IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 30 January 2020 (all times are UTC)
| 00:21 | Ark74 has joined IRC (Ark74!~Luis@177.238.145.140) | |
| 04:00 | <alkisg> !forget ltsp-initrd
| |
| 04:00 | <ltspbot> The operation succeeded.
| |
| 04:01 | <alkisg> !learn ltsp-initrd as Create the ltsp.img initrd add-on: https://ltsp.org/man/ltsp-initrd/
| |
| 04:01 | <ltspbot> The operation succeeded.
| |
| 04:01 | <vagrantc> heh :)
| |
| 04:03 | alkisg: i didn't test some of the newly fixed issues, but it didn't break anything for me as of 8aa2e8de8029d2db722254616f740620ca343c2b Restore live CD netbooting (#43)
| |
| 04:03 | <alkisg> Great
| |
| 04:03 | If you want, go ahead and release it; or wait for me to fix that partprobe thing today
| |
| 04:04 | <vagrantc> if you fix it today, i may as well wait till tomorrow to upload ... should have enough time
| |
| 04:04 | <alkisg> Let's try to make it 20.01 instead of 20.02 though, in case we find something grave and need an 20.02 release :)
| |
| 04:05 | I will fix it today
| |
| 04:06 | <vagrantc> still will fit the version scheme :)
| |
| 04:07 | ah, i did notice cups failing to load ina way that there might be some more elegant way to fix ... i presume it's disabled in ltsp somehow?
| |
| 04:07 | it spits out obnoxious failures at boot
| |
| 04:07 | <alkisg> Yes, to save RAM, as by default it's not used
| |
| 04:08 | It didn't for me... what were the failure messages?
| |
| 04:09 | I think that by default I just use `systemctl mask`, the same as MASK_SYSTEM_SERVICES...
| |
| 04:10 | * vagrantc boots a client to try and find what it says | |
| 04:12 | <vagrantc> Failed to start CUPS Scheduler ... and ... Failed to start CUPS Socket
| |
| 04:12 | <alkisg> Ah, then we probably just need to mask those too
| |
| 04:14 | <vagrantc> seems to be cups.path and cups.socket
| |
| 04:14 | but they're not .service files ... does MASK_SYSTEM_SERVICES work with that?
| |
| 04:15 | <alkisg> If you do the change, please also lower "re systemctl mask" to "rw systemctl mask"; I updated MASK_SYSTEM_SERVICES yesterday to "rw" but forgot to update cups...
| |
| 04:15 | Yes, it should work fine
| |
| 04:15 | client/init/55-various.sh
| |
| 04:17 | <vagrantc> yup, adding cups.path and cups.socket to MASK_SYSTEM_SERVICES made the noise go away
| |
| 04:18 | in config_cups_server ?
| |
| 04:22 | <alkisg> Yes, only when it's needed to disable cups
| |
| 04:22 | There's a logic there where if CUPS_SERVER is localhost, cups shouldn't be disabled etc
| |
| 04:23 | vagrantc: but also note that something else might be triggering cups.path etc in your setup, and disabling that something else instead might make more sense
| |
| 04:23 | * vagrantc shrugs | |
| 04:23 | * alkisg boots a buster vm to check... | |
| 04:23 | <vagrantc> this is on bullseye, so maybe doesn't even exist in buster
| |
| 04:24 | they both claim to be PartOf=cups.service
| |
| 04:25 | <alkisg> Do you see a red line on boot? Or does it show up on systemctl --status?
| |
| 04:26 | <vagrantc> alkisg: red line on boot ... "systemctl status" didn't show a problem ... systemctl status cups.path or cups.socket showed an issue
| |
| 04:26 | <alkisg> systemctl status cups.path => not found
| |
| 04:26 | I don't have a bullseye vm though
| |
| 04:27 | <vagrantc> do i need to run "ltsp initrd" every time i change lts.conf ?
| |
| 04:27 | hah!
| |
| 04:27 | ltsp.conf
| |
| 04:27 | <alkisg> Yeah
| |
| 04:27 | <vagrantc> ok
| |
| 04:28 | <alkisg> It's possible to fetch and apply ltsp.conf via nfs, but I haven't enabled it by default yet
| |
| 04:28 | <vagrantc> could also write a hook that re-runs it whenever ltsp.conf is modified
| |
| 04:28 | iwatch or whatever
| |
| 04:29 | <alkisg> `ltsp initrd` also sends all of /etc/ltsp and /usr/share/ltsp
| |
| 04:29 | So it will need some extra care if we use inotify etc to automate it
| |
| 04:30 | I guess that part might be implemented as an http service instead
| |
| 04:30 | wget server:ltsp.conf => also updates ltsp.img if needed on the server
| |
| 04:31 | <vagrantc> it's so nice that login on the console works now!
| |
| 04:32 | <alkisg> We should do a gsoc about implementing passwd changing with PAM too
| |
| 04:32 | I wonder if that can be done with pam_exec, or if .c will be needed
| |
| 04:33 | * alkisg tests with debian-live-testing-i386-mate.iso | |
| 04:33 | <alkisg> (booting .isos is cool too; otherwise it wouldn't be easy for me to test bullseye :))
| |
| 04:34 | <vagrantc> yeah, that's amazing :)
| |
| 04:35 | so far, i've tested sddm and lightdm and gdm3 ... although gdm3 is absurdly slow but works ... and lxdm doesn't work at all
| |
| 04:36 | well, lxdm comes up, but it doesn't successfully auth
| |
| 04:36 | <alkisg> LXDE has moved to lightdm a couple of years ago, right?
| |
| 04:36 | So I didn't write code to handle lxdm.conf at all
| |
| 04:36 | <vagrantc> could be
| |
| 04:37 | <alkisg> gdm3 is using wayland, and even when it fails and falls back to xorg, it takes a long time on VMs
| |
| 04:37 | <vagrantc> i figured the pam stuff should work, but apparently you need additional code?
| |
| 04:37 | <alkisg> It's a bit faster on real hardware
| |
| 04:37 | No, it should work, sorry I was talking about handling autologins etc
| |
| 04:37 | But DMs do have bugs, some don't work with ldap etc
| |
| 04:37 | Not all of them use pam properly
| |
| 04:38 | E.g. ubuntu unity 16.04 failed to unlock with the new ltsp. I googled it,and it had the same issue with ldap
| |
| 04:38 | Yet ubuntu mate 16.04 etc work fine, it was a unity bug
| |
| 04:38 | <vagrantc> heh
| |
| 04:39 | that said, supporting lightdm, sddm and gdm3 sounds to cover the most important ones ...
| |
| 04:39 | <alkisg> debian-live-testing-i386-mate.iso doesn't have cups.service nor cups.path...
| |
| 04:40 | focal does have cups.path, but it doesn't show up in red, just "dead inactive"
| |
| 04:41 | Sorry it does at boot, but not at systemctl status
| |
| 04:41 | <vagrantc> sysemctl --list-units, maybe
| |
| 04:41 | <alkisg> So I can reproduce locally
| |
| 04:41 | <vagrantc> yay
| |
| 04:41 | seems fairly minor
| |
| 04:41 | i've built a patched ltsp that also tries to disable those two ... but i guess it might go badly if those files aren't present
| |
| 04:42 | or, it will shift from one warning to another :)
| |
| 04:42 | <alkisg> mask isn't intrusive, it should show a "service not foudn but proceeding anyway", and just make a symlink to /dev/null
| |
| 04:42 | But not in red :)
| |
| 04:43 | gdm3 is a pain with epoptes, they refused to allow broadcasting over it as a valid use case, and it pops up one vt for itself and one for the user, so it's confusing
| |
| 04:43 | <vagrantc> ok, great
| |
| 04:43 | <alkisg> For now, I'll be recommending gnome users to use lightdm instead
| |
| 04:45 | <vagrantc> sounds reasonable
| |
| 04:48 | gdi2k has left IRC (gdi2k!~gdi2k@host86-185-148-176.range86-185.btcentralplus.com, Quit: Leaving) | |
| 04:49 | <alkisg> vagrantc: as far as I can see, yeah disabling cups.path and cups.socket when cups.service is disabled, is fine and won't cause other issues
| |
| 04:50 | I even have those in 18.04, they just didn't show up in red for some reason
| |
| 04:50 | If you want commit it and I'll test that too today along with partprobe
| |
| 04:50 | Or if you don't have time, I can do that too
| |
| 04:52 | <vagrantc> already have the patches locally, so will push :)
| |
| 04:52 | <alkisg> Great
| |
| 04:52 | <vagrantc> seemed to work for me
| |
| 04:53 | alkisg: any objection to moving /usr/sbin/ltsp to /usr/bin/ltsp ?
| |
| 04:53 | <alkisg> vagrantc: don't we put administrative programs in sbin?
| |
| 04:53 | <vagrantc> but ltsp info is useful without administrative privledges
| |
| 04:54 | <alkisg> Just for now :)
| |
| 04:54 | It should use ltsp mount in the future to get info from inside the images
| |
| 04:54 | <vagrantc> sure, but you could use archivemount for that :)
| |
| 04:54 | <alkisg> Or to access ltsp.conf
| |
| 04:55 | (which is not readable by users in the recommended setup)
| |
| 04:55 | We'd have to rewrite the ltsp code to use archivemount then, instead of reusing the existing mount code
| |
| 04:55 | <vagrantc> ah, i see
| |
| 04:55 | if it's really going to soon require root, i guess it's fine
| |
| 04:55 | <alkisg> The existing mount code runs under busybox too
| |
| 04:56 | <vagrantc> makes it harder to run ltsp --help but no big deal
| |
| 04:56 | <alkisg> But don't administrative programs belong in sbin anyway? All those do support --help
| |
| 04:56 | Is there any recommendation about moving things to bin that I'm not aware of?
| |
| 04:57 | <vagrantc> it's seemed to be a trend, but i'm not aware of any specific recommendations
| |
| 04:58 | * alkisg wouldn't even mind if sbin went away completely :) | |
| 04:58 | <vagrantc> i think that might happen
| |
| 04:58 | soemday
| |
| 04:58 | <alkisg> There are so many programs in bin that aren't useful without root rights or pkexec authorization anyway
| |
| 04:59 | vagrantc: so if you think that bin is more appropriate, I have no objections at all
| |
| 04:59 | Maybe we'll need to update the section of man pages, maybe not
| |
| 05:01 | <vagrantc> the thought came to me when i booted a vm that i *thought* had ltsp installed but then i ran "ltsp info" only to get command not found ...
| |
| 05:01 | <alkisg> Ah vagrantc one thing only... we're using things like `modprobe`; if ltsp isn't in sbin, and sbin isn't in path, those won't work
| |
| 05:01 | <vagrantc> true enough
| |
| 05:03 | alkisg: it's a pleasure to be working on ltsp again ... even if i still don't personally have a real-world use at the moment :)
| |
| 05:03 | <alkisg> It's a fun little beast; challenging but rewarding
| |
| 05:04 | OK off to work; will start coding in an hour
| |
| 05:04 | <vagrantc> the latest stuff seems so much easier to debug and work with and tweak :)
| |
| 05:05 | still don't know where everything is off the top of my head, but that might come with time :)
| |
| 05:05 | <alkisg> It was horrible when `ltsp-update-image` was needed to test a one-line changr
| |
| 05:05 | <vagrantc> yes!
| |
| 05:05 | the LTSP_INIT_* tweaks are astoundingly flexible
| |
| 05:06 | er ... POST_INIT_*
| |
| 05:06 | <alkisg> Yeah, PRE_* too
| |
| 05:06 | E.g. PRE_INITRD_BOTTOM => gets a shell inside the initramfs
| |
| 05:46 | What would be the "mid" directive? IN_*? E.g. IN_KERNEL_DEBUG_SHELL=debug_shell, to get a shell at the time when the image is mounted, not after the work is done...
| |
| 05:47 | (for the future of course, not now...)
| |
| 05:48 | kjackal has joined IRC (kjackal!~quassel@2a02:587:3115:3000:e5a1:8ff8:992d:2899) | |
| 06:01 | statler has joined IRC (statler!~Georg@p5B30E4C3.dip0.t-ipconnect.de) | |
| 06:15 | <alkisg> !learn unmkinitramfs as unmkinitramfs extracts an initrd, but it doesn't always work, so: kvm -m 512 -kernel /srv/ltsp/focal/boot/vmlinuz -initrd /srv/ltsp/focal/boot/initrd.img -append "break=top"
| |
| 06:15 | <ltspbot> The operation succeeded.
| |
| 06:16 | <vagrantc> can't you just zcat | cpio -i ?
| |
| 06:16 | <alkisg> Nah, those were the good old days
| |
| 06:16 | <vagrantc> hah
| |
| 06:16 | well, there are other compression types, sure
| |
| 06:16 | <alkisg> Now they put the microcode in front and it doesn't work anymore
| |
| 06:17 | Hence unmkinitramfs,
| |
| 06:17 | * vagrantc sighs | |
| 06:17 | <alkisg> But then Ubuntu does something more weird and breaks concatenated initrd traversal, hence... kvm
| |
| 06:18 | <vagrantc> eesh
| |
| 06:18 | <alkisg> I should file a bug for unmkinitramfs in ubuntu, but ... no... time... for all those bugs :D
| |
| 06:19 | <vagrantc> i tried making an x86_32 chroot ... and it failed until i figured out i needed to get the overlay module added
| |
| 06:19 | and the easiest way to do that was to add the ltsp package
| |
| 06:19 | <alkisg> Hehe
| |
| 06:20 | O
| |
| 06:20 | <vagrantc> but mmdebstrap works really nicely to make a chroot ... much faster than debootstrap
| |
| 06:20 | <alkisg> I'll rewrite the systemctl mask cups in a single line, no need to do 3 invocations there
| |
| 06:21 | * alkisg would love to see cross-distro chroot creation tools | |
| 06:21 | <vagrantc> it can even make a squashfs image ... can ltsp image read in a squashfs?
| |
| 06:21 | <alkisg> Sure
| |
| 06:21 | It can even read a squashfs from inside a partitioned image
| |
| 06:22 | mmdebstrap available since buster, not bad for recommending it in https://github.com/ltsp/community/wiki/chroots
| |
| 06:24 | Heh, it even has a chrootless mode; I guess that's a word then :P
| |
| 06:24 | (when we changed ltsp-pnp to chrootless ltsp)
| |
| 06:25 | I wonder if it can create an armhf chroot that way
| |
| 06:26 | kjackal has left IRC (kjackal!~quassel@2a02:587:3115:3000:e5a1:8ff8:992d:2899, Ping timeout: 272 seconds) | |
| 06:28 | <vagrantc> anything well supported by qemu-user-static should work fine
| |
| 06:32 | <alkisg> It says chrootless, I imagine it doesn't even use the chroot command at all
| |
| 06:33 | So no qemu either
| |
| 06:33 | So it should work even in architectures that qemu doesn't support, or even to create a x86 chroot while the server is armhf, etc etc
| |
| 06:41 | <vagrantc> so, most of those modes are quirky
| |
| 06:41 | i've only had any luck with sudo/root and unshare modes
| |
| 06:43 | it would be interesting to try it, though :)
| |
| 06:44 | <alkisg> There's a trend in fedora to avoid postinst in favor of declarative things
| |
| 06:44 | E.g. to drop a file with the group you want, instead of calling adduser
| |
| 06:44 | That would be awesome, if packages were able to get installed without running any code at postinst
| |
| 06:44 | <vagrantc> yeah, i've started seeing that in debian too
| |
| 06:45 | <alkisg> Then we could manage chroots without the chroot command, with just an apt --root=there switch
| |
| 06:45 | <vagrantc> yes, debian's maintainer scripts are a big strike against debian (and derivatives)
| |
| 06:45 | that's basically what mmdebstrap's chrootless attempts to do...
| |
| 06:47 | <alkisg> OK I'll assume that `losetup -P` is everywhere available *except* for Ubuntu's initramfs'es
| |
| 06:47 | It's provided by the mount package, and it's there in debian initramfs'es
| |
| 06:48 | (about the partprobe thing)
| |
| 06:48 | So it's even there on a `deboostrap`'ed chroot with no other packages
| |
| 07:02 | This is the diff for partprobe, testing it: https://termbin.com/jkir
| |
| 07:02 | I.e. we try to pass as much of "-rP" as we can to losetup
| |
| 07:03 | And we expect it to work in all cases, even on initramfs/klibc which doesn't support either of them
| |
| 07:04 | <vagrantc> so ... with mmdebstrap and squashfs-tools-ng ... managed to create a bootable ltsp image:
| |
| 07:04 | sudo mmdebstrap --aptopt='Acquire::http::Proxy "http://192.168.122.1:8000";' --aptopt='Apt::Install-Recommends "true"' --aptopt='Acquire::Languages "none"' --include=ltsp,linux-image-686,systemd-sysv,lxde --arch=i386 bullseye /srv/ltsp/x86_32.squashfs http://deb.debian.org/debian
| |
| 07:05 | the proxy is obviously optional
| |
| 07:05 | <alkisg> Yey! Wiki time! :)
| |
| 07:05 | <vagrantc> and the acquire::languages is optional as well ... but it skips downloading some needless cruft
| |
| 07:05 | though package searches won't have full descriptions from apt
| |
| 07:06 | oh, then i had to ln -s /srv/ltsp/x86_32.squashfs /srv/ltsp/images/x86_32
| |
| 07:06 | <alkisg> VMs go to /srv/ltsp/vm.img, exported images to /srv/ltsp/images/ex.img
| |
| 07:07 | VMs can be symlinked, images shouldn't due to nfs
| |
| 07:07 | (although our nfs root is /srv/ltsp, so it'll work under that)
| |
| 07:07 | <vagrantc> i guess i could've ln -s x86_32.squashfs x86_32.img and then ltsp image x86_32 ?
| |
| 07:08 | <alkisg> Why do you want to create a squashfs from the chroot instead of letting ltsp do it?
| |
| 07:08 | Put the chroot in /srv/ltsp/chroot and run ltsp image chroot
| |
| 07:08 | <vagrantc> there's no chroot
| |
| 07:08 | <alkisg> mmdebstrap creates a squashfs without a directory?
| |
| 07:08 | <vagrantc> it produces a usable squashfs image
| |
| 07:08 | <alkisg> But only if you pass some option, right?
| |
| 07:09 | Can't you just tell it to create a chroot instead?
| |
| 07:09 | i.e. omit that option?
| |
| 07:09 | <vagrantc> yes, it can also create a chroot ... just wanted to see if it woudl work this way
| |
| 07:09 | <alkisg> In that case, you'll put it directly inside images/
| |
| 07:09 | As the target dir, with no symlinks etc
| |
| 07:09 | <vagrantc> mmdebstrap assumes it will be a squashfs if the target ends in .squashfs
| |
| 07:09 | <alkisg> Ah, no option to make this .img?
| |
| 07:10 | <vagrantc> would have to rename it
| |
| 07:10 | <alkisg> Renaming sounds better than ln in this case
| |
| 07:10 | <vagrantc> sure
| |
| 07:10 | but it's pretty cool
| |
| 07:10 | it does create a chroot directory but then throws it out
| |
| 07:11 | the advantage being, no intermediary uncompressed directory ... the disadvantage being you can't update it as easily ... but it's pretty fast
| |
| 07:12 | mmdebstrap produces a tarball on standard output if you don't give it a target ... which then uses a tar2sqfs to generate the squashfs image
| |
| 07:12 | so i guess i could do the tar2sqfs part myself ... but was just experimenting with the options and ideas
| |
| 07:13 | it'd be nice to be able to add the initramfs-tools module hooks without installing all of ltsp ... not that it's very large :)
| |
| 07:14 | <alkisg> I think all the ltsp dependencies are already there in a plain deboostrap
| |
| 07:14 | So it just adds a few kb of code under /usr, nothing more
| |
| 07:14 | <vagrantc> i was more thinking about if you have an old version of ltsp in the booted image, and it has some leftover files that the initramfs copy of the ltsp files doesn't delete
| |
| 07:15 | <alkisg> The new and old ltsp can coexist without issues, including the initramfs code
| |
| 07:15 | <vagrantc> or does it not use /usr/share/ltsp ... just from the initramfs directly?
| |
| 07:15 | cool.
| |
| 07:15 | <alkisg> Some tricks were needed, but it works fine
| |
| 07:16 | <vagrantc> with --mode=unshare, you can create the ltsp image without root access ... though it requires enabling user namespaces on Debian (not sure if ubuntu has it enabled by default or not)
| |
| 07:21 | kjackal has joined IRC (kjackal!~quassel@athedsl-299999.home.otenet.gr) | |
| 07:24 | woernie has joined IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de) | |
| 07:49 | <alkisg> vagrantc: I'm ready, you may release now if you like
| |
| 07:50 | <vagrantc> alkisg: not until i get a good night's sleep :)
| |
| 07:50 | <alkisg> Sure :)
| |
| 07:54 | * vagrantc waves | |
| 07:55 | <alkisg> gn vagrantc
| |
| 07:55 | vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving) | |
| 09:24 | kjackal has left IRC (kjackal!~quassel@athedsl-299999.home.otenet.gr, Ping timeout: 268 seconds) | |
| 09:24 | kjackal has joined IRC (kjackal!~quassel@195.97.13.252) | |
| 09:46 | statler has left IRC (statler!~Georg@p5B30E4C3.dip0.t-ipconnect.de, Remote host closed the connection) | |
| 10:07 | kjackal has left IRC (kjackal!~quassel@195.97.13.252, Ping timeout: 268 seconds) | |
| 10:21 | statler has joined IRC (statler!~Georg@gwrz.lohn24.de) | |
| 10:38 | kjackal has joined IRC (kjackal!~quassel@2a02:587:3115:3000:e5a1:8ff8:992d:2899) | |
| 11:26 | woernie has left IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de, Remote host closed the connection) | |
| 11:27 | woernie has joined IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de) | |
| 11:31 | section1 has joined IRC (section1!~section1@178.33.109.106) | |
| 11:46 | enaut[m] has left IRC (enaut[m]!enauttchnc@gateway/shell/matrix.org/x-iwrgghksxacdlaka, Quit: User has been idle for 30+ days.) | |
| 12:35 | Faith has joined IRC (Faith!~Paty_@unaffiliated/faith) | |
| 15:38 | statler has left IRC (statler!~Georg@gwrz.lohn24.de, Ping timeout: 260 seconds) | |
| 15:51 | statler has joined IRC (statler!~Georg@gwrz3.lohn24.de) | |
| 16:17 | GodFather has joined IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net) | |
| 17:28 | kjackal has left IRC (kjackal!~quassel@2a02:587:3115:3000:e5a1:8ff8:992d:2899, Ping timeout: 272 seconds) | |
| 17:29 | mnevans has joined IRC (mnevans!8008b09d@zero.geol.umd.edu) | |
| 17:30 | <mnevans> Hell LTSP World. I would like to set up internet passthrough for a laptop computer that is on my LTSP LAN. 2 network interfaces, chrootless install.
| |
| 17:30 | Following instructions here, adapted for LTSP19:
| |
| 17:30 | https://help.ubuntu.com/community/UbuntuLTSP/ThinClientHowtoNAT
| |
| 17:31 | so I was thinking to add the line, to /etc/dnsmasq.d/local-dhcp.conf, such as:
| |
| 17:31 | # set router for setting up NAT? dhcp-option=option:routers,192.168.67.254
| |
| 17:32 | But this is the wrong syntax, as reported when I restart dnsmasq.
| |
| 17:32 | Any suggestions - appreciated. I will check back in 28 mins... thanks in advance!
| |
| 17:32 | <alkisg> mnevans: please don't follow ltsp5 instructions on ltsp20 :)
| |
| 17:32 | Read the man page and set NAT=1, nothing more
| |
| 17:32 | !ltsp.conf
| |
| 17:32 | <ltspbot> ltsp.conf: Configuration file for LTSP: https://ltsp.org/man/ltsp.conf/
| |
| 17:32 | <mnevans> Haha! Of course.
| |
| 17:33 | Sorry, I missed that. Which man page, exactly, please?
| |
| 17:33 | <alkisg> Read 2 lines above
| |
| 17:33 | <mnevans> got it...
| |
| 17:33 | man ltsp.conf
| |
| 17:33 | <alkisg> Or the link I pasted, online
| |
| 17:34 | Read the NAT parameter there
| |
| 17:34 | <mnevans> Is there a template ltsp.conf? I don't see one in /etc/ltsp/ in my installation.
| |
| 17:47 | mnevans2 has joined IRC (mnevans2!8102b419@129-2-180-25.wireless.umd.edu) | |
| 17:55 | <alkisg> mnevans, again read the man page
| |
| 17:55 | The first line there tells you how to install the template
| |
| 17:55 | Really, man pages are the best source of information
| |
| 17:55 | mnevans2 ^
| |
| 17:55 | <mnevans2> Got it ... I think the lines in ltsp.conf:
| |
| 17:55 | [server]
| |
| 17:55 | NAT=1
| |
| 17:55 | and restarted dnsmasq seemed not to complain. I can test in 5 mins... Then I will figure out and add the settings for the client displays, based on the examples in the man pages.
| |
| 17:55 | ...Thank you!
| |
| 17:56 | <alkisg> mnevans2, you need to restart the ltsp service, NOT dnsmasq, for NAT=1
| |
| 17:56 | sudo service ltsp restart
| |
| 17:56 | dnsmasq is not related to NAT
| |
| 17:56 | NAT is done by an ltsp script
| |
| 17:56 | <mnevans2> thank you - reading that... and also ltsp initrd?
| |
| 17:56 | <alkisg> [server] doesn't need ltsp initrd
| |
| 17:56 | Only clients do
| |
| 17:57 | vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc) | |
| 17:57 | <mnevans2> Ah, I missed that too. So you run ltsp initrd on each client?
| |
| 17:57 | to update them to the server installation?
| |
| 18:04 | * vagrantc would guess not, unless there's some unusual context | |
| 18:04 | mnevans2 has left IRC (mnevans2!8102b419@129-2-180-25.wireless.umd.edu, Ping timeout: 260 seconds) | |
| 18:05 | <vagrantc> alkisg: going for a release soonish :)
| |
| 18:11 | <alkisg> mnevans: no, i meant "ltsp initrd is only needed if you make changes in ltsp.conf under [clients] sections, not under [server] section"
| |
| 18:11 | vagrantc: great, it should be good to go
| |
| 18:12 | <vagrantc> i must admit it's a bit of a learning curve knowing what has to be run when
| |
| 18:13 | i don't know if there's some way to "check" so the end-user doesn't have to track which commands to run
| |
| 18:14 | ltsp do-the-things-and-don't-bother-me :)
| |
| 18:15 | alkisg: i struggled to get a working ubuntu 19.10 .iso to work ... but wasn't sure what the steps were ... remember doing it in the past
| |
| 18:15 | <alkisg> You should be able to just put it to images/ubuntu.img and run ltsp kernel ubuntu
| |
| 18:16 | <vagrantc> it was missing sshfs, i think ... there's somewhere to drop an sshfs binary
| |
| 18:17 | i also built an image with mmdebstrap, and it seemed to successfully authenticate (and login from the console worked), but gdm didn't start
| |
| 18:18 | er... gdm started ... but it failed to start a session for the user
| |
| 18:18 | who knows what i was missing :)
| |
| 18:20 | same process worked with debian bullseye, though
| |
| 18:33 | <alkisg> if you put sshfs in /etc/ltsp/ssh-arch, or if you use nfs
| |
| 18:34 | (or you create a local home with POST_INIT_somthing
| |
| 18:34 | <vagrantc> very clever :)
| |
| 18:49 | tagged, pushed ...
| |
| 18:49 | and uploaded to Debian
| |
| 18:49 | should be in the archive in 6 or so hours
| |
| 18:50 | alkisg: and then you file a sync request for ubuntu? or do you have to wait till it lands in debian testing?
| |
| 18:59 | section1 has left IRC (section1!~section1@178.33.109.106, Quit: Leaving) | |
| 19:00 | kjackal has joined IRC (kjackal!~quassel@2a02:587:3115:3000:e5a1:8ff8:992d:2899) | |
| 19:00 | <alkisg> vagrantc: nice! It should sync from unstable in a couple of days
| |
| 19:02 | statler has left IRC (statler!~Georg@gwrz3.lohn24.de, Ping timeout: 240 seconds) | |
| 19:06 | vagrantc_ has joined IRC (vagrantc_!~vagrant@unaffiliated/vagrantc) | |
| 19:14 | statler has joined IRC (statler!~Georg@gwrz.lohn24.de) | |
| 19:18 | statler has left IRC (statler!~Georg@gwrz.lohn24.de, Remote host closed the connection) | |
| 19:52 | <vagrantc_> alkisg: the release announcement seems to be missing some changelog entries ... or i'm not reading carefully
| |
| 20:01 | <alkisg> I left out 2 of mine that were regression fixes
| |
| 20:01 | so they shouldn't really be there
| |
| 20:02 | <vagrantc_> got it
| |
| 20:03 | i think adding "Gbp-Dch: Ignore" would make them not appear in debian/changelog when running "gbp dch" ... but i don't usually bother, and just manually edit it anyways.
| |
| 20:05 | * alkisg hears about gbp for the first time | |
| 20:05 | <vagrantc_> it's in git-buildpackage ...
| |
| 20:05 | <alkisg> I see
| |
| 20:06 | Is that what made two sections for each of us?
| |
| 20:06 | <vagrantc_> it's a full suite of stuff to generate debian packages from git, but i typically only use "dbp dch"
| |
| 20:07 | alkisg: yes. which isn't wrong, per se, as sometimes commits depend on previous commits and thus chronological makes more sense
| |
| 21:05 | mnevans has left IRC (mnevans!8008b09d@zero.geol.umd.edu, Remote host closed the connection) | |
| 21:26 | woernie has left IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de, Remote host closed the connection) | |
| 21:40 | vagrantc_ has left IRC (vagrantc_!~vagrant@unaffiliated/vagrantc, Quit: leaving) | |
| 21:48 | kjackal has left IRC (kjackal!~quassel@2a02:587:3115:3000:e5a1:8ff8:992d:2899, Ping timeout: 246 seconds) | |
| 22:16 | Faith has left IRC (Faith!~Paty_@unaffiliated/faith, Quit: Leaving) | |