IRC chat logs for #ltsp on irc.libera.chat (webchat)


Channel log from 15 August 2019   (all times are UTC)

01:30mmarconm has joined IRC (mmarconm!~mmarconm@unaffiliated/mmarconm)
01:47mmarconm has left IRC (mmarconm!~mmarconm@unaffiliated/mmarconm, Quit: Leaving)
03:25adrianorg has joined IRC (adrianorg!~adrianorg@179.187.25.90.dynamic.adsl.gvt.net.br)
04:24vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
05:41os_a has joined IRC (os_a!~Thunderbi@195.112.116.22)
05:50ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
07:18kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr)
07:55kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr, Ping timeout: 248 seconds)
08:25kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr)
08:27statler has joined IRC (statler!~Georg@gwrz.lohn24.de)
10:37GodFather_ has left IRC (GodFather_!~rcc@d53-64-7-141.nap.wideopenwest.com, Ping timeout: 245 seconds)
10:38GodFather has left IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com, Ping timeout: 246 seconds)
11:28kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr, Quit: No Ping reply in 180 seconds.)
11:30kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr)
12:17os_a has left IRC (os_a!~Thunderbi@195.112.116.22, Remote host closed the connection)
12:20section1 has joined IRC (section1!~section1@178.33.109.106)
12:31vsuojanen has left IRC (vsuojanen!~vsuojanen@cable-hml-585686-205.dhcp.inet.fi, Ping timeout: 268 seconds)
12:43Filipe96 has joined IRC (Filipe96!c8118d03@aruana.ufs.br)
12:43
<Filipe96>
alkisg, you know any tutorial to installation ltsp for rapsberry clients?
12:50
someone know any tutorial for installation of ltsp for raspberry clients?
12:53gbaman has joined IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
12:53
<alkisg>
!raspberrypi
12:53
<ltsp>
raspberrypi: (#1) Ubuntu/LTSP on Pi 2: https://help.ubuntu.com/community/UbuntuLTSP/RaspberryPi, or (#2) Debian/LTSP (with raspbian chroot) on Pi: http://cascadia.debian.net/trenza/Documentation/raspberrypi-ltsp-howto/, or (#3) unofficial Ubuntu/LTSP (with raspbian chroot) on Pi: http://pinet.org.uk/
12:55
<gbaman>
Well, isn't that an oddly timed coincidence alkisg :)
12:56
One thing worth keeping in mind is don't think any of the options work with Pi 4 just yet though
12:56
(PiNet certainly doesn't... Still stuck up against what I am assuming are graphics issues with LDM instantly crashing)
13:05
<alkisg>
gbaman: I have the first ltsp19 packages in the ppa ready, you might want to rebase pinet on them
13:06
<gbaman>
ltsp19?
13:06
<alkisg>
Yes, the new ltsp, as the old one is not maintained anymore
13:06
!installation
13:06
<ltsp>
installation: To install the new LTSP, start there: https://github.com/ltsp/ltsp/wiki/installation
13:06
<gbaman>
Completely new LTSP that will break everything?...
13:06
<alkisg>
Yes, or fix everything, it depends on if you're optimist or not
13:06
<gbaman>
:)
13:07
I guess at this point, with so much breaking PiNet in Buster, not much to loose...
13:08
<alkisg>
gbaman: you're not using ubuntu, right? So you don't care if ubuntu 20.04 ships with ltsp5 or ltsp20, correct?
13:08
<gbaman>
On the client or server? PiNet recommends Ubuntu for the server side
13:09
(mainly given general ease of use of Ubuntu for none Linux experienced teachers)
13:09
<alkisg>
We can't have separate packages on ubuntu for ltsp server vs client
13:09
They'll either be both ltsp5,or ltsp19
13:09
(or ltsp20)
13:09
But note that noone will maintain ltsp5
13:09
So someone should step up and maintain it if he wants ltsp5 over ltsp19
13:10
<gbaman>
Well, if ltsp19 fixes what I have broken in front of me (without breaking too much else), then I am happy to migrate over
13:10
<alkisg>
I'm pretty sure ltsp19 will fix things BUT it will require integration work
13:10
As it doesn't care at all about backwards compatibility etc
13:11
<gbaman>
Is it at the point of being ready to actually use though?
13:12
<alkisg>
Now it's alpha; in a couple of weeks it'll be beta; in September it'll be used in schools
13:12
<gbaman>
Ohh, I see the timeline now
13:12
<alkisg>
Currently it can boot any .deb-based distribution after and including ubuntu 16.04 and debian jessie
13:12
<gbaman>
Is the LDM replacement bit done?
13:12
<alkisg>
Yes
13:13
It's all done, except for half of the lts.conf parameters
13:13
<gbaman>
Then that sounds like a lot of the stuff I need is there
13:13
Aww, shame to see ltsp-remoteapps go
13:13
<alkisg>
Most lts.conf parameters will be implemented over the next couple of weeks, and a few will be implemented later, or none at all
13:13
ltsp-remoteapps can be implemented with ssh keys and ssh -X
13:13
nothing important there
13:14
Although, that won't work with wayland, as there's no ssh -X there
13:14
<gbaman>
Haha yeah I know, is what I was doing for some of the PiNet pieces anyway
13:14
Raspbian dropped Wayland plans a while back
13:14
<alkisg>
I'm sure that means "postponed"
13:14
In 10 years when no xorg exists at all, I doubt they'll be the ones that maintain it
13:15
<gbaman>
Haha, I think long term postponed is best way to put it
13:15
<alkisg>
I'm guessing in 5 years many apps will depend on wayland and won't run on xorg
13:15
We'll see; in any case trying to support wayland is a good thing
13:15
<gbaman>
Yip
13:19gbaman has left IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com, Read error: Connection reset by peer)
13:19gbaman_ has joined IRC (gbaman_!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
13:20
<gbaman_>
So is login done via a custom LTSP ltsp application like LDM, or via pam module in this new version?
13:20
<alkisg>
PAM
13:21
So you'd be using the stock lightdm of rasbpian
13:21
<gbaman_>
Ahh cool
13:33
So am I right in saying there is no ltsp-client package now?
13:52Filipe96 has left IRC (Filipe96!c8118d03@aruana.ufs.br, Ping timeout: 260 seconds)
13:58
<alkisg>
gbaman_: the ltsp-client package is now named ltsp
13:58
<gbaman_>
So all merged into 1 package then
13:58
<alkisg>
There's no ltsp-server package; one would install: apt install ltsp dnsmasq nfs-kernel-server etc
13:58
Yes
13:58
And all interpreted, so it's sent from the server, it doesn't even need to be installed on the clients
13:58
I.e. you could directly boot the raspbian image without ever even installing it anywhere
13:59
<gbaman_>
So no additional packages are required on the Raspbian image?...
13:59
I was about to go install the ltsp package on it
14:00
<alkisg>
I helps if it's there
14:00
It save 1 MB RAM :P and it ensures that squashfs and overlayfs are there
14:00
But squashfs and overlayfs are usually there anyway, and 1 MB isn't important, so...
14:02
<gbaman_>
So what has changed around how I actually get into an LTSP boot mode from the Raspbian standpoint? In the past the entrance in was an LTSP specific init file that I was calling from the cmdline parameters, that now changed then?
14:03
<alkisg>
Yes, the fact that the ltsp initrd is sent, signals an ltsp boot mode
14:03
!ltsp-initrd
14:03
<ltsp>
Error: "ltsp-initrd" is not a valid command.
14:03
<gbaman_>
Ahh, so it is just hard coded in the initrd
14:03
<alkisg>
https://github.com/ltsp/ltsp/blob/master/man/ltsp-initrd.8.md
14:03
If you boot kernel, initrd,it's normal mode
14:04
if you boot kernel, initrd, ltsp.img, it's ltsp mode
14:04
It's not "hardcoded", it's just a cleaner way to do it
14:05
!learn ltsp-initrd as Creates the ltsp.img initrd add-on: https://github.com/ltsp/ltsp/blob/master/man/ltsp-initrd.8.md
14:05
<ltsp>
The operation succeeded.
14:09kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr, Quit: No Ping reply in 180 seconds.)
14:11kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr)
14:13
<gbaman_>
So alkisg ltsp-chroot is gone then?
14:14
(or an equivalent that is)
14:14
<alkisg>
Yes, chroots have problems, they shouldn't really be used with the "chroot" command
14:14
One can use VM, qemu, nfs-rw booted clients etc
14:14
But of course if one wants to maintain a chroot, then ltsp-image is happy to use it
14:15
<gbaman_>
What was wrong with the old approach? Worked great for PiNet
14:15
<alkisg>
We just don't provide tools to generate or maintain chroots; but we can use them if the user does maintain them
14:15
There are many packages that can't be installed in chroots, because they break, because they don't see systemd or dbus or sysvinit etc running
14:15
Or they leave mounts running, breaking even the host
14:16
<gbaman_>
So the solution is actually "booting" them up
14:16
<alkisg>
Right; or use lxc or whatever else is more stable than plain "chroot"
14:16
Unfortunately developers don't care a lot if their packages don't work in chroots
14:20kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-197.home.otenet.gr, Ping timeout: 246 seconds)
14:26
<gbaman_>
So, am I right in saying then the kernel/initrd are standard, with no extra stuff injected inside? Then the ltsp.img has all the extra bits in it?
14:26
<alkisg>
Yes, ...except...
14:27
Live CDs do have the squashfs module in the initramfs. So one can use the stock kernel/initrd, and append the ltsp.img
14:27
But if one has a VM without the ltsp package, its initramfs doesn't have the squashfs module
14:27
<gbaman_>
See, the bit I am missing I think is how to get it loading the ltsp.img?
14:28
<alkisg>
So, if he runs `ltsp image` to generate vm.img, then the stock kernel/initrd can't read it
14:28
In that case, one would install the ltsp package inside the vm, which would run `update-initramfs` on postinst, and put the squashfs module in the initrd
14:28
So the ltsp package inside the chroot/vm/etc is necessary only for when the squashfs module is missing
14:29
Are you using NFS or NBD with pinet?
14:29
<gbaman_>
Makes sense
14:29
By default, NBD with squashfs images, but also do support NFS as the share
14:29
(I generally debug with the NFS option so don't need to recompress each time)
14:30
<alkisg>
You can use a VM instead, but anyway,
14:30
by default raspbian doesn't have an initramsf
14:30
So you'd need the ltsp package in order to automatically generate one
14:30
<gbaman_>
Well, had numerous issues with trying to make Raspbian play ball inside a booted Qemu VM
14:31
<alkisg>
You can maintain it from a client, by booting the VM from the client in rw mode
14:31
<gbaman_>
Yeah, currently got a test set of boot files with an initramfs that has squashfs included :)
14:32
Along with the ltsp.img file sitting on the sd card as well
14:33
So I am guessing, all that I have left is the right kernel cmdline
14:33
<alkisg>
Eh, sure, although I'd put bootcode.bin in the card and have the others on the server
14:34
<gbaman_>
Think to make that work, that would take a little bit more tweaking? Unless you have their custom dhcp option in dnsmasq default?
14:34
Given it doesn't implement normal pxe
14:35
<alkisg>
dhcp-mac=set:rpi,b8:27:eb:*:*:*
14:35
pxe-service=tag:rpi,X86PC,"Raspberry Pi",unused
14:35
!ltsp-dnsmasq
14:35
<ltsp>
Error: "ltsp-dnsmasq" is not a valid command.
14:35
<gbaman_>
And actually, in saying all that, Pi 4 is broke anyway for network booting right now :)
14:35
<alkisg>
It should work with a card with bootcode.bin
14:35
But not without the card
14:35
<gbaman_>
My understanding was Gordon hadn't even written the bootcode.bin bit yet for Pi 4?
14:36
<alkisg>
!learn ltsp-dnsmasq as Configures dnsmasq for LTSP: https://github.com/ltsp/ltsp/blob/master/man/ltsp-dnsmasq.8.md
14:36
<ltsp>
The operation succeeded.
14:36
<gbaman_>
But, I do conveniently have a Pi 3B+ in front of me as well
14:36
<alkisg>
I don't know, last time I checked they were saying that it runs fine, and I tested it in my rpi2 at the time and it worked
14:36
<gbaman_>
Yeah, they released Pi 4 without support for network or USB booting of any sort
14:36
<alkisg>
But I only work on maintaining ltsp for raspberries when someone hires me to do so, so... it's been more than a year now
14:37
<gbaman_>
But difference is the bootrom this time is stored on an SPI flash chip, so it can be upgraded
14:37
So they have said they will get around to it at some point, just ran out of time to rush the Pi 4 out
14:38
<alkisg>
...as long as they manage to create a working bootcode.bin later on, all is fine
14:38
<gbaman_>
Haha, yeah it's coming
14:38
But ok, I will try out on a Pi 3B+ today then
14:38
Does mean I need to fix the kernel situation though first
14:39
Actually, maybe will stick with just it all on the card just to make sure it works first
14:39
<alkisg>
Sure, that's more certain to work
14:39
<gbaman_>
Yeah
14:39
<alkisg>
And it gives better error messages :)
14:39
I don't have automatic kernel updates implemented yet
14:39
<gbaman_>
Haha, you mean "actual error messages"
14:40
Well, PiNet previously handled that itself
14:40
Given don't think the LTSP version existed when I initially needed it
14:41
<alkisg>
I think I implemented that before pinet
14:41
<gbaman_>
I am a long way from any sort of public release though, so intention would be eventually to release with bootcode.bin mode only anyway
14:41
<alkisg>
To boot some old compex cards
14:41
<gbaman_>
So what cmdline parameters are you passing in when it boots over pxe?
14:42
<alkisg>
!ltsp-ipxe
14:42
<ltsp>
I do not know about 'ltsp-ipxe', but I do know about these similar topics: 'ltsp-pnp', 'ltsp-info', 'ltsp-pam', 'ltsp-site'
14:42
<alkisg>
Meh I don't have anything :P
14:42
<gbaman_>
I literally just found my own way there anyway at same time
14:42
<alkisg>
!learn ltsp-ipxe as Install iPXE binaries and configuration in TFTP: https://github.com/ltsp/ltsp/blob/master/man/ltsp-ipxe.8.md
14:42
<ltsp>
The operation succeeded.
14:43
<alkisg>
https://github.com/ltsp/ltsp/blob/master/ltsp/server/ipxe/ltsp.ipxe
14:43
The parameters are there ^
14:45
And from a live client: termbin.com/b2fu
14:45
You don't need max_part since you don't have ubuntu clients
14:52
<gbaman_>
Yeah, it doesn't like that given it wants an init=
14:52
(although that might be coming from the main initrd coming from an LTSP5 build)
14:53
<alkisg>
init=/sbin/init is assumed when init isn't present
14:53
It's from a running client, the syntax is fine
14:54
Are raspberries able to load 2 initrds?
14:54
<gbaman_>
See that is an important question
14:54
<alkisg>
Otherwise you need to concat them
14:55
<gbaman_>
Because I have never done it in the past
14:58
So the main way I have done it in the past is using the initramfs parameter in the config.txt instead of in the cmdline.txt file
14:58
https://www.raspberrypi.org/documentation/configuration/config-txt/boot.md
14:58
But I just gave it a go there with both in the config.txt file, funny feeling it is only running the 2nd one
15:03bengoa has joined IRC (bengoa!~alberto@194.50.55.200)
15:13vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
15:19
<alkisg>
gbaman_: that might be; start with concatenating the two initrds into one
15:24ZAJDAN has left IRC (ZAJDAN!~zdenek@77.48.149.75, Quit: Konversation terminated!)
15:24ZAJDAN has joined IRC (ZAJDAN!~zdenek@77.48.149.75)
16:22statler has left IRC (statler!~Georg@gwrz.lohn24.de, Ping timeout: 248 seconds)
16:47statler has joined IRC (statler!~Georg@gwrz.lohn24.de)
17:28pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
18:15pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Remote host closed the connection)
18:16pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
18:22shored has left IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi, Read error: Connection reset by peer)
18:22shored has joined IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi)
19:01shored has left IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi, Read error: Connection reset by peer)
19:01shored has joined IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi)
19:30kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-29.home.otenet.gr)
19:53section1 has left IRC (section1!~section1@178.33.109.106, Remote host closed the connection)
19:56statler has left IRC (statler!~Georg@gwrz.lohn24.de, Remote host closed the connection)
19:56pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Remote host closed the connection)
20:00pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
20:07pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Remote host closed the connection)
20:24kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-29.home.otenet.gr, Ping timeout: 244 seconds)
20:27pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
20:28kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-29.home.otenet.gr)
21:01pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Remote host closed the connection)
21:01pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
21:06pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Ping timeout: 268 seconds)
21:10pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
21:37ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
21:40pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Remote host closed the connection)
22:17kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-29.home.otenet.gr, Ping timeout: 272 seconds)
22:54gbaman_ has left IRC (gbaman_!~gbaman@host81-142-46-233.in-addr.btopenworld.com, Read error: Connection reset by peer)
23:30vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
23:41pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
23:45pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Ping timeout: 248 seconds)