: Removing essential system-critical packages is not permitted. This might break the system." | 05:01 |
gnstaxo | But the packages I want to remove were installed along with wine | 05:01 |
gnstaxo | I checked the log | 05:01 |
gnstaxo | I uninstalled wine and I was left with deps I cannot remove | 05:02 |
gnarface | gnstaxo: usually this happens when an alternate package is swapped in with different dependencies, you have to figure out what got replaced when you installed wine, and switch it back | 05:07 |
gnarface | (is my guess, anyway) | 05:07 |
gnarface | it helps to not install "recommends" | 05:07 |
gnarface | just helps make things easier to keep track of | 05:08 |
gnstaxo | gnarface: I have both the 32bit and the 64bit version of each package | 05:14 |
gnstaxo | gcc-12-base:i386 libgcc-s1:i386 libc6:i386 libidn2-0:i386 zlib1g:i386 | 05:15 |
gnstaxo | Is it not possible to force removal? | 05:16 |
gnstaxo | But I'm gonna update the whole system before forcing anything | 05:16 |
gnarface | gnstaxo: i'm sure it's possible to force removal | 05:39 |
gnarface | gnstaxo: you definitely need libc6 though | 05:41 |
gnarface | you might only need the 64-bit version but none of the 32-bit stuff will work without the i386 version | 05:41 |
cws6969 | I have two OS , One is devuan_chimaera_4.0.0_amd64_desktop-live , It's called LIVE_OS | 11:08 |
cws6969 | The other is built using debootstrap , It's called DB_OS | 11:08 |
cws6969 | Both systems can be accessed via wired network and can open any web page | 11:08 |
cws6969 | both systems can connect to wifi , But LIVE_OS can open any web page, while DB_OS cannot open any web page | 11:08 |
cws6969 | This is strange, is DB_OS missing some software? | 11:08 |
xinomilo | is the clock right? what's the output of `ping $somewebpage ` from db_os ? | 11:12 |
cws6969 | The clock is right , ping wechat.com >>"Domain name resolution fails temporarily" . It's the same for any URL | 11:28 |
xinomilo | so, dns issue | 11:30 |
xinomilo | if you can ping some public resolver (eg. ping 94.16.114.254 - opennic) , then add this ip in /etc/resolv.conf : "nameserver 94.16.114.254" and retry ... | 11:32 |
cws6969 | resolv.conf Configuration error ?? | 11:33 |
xinomilo | you can add any resolver ip you prefer, just an example.. | 11:33 |
xinomilo | what's in that file now? | 11:33 |
xinomilo | configuration error, or broken dns | 11:33 |
xinomilo | don't know what dns debootstrap sets up ... | 11:34 |
cws6969 | The odd thing is that the wired network is ok , The braincase hurts | 11:42 |
Ryushin | Does Devuan include a systemd unit file to sysv init script tool? I found this: http://www.trek.eu.org/devel/sysd2v/ | 12:52 |
Ryushin | It is supposed to be part of the sysv source now in the contrib directory. | 12:54 |
Necrodiver | i have one really quick question before i need to get to bed: I have a pc arriving tomorrow with different specs than the one im on currently. Can i just take out the drives from this pc and put them in the new one and everything will work okay? | 14:06 |
Necrodiver | or will i need to do a clean install? | 14:08 |
djph | Necrodiver: "maybe" | 14:18 |
djph | Necrodiver: worst thing will be graphics and wifi. Other than that, it should be pretty okay | 14:18 |
xinomilo | install all firmware drivers before moving drives, just to be sure everything will work.. (or the ones needed for new pc if you know exactly what...) | 14:22 |
Wonka | meh... $ freecad | 22:37 |
Wonka | freecad: error while loading shared libraries: libpyside2.cpython-310-x86_64-linux-gnu.so.5.15: cannot open shared object file: No such file or directory | 22:37 |
gnarface | Wonka: do you know which package it's supposed to come from? | 22:40 |
Wonka | I guess freecad has some dependency broken; i've upgraded from Python 3.10 to Python 3.11 packages recently | 22:45 |
gnarface | i'd try an apt-file search for libpyside | 22:47 |
gnarface | maybe it's still there but just with a slightly different name and you can fix it with a symlink | 22:48 |
Jjp137 | this could be kinda related? not the exact versions though: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1013881 | 22:48 |
Wonka | this exists: /usr/lib/x86_64-linux-gnu/libpyside2.cpython-311-x86_64-linux-gnu.so.5.15 | 22:48 |
gnarface | looks definitely related | 22:48 |
gnarface | i vote you symlink it and see what happens | 22:49 |
gnarface | backup your freecad data first just in case | 22:49 |
Wonka | yep, as I said, freepython should depend on the python 3.10 version and have a newer package that depends on the python 3.11 version... | 22:49 |
Wonka | Jjp137: yep, same issue with previous python version | 22:51 |
Wonka | waiting for aptitude to finish trivial updates first... | 22:51 |
Wonka | breakage with libreoffice again because something is unavailable O.o | 22:52 |
gnarface | this is all happening on sid? | 22:52 |
gnarface | er, ceres? | 22:52 |
Wonka | unstable | 22:52 |
Wonka | libreoffice-java-common (>= 2:7.4.4~) (UNAVAILABLE) | 22:58 |
Jjp137 | hm I see 4:7.4.4-7 here: https://packages.debian.org/sid/libreoffice-java-common | 23:05 |
fsmithred | olicy | 23:08 |
Wonka | libfreecad-python3-0.20 0.20.1+dfsg1-2+b1 is installed and depends on libpyside2-py3-5.15 (>= 5.15.2), which is satisfied by 5.15.8-2+b1, which contains /usr/lib/x86_64-linux-gnu/libpyside2.cpython-311-x86_64-linux-gnu.so.5.15 | 23:11 |
Wonka | libfreecad-python3-0.20 0.20.1+dfsg1-2+b1 also depends on libpython3.10 (>= 3.10.0) | 23:11 |
Necrodiver | djph: I dont use wifi so that wont be an issue, and my current pc has an integrated cpu, the new one will have a dedicated one, but the pc im moving it to is a "downgrade" because I wanted a core 2 quad and not an i3. | 23:12 |
Wonka | oh f*ck... did another "apt update", now libreoffice upgrade fails on something else... libreoffice-common 4:7.4.4-7 breaks libreoffice-core >= 1:7.5~, which includes libreoffice-core 4:7.4.4-7, which every other part of libreoffice depends on... | 23:18 |
Wonka | someone missed a place where the 1: should've been replaced with 4: | 23:18 |
Jjp137 | yup: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1029534 | 23:23 |
Wonka | oh my... I'm on a go today. | 23:29 |
Wonka | 0% [Working]E: Failed to fetch http://deb.devuan.org/merged/pool/DEBIAN/main/y/yq/yq_3.1.0-3_all.deb: 404 Not Found [IP: 131.188.12.211 80] | 23:29 |
Wonka | related: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1029509 | 23:30 |
bb|hcb | Wonka: That comes from deb.debian.org, retry with another mirror and it will work | 23:36 |
bb|hcb | Looks like lately fastly does not stand up to its name | 23:37 |
Wonka | wget http://deb.debian.org/debian/pool/main/y/yq/yq_3.1.0-3_all.deb worked... | 23:39 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!