IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 3 June 2021 (all times are UTC)
| 00:28 | eu^itninja has left IRC (eu^itninja!~eu^itninj@riddle.wizard.ca, Quit: Client closed) | |
| 00:56 | lucas_ has joined IRC (lucas_!~lucas@177-185-133-236.dynamic.isotelco.net.br) | |
| 00:56 | lucascastro has left IRC (lucascastro!~lucas@177-185-133-236.dynamic.isotelco.net.br, Remote host closed the connection) | |
| 04:22 | quinox has left IRC (quinox!~quinox@ghost.qtea.nl, Quit: WeeChat 3.0.1) | |
| 04:25 | quinox has joined IRC (quinox!~quinox@ghost.qtea.nl) | |
| 05:48 | shored has left IRC (shored!~shored@user/shored, Read error: Connection reset by peer) | |
| 05:48 | shored has joined IRC (shored!~shored@user/shored) | |
| 05:53 | ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz) | |
| 07:35 | ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Ping timeout: 245 seconds) | |
| 07:56 | ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz) | |
| 09:16 | hark_ has joined IRC (hark_!~hark@94.142.244.66) | |
| 09:20 | alkisg has left IRC (alkisg!~alkismatr@2001:470:69fc:105::2d3, Ping timeout: 272 seconds) | |
| 10:01 | ogra has left IRC (ogra!~ogra_@ubuntu/member/ogra, *.net *.split) | |
| 10:01 | jgee has left IRC (jgee!~jgee@186.80.49.20, *.net *.split) | |
| 10:11 | ogra has joined IRC (ogra!~ogra_@ubuntu/member/ogra) | |
| 10:11 | jgee has joined IRC (jgee!~jgee@186.80.49.20) | |
| 11:14 | alkisg has joined IRC (alkisg!~alkisg@2001:470:69fc:105::2d3) | |
| 12:34 | adrianorg has left IRC (adrianorg!~adrianorg@177.204.79.212.dynamic.adsl.gvt.net.br, Ping timeout: 264 seconds) | |
| 12:35 | adrianorg has joined IRC (adrianorg!~adrianorg@186.213.159.101) | |
| 13:24 | alkisg has left IRC (alkisg!~alkisg@2001:470:69fc:105::2d3, Quit: node-irc says goodbye) | |
| 13:26 | alkisg has joined IRC (alkisg!~alkisg@2001:470:69fc:105::2d3) | |
| 15:55 | vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100b) | |
| 21:39 | ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving) | |
| 22:29 | hark_ has left IRC (hark_!~hark@94.142.244.66, Ping timeout: 272 seconds) | |
| 22:29 | hark_ has joined IRC (hark_!~hark@94.142.244.66) | |
| 22:50 | itninja has joined IRC (itninja!~itninja@riddle.wizard.ca) | |
| 22:58 | <itninja> Hello! I still can't join the Matrix room, so I am posting my question here. Does the current LTSP allow for thin-client style of setup, where all a thin client runs is something like "ldm" that displays X/Gnome interface from the LTSP server as if they are on working on the LTSP server? (all applications are run on the LTSP server)
| |
| 23:03 | <vagrantc> itninja: not out of the box. you could implement something like that using RDP, vnc or x2go on top of LTSP
| |
| 23:04 | itninja: unfortunately, the methods LDM used are loosing support in various upstream projects; many projects do not support network-transparent X11 that LDM relied on for thin clients
| |
| 23:04 | or rather, they do not support it well; there have been increasing bugs over the last 10-15 years
| |
| 23:05 | it is not a well tested use-case, and many have switched to wayland, so that method suffers from lack of maintenance
| |
| 23:06 | there's nothing LTSP can do to fix the fundamental problem, as it is various applications such as firefox, libreoffice, etc. that have had issues
| |
| 23:08 | the x2go stuff is pretty actively maintained, though; i think it takes a bit more resources on the server-side, but might meet your needs
| |
| 23:08 | itninja: good luck!
| |
| 23:09 | <itninja> I see. so the current LTSP design is for working with "fat clients", and I have to ensure that the thin-computers are upgraded to have enough RAM and CPU power to run all those apps locally...
| |
| 23:09 | x2go... good to know. I will check it out!
| |
| 23:09 | <vagrantc> i think the x2go maintainer in Debian at least actually uses it with LTSP, if i recall correctly
| |
| 23:10 | <itninja> I see. thanks for the info!
| |
| 23:11 | btw, does anyone know why I can't "Join the conversation" in Matrix (#ltsp:matrix.org)?
| |
| 23:11 | *Join the discussion
| |
| 23:11 | <vagrantc> i was sad to see the transparent network stuff go too, but ... it's hard to fight against the industry-wide momentum moving away from it
| |
| 23:11 | <itninja> yeah... sad indeed.
| |
| 23:11 | <vagrantc> not familiar with matrix yet
| |
| 23:13 | <itninja> lol.. I just saw on the x2go web site...that they moved to Libera.chat and Matrix as well due to Freenode team change.. (didn't realize that Freenode team change was so major...)
| |
| 23:15 | re. "chrootless" and "chroot" set up in the current LTSP system... where can I find more explanation or use-case or documentation between the two? I am confused by how they are different..and from my generating the client-images for each based on the ltsp.org site's instructions, the results on the 'thin'-client side look the same to me....
| |
| 23:17 | <vagrantc> yeah, not sure if there are any legitimate projects left on freenode
| |
| 23:19 | itninja: chrootless builds the LTSP image from the host OS's files, a chroot set up would be if you install an operating system in a subdirectory and tell LTSP to build your image from that
| |
| 23:20 | from LTSP's perspective, they're pretty much the same ... "/" is used in chrootless mode, and "/some/other/directory" would be used in chroot mode.
| |
| 23:20 | where /some/other/directory contains the complete OS you want to build an LTSP image for
| |
| 23:25 | <itninja> ah I see. so the "chroot" just refers to how you build the client image. But in the client side, how it loads the image and the resulting environment are the same (provided that nothing is changed in the 'chroot' image since creation by ltsp)
| |
| 23:26 | <vagrantc> right
| |
| 23:27 | <itninja> Thank you very much for the clarification! I am enlightened! :)
| |
| 23:28 | <vagrantc> i don't have much of a hand in the new LTSP, but i've maintained LTSP in Debian for many years and at least kept an eye on alkisg with the rewrite :)
| |
| 23:32 | <itninja> I see! Did you use ldm-style transparent-X11-tunneling thin-client setup before and had to move to the new LTSP system with 'fat clients'? (I am worried that I might have to get several thin-client replaced with new more powerful boxes to upgrade from an old Ubuntu16 LTSP system to the current one..)
| |
| 23:33 | <vagrantc> i haven't used LTSP in the real world for many years, so I haven't really done such a transition
| |
| 23:34 | long before LTSP dropped support for thin clients, fat clients were pretty much the recommended way anyways
| |
| 23:35 | how much ram/cpu do your old clients have?
| |
| 23:37 | rough estimate, no need to get exact :)
| |
| 23:37 | <itninja> I see. they are most Zotac boxes.. 2GB~4GB RAM... dual-core Celerons
| |
| 23:37 | <vagrantc> might be ok
| |
| 23:38 | <itninja> oh? :D
| |
| 23:38 | <vagrantc> can you set up a test network and try a few of them on it?
| |
| 23:39 | obviously, more ram is better
| |
| 23:39 | there might be some options for network swap, too ... not sure what the current recommendation is
| |
| 23:40 | might even perform better
| |
| 23:40 | the overhead with LDM was pretty high on both the client and server
| |
| 23:41 | one nice thin with LTSP fat clients is it requires way fewer resources on the server
| |
| 23:42 | main thing i would guess is web browsers these days have really grown a lot ... but a few years ago 2-4GB ram was a solid indication to encourage switching to fat clients
| |
| 23:43 | but it will be better able to take advantage of the local GPU, and it won't require network traffic for every mouse movement and keystroke
| |
| 23:44 | <itninja> yes. I will be doing some testing. the current set up has a beefy LTSP server.. lol .. cause the idea then was FAT-server + thin-client structure.
| |
| 23:45 | <vagrantc> cut the server in half and serve multiple networks :)
| |
| 23:45 | <itninja> true. there will also be several changes required to switch to fat-client set up.. due how user environments are set up in the server and how certain resources and permission are all set up /controlled in the server...
| |
| 23:56 | hmm..one question: with the current LTSP set up with fat clients.. do the clients need a local hard drive?
| |
| 23:56 | <vagrantc> no
| |
| 23:57 | <itninja> ok. I assumed so. good!
| |
| 23:57 | <vagrantc> it wouldn't be LTSP if it did :)
| |
| 23:57 | that's the fundamental problem LTSP solves ... network booting your OS
| |