00:02 | Faithful has quit IRC | |
00:04 | Faithful has joined #ltsp | |
01:01 | cyberorg has quit IRC | |
01:01 | cyberorg has joined #ltsp | |
01:09 | pembo13 has joined #ltsp | |
01:10 | <pembo13> I tried out K12Linux, worked very well
| |
01:10 | i ran into the multi network card issue however
| |
01:10 | ie. connecting to a secondary NIC and having the initrd later fail
| |
01:11 | in some cases I simply couldn't PXE at all, in other cases initrd would fail
| |
01:11 | is the solution to just use the primary card?
| |
01:16 | alkisg has quit IRC | |
01:30 | Faithful has quit IRC | |
01:43 | alkisg has joined #ltsp | |
01:51 | shrek_ has joined #ltsp | |
02:09 | shrek has quit IRC | |
02:21 | pembo13 has quit IRC | |
02:31 | shrek has joined #ltsp | |
02:36 | Faithful has joined #ltsp | |
02:39 | alkisg has quit IRC | |
03:16 | tjikkun_work has joined #ltsp | |
03:27 | ScorpKing has joined #ltsp | |
03:30 | craver has joined #ltsp | |
03:34 | try2free has quit IRC | |
03:34 | nicros_ has quit IRC | |
03:55 | dirigeant has joined #ltsp | |
04:05 | shrek_ has joined #ltsp | |
04:06 | shrek_ has quit IRC | |
04:21 | bobby_C has joined #ltsp | |
04:23 | shrek has quit IRC | |
04:24 | gfarbal2 has joined #ltsp | |
04:25 | ScorpKing has left #ltsp | |
04:33 | bobby_C has quit IRC | |
04:42 | gfarbal2 has left #ltsp | |
04:43 | alkisg has joined #ltsp | |
04:49 | bobby_C has joined #ltsp | |
05:05 | alkisg has quit IRC | |
05:31 | shrek has joined #ltsp | |
05:51 | EXP__ has joined #ltsp | |
05:51 | <EXP__> has anyone tried to play dvd on client
| |
05:52 | i can getit work, but I can play obly single files. Menu does not work
| |
06:01 | greenmang0 has joined #ltsp | |
06:01 | <greenmang0> cyberorg:
| |
06:03 | EXP__ has quit IRC | |
06:03 | <greenmang0> what is the maximum number of client supported by LTSP?
| |
06:03 | <ogra> there is none
| |
06:04 | though there is a limit of sessions you can run per amount of ram on a server
| |
06:04 | <greenmang0> ogra: can I know what's that limit?
| |
06:05 | <ogra> for a gnome or KDE session you can count 128M x running sessions plus about 256M to have the server itself running
| |
06:05 | <greenmang0> basically i want to decide the configuration of ltsp server
| |
06:05 | <ogra> i.e. 10 sessions would mean 1280M+256M
| |
06:06 | <greenmang0> ogra: what about openbox ? :D
| |
06:06 | <ogra> if you pick smaller desktops you indeed get a smaller footprint ...
| |
06:06 | no idea, i havent measured how much a single session eats ... install htop and run a single session ;)
| |
06:07 | <greenmang0> ogra: ok
| |
06:07 | <ogra> and sum up how the ram used for processes the user claims, that should get you an idea
| |
06:18 | pmatulis has joined #ltsp | |
06:30 | dipeshmehta has joined #ltsp | |
06:31 | <dipeshmehta> hello all, at my PXE client, I am able to LDM login screen, when I issue username/passwd, it gets authenticated (I checked the auth.log) but then the login screen re-appears... there is error message like "unable to start X session --- no "/home/user1/.xsession"" in user's home folder, can any body help me please?
| |
06:32 | <ogra> do you have a desktop installed ?
| |
06:37 | dirigeant has quit IRC | |
06:40 | <dipeshmehta> orga:I am installing LTSP on 8.04 server, followed https://help.ubuntu.com/community/ThinClientHowto
| |
06:40 | <ogra> did you read the top note on that page ?
| |
06:40 | please dont use it
| |
06:41 | and you cant install on a server, you need a desktop
| |
06:41 | i mean, you can install on a server, but you will still need a desktop to log in to
| |
06:42 | <dipeshmehta> how do I have a desktop on my server?
| |
06:42 | <ogra> use: https://help.ubuntu.com/community/UbuntuLTSP/LTSPQuickInstall
| |
06:44 | <dipeshmehta> thanks for link, but it also mentioned to 1. apt-get install ltsp-server, and 2. ltsp-build-client, I have done the same.... however, I do not have desktop installed on my server, it is CLI only
| |
06:49 | <ogra> you can install ubuntu-desktop on your server, but the results might differ from using a real desktop CD for installation
| |
06:52 | <dipeshmehta> in fact, I do not want to use desktop on my server, but as you said I would be needing to have desktop installed to run LTS, I want to install it... I just want thinclients to work, without disturbing my server's actual jobs...
| |
06:57 | dirigeant has joined #ltsp | |
07:09 | Nubae has joined #ltsp | |
07:14 | Egyptian[Home] has quit IRC | |
07:19 | The_Code has joined #ltsp | |
07:20 | dirigeant has quit IRC | |
07:22 | cliebow has joined #ltsp | |
07:22 | craver has quit IRC | |
07:24 | CAN-o-SPAM has joined #ltsp | |
07:25 | dirigeant has joined #ltsp | |
07:28 | roblimo has quit IRC | |
07:34 | dipeshmehta has left #ltsp | |
07:35 | Egyptian[Home] has joined #ltsp | |
07:58 | ScorpKing has joined #ltsp | |
08:00 | ScorpKing has quit IRC | |
08:01 | ScorpKing has joined #ltsp | |
08:04 | Nubae1 has joined #ltsp | |
08:05 | Nubae has quit IRC | |
08:09 | ScorpKing has quit IRC | |
08:09 | pmatulis has quit IRC | |
08:10 | ScorpKing has joined #ltsp | |
08:10 | pmatulis has joined #ltsp | |
08:14 | dirigeant has quit IRC | |
08:14 | dirigeant has joined #ltsp | |
08:18 | bobby_C has quit IRC | |
08:18 | Gadi has joined #ltsp | |
08:37 | jammcq has joined #ltsp | |
08:37 | <jammcq> good morning everybody
| |
08:37 | shrek has quit IRC | |
08:46 | shogunx has quit IRC | |
08:47 | <stgraber> morning jammcq
| |
08:50 | <jammcq> hey stgraber
| |
08:50 | Faithful1 has joined #ltsp | |
08:51 | alkisg has joined #ltsp | |
08:59 | shogunx has joined #ltsp | |
09:02 | Syntony has joined #ltsp | |
09:02 | <Syntony> hello....only english speaking here`?
| |
09:03 | perhaps someone can help me...
| |
09:04 | how can i manage userrights for thin clients on my ltsp-server?
| |
09:05 | <ScorpKing> Syntony: what kind of userights?
| |
09:05 | userrights*
| |
09:06 | <Syntony> hm?...for the terminal-user
| |
09:08 | <ScorpKing> Syntony: restrict direcoties? restrict site?
| |
09:08 | programs?
| |
09:08 | directories*
| |
09:08 | * ScorpKing can't spell today.. | |
09:08 | <Syntony> aaah okay...for programms and directories
| |
09:08 | sorry i missunderstood :)
| |
09:08 | <ScorpKing> np
| |
09:08 | <Syntony> my english is not so good :D
| |
09:09 | <ScorpKing> can you explain in detail what you want to do?
| |
09:09 | Syntony: neither are myne ;)
| |
09:09 | <Syntony> i can try :)
| |
09:10 | greenmang0 has quit IRC | |
09:11 | <Syntony> alright...the users, that r using my thin-clients should not allowed to do everything...okay?...they should only surfing, writing some documents and so on...
| |
09:12 | but they should not allowed to change something on the system...
| |
09:12 | alkisg has quit IRC | |
09:13 | <Syntony> i hope u know what i mean :D
| |
09:13 | <ScorpKing> Syntony: you can restrict users in derectories with filepermissions.
| |
09:13 | <Syntony> aha...okay...on the server-site?
| |
09:13 | <ScorpKing> Syntony: you are using it for an internet kiosk right?
| |
09:13 | yes
| |
09:14 | <Syntony> no its an study-project
| |
09:14 | <ScorpKing> ah
| |
09:14 | <Syntony> for a cluster-room
| |
09:14 | <ScorpKing> hmm..
| |
09:14 | <Syntony> 15 thin clients
| |
09:14 | 1 server
| |
09:15 | <ScorpKing> you only need to fiddle with the server
| |
09:15 | i'm not sure if it is possible to chroot the clients. i've never done it myself
| |
09:16 | <Syntony> yeah...hmm....but the clients must be manage
| |
09:16 | nicoAMG has joined #ltsp | |
09:16 | <ScorpKing> users will not be able to change the filesystem unless they are in the admin group or root
| |
09:16 | <Syntony> ok
| |
09:16 | hmm....
| |
09:16 | <ScorpKing> Syntony: you can use lts.conf file for special client config's
| |
09:17 | configurations*
| |
09:17 | <Syntony> the ltsp.conf or lts.conf?
| |
09:17 | <ScorpKing> lts.cont in the tftboot directory
| |
09:18 | <Syntony> hehe u have some fast fingers :D
| |
09:18 | <ScorpKing> not really. haha
| |
09:18 | <Syntony> :)
| |
09:19 | also...i can edit the restrict rights for the users in the lts.conf in the tftboot dircetory...right?
| |
09:19 | <ScorpKing> i have to go to gym now but i'm sure if you hang around here someone else might add something usefull or give you a better option
| |
09:20 | <Syntony> hehe okay...thanks alot :)...
| |
09:20 | have fun :)
| |
09:20 | <ScorpKing> Syntony: not restrictions. look on the ltsp website documentation for all the things you can put in lts.conf
| |
09:20 | cheers
| |
09:20 | ScorpKing has left #ltsp | |
09:35 | rjune has joined #ltsp | |
09:43 | evilx_ has quit IRC | |
09:56 | Syntony has quit IRC | |
09:59 | mikkel has joined #ltsp | |
10:06 | bobby_C has joined #ltsp | |
10:08 | Faithful1 has quit IRC | |
10:22 | Ryan52 has quit IRC | |
10:22 | F-GT has quit IRC | |
10:23 | mikkel has quit IRC | |
10:23 | Ryan52 has joined #ltsp | |
10:24 | F-GT has joined #ltsp | |
10:26 | dirigeant has quit IRC | |
10:27 | dirigeant has joined #ltsp | |
10:39 | Comete has joined #ltsp | |
10:39 | <Comete> hi
| |
10:41 | mikkel has joined #ltsp | |
10:41 | hanthana has joined #ltsp | |
10:43 | <Comete> i've set up a small network of 16 thinclients and a server with Edubuntu 8.10 but i can't boot correctly more than 3 thinclients, all the others hangs during the boot process. It's well explain in this post:http://www.mail-archive.com/ltsp-discuss@lists.sourceforge.net/msg35552.html
| |
10:44 | my problem is exactly the same
| |
10:45 | as this guy did, i've tested the network components and the server memory, all is ok.
| |
10:45 | i even replaced the NIC on the server
| |
10:47 | i can boot without problem each thinclient separately
| |
10:49 | <NeonLicht> so, what's your problem?
| |
10:52 | <Comete> NeonLicht: network connections stop working after more than 3 thinclients were started
| |
10:53 | <NeonLicht> any of them? I mean, you can start any 3 and they work and the 4th doesn't?
| |
10:55 | <Gadi> Comete: if you enabled remote logging, turn it off
| |
10:55 | <Comete> yes
| |
10:55 | Gadi: remote login ? what's this ?
| |
10:55 | <Gadi> ie: /etc/default/syslogd (or whatever its called)
| |
10:55 | adding "-r" to syslog
| |
10:56 | to have it accept logging from the terminals
| |
10:56 | that floods the network with traffic
| |
10:56 | and can interfere with everything
| |
10:56 | some people turn this on for debugging and then forget to turn it off
| |
11:00 | <Comete> Gadi: no, no remote logging
| |
11:12 | Gadi: you give me an idea... :) that was LDM_SYSLOG=true in lts.conf
| |
11:12 | Gadi: thank a lot !
| |
11:13 | <NeonLicht> do more than 3 terminals come up now?
| |
11:14 | <Comete> yes almost all of them
| |
11:14 | but it's still slow
| |
11:14 | <NeonLicht> excellent, so remote logging was the problem, then
| |
11:16 | <Comete> some of them still hangs
| |
11:17 | etyack has joined #ltsp | |
11:17 | <NeonLicht> have you tried not to boot all of them at once?
| |
11:18 | <etyack> Is is possible to disable Select Session and Select Language in LDM? Running ubuntu 8.10, ldm 2.0.14
| |
11:18 | <NeonLicht> when I run 20 clients out of one server (4.x times), sometimes a few of the clients couldn't finish booting when all the students on the classroom fired up the boxes at once
| |
11:19 | <Comete> NeonLicht: i will try it tomorrow, i must leave now... thanks !
| |
11:19 | <NeonLicht> good luck
| |
11:19 | Comete has quit IRC | |
11:27 | japerry has joined #ltsp | |
11:29 | carniel has joined #ltsp | |
11:31 | <Shingoshi> Is there anything that would prevent a host to network boot from it's own disks? Where the host was just another client of the system itself?
| |
11:32 | It wouldn't actually need to boot from the network.
| |
11:32 | I'm just curious how global the LTSP configuration can be made.
| |
11:34 | <Gadi> etyack: I don't believe so - but, it is one of the wishlist features. For now, you would need to change greeter.c
| |
11:34 | vagrantc has joined #ltsp | |
11:35 | <Shingoshi> Thanks!! I was hoping no one would think I was CRAZY!!
| |
11:35 | LOL
| |
11:36 | Where can I get explicit details on the filesystem itself?
| |
11:36 | <etyack> Gadi: ok. I c editor I shall become
| |
11:37 | evilx has joined #LTSP | |
11:38 | <Shingoshi> Gadi: Oh wait. Maybe I am CRAZY!! Your answer was for etyack!
| |
11:38 | * Gadi wasn't going to be the one to point that out | |
11:38 | * Gadi slowly steps away | |
11:38 | <Gadi> :)
| |
11:39 | Shingoshi: you can boot a virtual machine thin client to the server real easily
| |
11:40 | you could also run ldm on its own without *too* much hacking
| |
11:40 | <Shingoshi> Ok. So I'm possibly, marginally crazy!
| |
11:40 | lol
| |
11:42 | The thing that I realize, is that if the main server boots from the network, it wouldn't (might not) have access to files on it's own drive outside of it's own root.
| |
11:42 | carniel has quit IRC | |
11:42 | <Gadi> damn it, Jim, I'm a doctor not a bricklayer
| |
11:42 | <Shingoshi> lol
| |
11:43 | * Gadi has no clue what you are saying | |
11:43 | * Gadi smiles and nods | |
11:43 | <Shingoshi> Let me try again.
| |
11:44 | I'm thinking of creating a root filesystem, where each ARCH has it's own root.
| |
11:44 | Given that case, the main server could either boot locally or from the network.
| |
11:45 | If locally, it could create a mount in memory, giving it access to all the files on the system.
| |
11:45 | If by network, it might be able to do the same thing. But not sure yet.
| |
11:46 | By "mount in memory", I mean something like squashfs in memory.
| |
11:47 | Or a loop mount.
| |
11:47 | <Gadi> the *server* network boots?
| |
11:47 | or the client?
| |
11:48 | <Shingoshi> The server would be just another client. So the system would have the same boot procedure for all ARCHs.
| |
11:48 | <Gadi> The server is a client to what?
| |
11:48 | another server?
| |
11:49 | <Shingoshi> To the files served on it's own disks.
| |
11:49 | It's kind of like "the snake eating it's tail"
| |
11:50 | Just trying to relate the concept.
| |
11:50 | tjikkun_work has quit IRC | |
11:50 | <Gadi> what is the goal here?
| |
11:50 | why would you want the i386 client filesystem to have i386 server tools, as well?
| |
11:51 | just so a client could be a server to another client?
| |
11:51 | <Shingoshi> Having the same boot procedure copied to all machines. So no one machine would boot differently from another.
| |
11:51 | How many different ARCHs are supported by LTSP?
| |
11:52 | <Gadi> as many as the distro supports
| |
11:52 | <Shingoshi> Where is the list so that I can look at it.
| |
11:52 | ??
| |
11:52 | <Gadi> pick a distro
| |
11:52 | <vagrantc> not exactly true- debian only supports it on i386, amd64 and powerpc, currently
| |
11:52 | <Gadi> LTSP uses the host distro's tools to create a chroot
| |
11:53 | <vagrantc> although it used to support all arches supported by debian, and it's really just a matter of "turning on" all the other arches
| |
11:53 | <Gadi> vagrantc: I suppose it depends how you define "support"
| |
11:53 | :)
| |
11:53 | <ogra> for ubuntu officially its only i386 and amd64
| |
11:53 | <Shingoshi> I would mean the ability to boot them.
| |
11:54 | alkisg has joined #ltsp | |
11:54 | <vagrantc> Gadi: yeah, i basically disabled all the obscure architectures that i had no way of testing, as it was holding up progress for the other arches...
| |
11:54 | <ogra> if i finally find the time (very unlikely) i will implement ltsp for arm on ubuntu
| |
11:54 | <Shingoshi> I'm looking for: i386, amd64, powerpc(64), sparc, etc/
| |
11:55 | <ogra> (which debian should be able to just take then)
| |
11:56 | <Shingoshi> Hi ogra! Long time since I've chatted with you. Years in fact.
| |
11:56 | <vagrantc> i've got this mipsel machine sitting around...
| |
11:56 | there's some mips based laptops and thin clients from lemote.com
| |
11:57 | the ltsp upload in debian/unstable actually allows any architecture, but the buildd's don't build it...
| |
11:58 | <ogra> Shingoshi, yeah, and i'm not doing much in ltsp anymore apart from annoying Gadi since he needs that like bread :)
| |
11:59 | <vagrantc> i should poke the mips*/armel buildd's and ask them to start building ltsp again ... of course, those architectures are often the hold-ups.
| |
11:59 | Shingoshi: i actually get sparc hardware now and again, but there's one outstanding bug that hasn't gotten fixed that makes ltsp fail to work on debian sparc
| |
11:59 | <Shingoshi> lol
| |
11:59 | <Gadi> its a drug I cannot seem to do without
| |
11:59 | <vagrantc> http://bugs.debian.org/444087
| |
12:00 | <ogra> :)
| |
12:00 | <vagrantc> without that fixed, i have little energy for working with sparc
| |
12:00 | <Shingoshi> vagrantc: Is that limited to Debian?
| |
12:00 | <Gadi> vagrantc: that a driver bug
| |
12:00 | <vagrantc> Shingoshi: probably any distro using klibc for ipconfig
| |
12:00 | <Shingoshi> Does Suse work.
| |
12:01 | <vagrantc> Gadi: in the network card driver?
| |
12:01 | <Shingoshi> Can klibc be replaced with something else instead?
| |
12:01 | <vagrantc> Gadi: it works fine with dhclient
| |
12:01 | Shingoshi: yes, but that's more work than i want to bother with.
| |
12:01 | <Gadi> oh? nm - I missed the: DonUnimplemented SPARC system call 97
| |
12:01 | :)
| |
12:01 | dirigeant has quit IRC | |
12:02 | <Gadi> stgraber reports udhcpc as being much better than ipconfig, anyway
| |
12:02 | faster
| |
12:02 | only its in universe
| |
12:02 | :P
| |
12:02 | * vagrantc suffers a flashback | |
12:02 | <Gadi> seems klibc needs more loving in general
| |
12:02 | <vagrantc> used udhcpc with lessdisks
| |
12:03 | <Gadi> bak to the future, baby
| |
12:03 | *back
| |
12:03 | <vagrantc> overall, initramfs-tools has "just worked" for the vast majority of cases, and i'd like to stick with whatever it's using.
| |
12:07 | * Shingoshi looks at udhcpc. | |
12:07 | <Gadi> indeed - who cares if its slow and poorly written
| |
12:07 | :)
| |
12:07 | like everything else, we can make options
| |
12:08 | * Gadi recalls stgraber whipping up some code to use udhcpc if it is installed in the chroot, and ipconfig otherwise | |
12:09 | <SDuensin> Anybody have a recommendation for logging in to a LTSP server remotely (as in, across the internet where PXE is insane)?
| |
12:09 | <vagrantc> well, i'd hate to make a change to fix a corner case if it breaks things for a larger number of cases.
| |
12:10 | <NeonLicht> SDuensin, gPXE
| |
12:10 | <SDuensin> Yea, but how crazy is that? Would it be useable, or just a glorified tech demo?
| |
12:11 | <NeonLicht> SDuensin, no idea, haven't tried it yet
| |
12:11 | * SDuensin is reading. | |
12:12 | <SDuensin> My other idea was to come up with some minimal Linux install of X11 and ssh and run it in a VM.
| |
12:12 | * NeonLicht looks forwards to SDuensin to stop reading and start trying and telling us the results XDD | |
12:12 | <Gadi> SDuensin: dont do *X* over the internet
| |
12:12 | thats the insane part
| |
12:12 | use NX
| |
12:12 | <SDuensin> Yea, I know. But nothing else seems stable.
| |
12:12 | Lns has joined #ltsp | |
12:12 | <Shingoshi> NX?
| |
12:13 | <Gadi> www.nomachine.com
| |
12:13 | <NeonLicht> use FreeNX instead :-)
| |
12:13 | <SDuensin> I've used NX and it's pretty zippy. I didn't like all the server side stuff. Never seemed to work right. So maybe just the NX proxy part and LTSP behind it. But then I still need a way to get that on the remote machine.
| |
12:14 | <NeonLicht> I hate the server side of NX
| |
12:15 | <SDuensin> Me too. Which is why I don't want to use it.
| |
12:15 | <Shingoshi> I have freenx installed here.
| |
12:15 | <Gadi> you mean server-side tools?
| |
12:15 | <NeonLicht> starting by the fact that it installs itself on /usr/NX :(((
| |
12:15 | <Shingoshi> freenx-server.
| |
12:16 | <SDuensin> I always have issues with it not catching disconnects and just generally being weird.
| |
12:21 | spectra has joined #ltsp | |
12:26 | din_os has joined #ltsp | |
12:27 | The_Code has quit IRC | |
12:27 | F-GT has quit IRC | |
12:32 | davidj has joined #ltsp | |
12:32 | <davidj> Hey, guys.
| |
12:34 | <SDuensin> Howdy.
| |
12:35 | Subhodip has joined #ltsp | |
12:36 | Appiah_ has joined #ltsp | |
12:36 | Appiah has quit IRC | |
12:37 | <Lns> morning
| |
12:37 | <davidj> Anything going on?
| |
12:37 | * SDuensin is trying to gpxe over the internet | |
12:38 | * NeonLicht impatiently waits for SDuensin's results | |
12:38 | <Lns> might take a while to broadcast across the whole inet ;)
| |
12:39 | <NeonLicht> Lns, he's not broadcasting, that's why he uses gPXE instead of plain PXE :-)
| |
12:39 | <Lns> NeonLicht: i know.. :) just joshin
| |
12:39 | * Lns steps back from his outdated term | |
12:40 | <NeonLicht> hahaha
| |
12:41 | Appiah_ has quit IRC | |
12:41 | Appiah has joined #ltsp | |
12:44 | F-GT has joined #ltsp | |
12:44 | <SDuensin> Well, I can get the image to load. Dunno how to boot it yet.
| |
12:45 | Whoa! Got it. Except it's the wrong image. :-)
| |
12:45 | <NeonLicht> hahahahaha
| |
12:45 | * SDuensin has two PXE servers - trying to get to the "non-standard" one. | |
12:46 | Appiah has quit IRC | |
12:46 | <Lns> SDuensin: how's the boot/load speed? what kind of wan link are you on?
| |
12:46 | johnny has left #ltsp | |
12:46 | <davidj> SDuensin: How does your client find the gpxe server?
| |
12:47 | <SDuensin> Right now, I'm just bangin' commands into the gPXE command line.
| |
12:47 | Lns - This first test is over my wifi, so it doesn't really count for speed. If I can get it to work at all, then I'll move on.
| |
12:47 | <NeonLicht> davidj: there are no gPXE servers, it's just a PXE server, it's the gPXE the one which is instructed to query a (particular) PXE server
| |
12:48 | <Lns> SDuensin: cool. what's your intention of doing this in the long term?
| |
12:49 | <SDuensin> I want remote access to my LTSP box.
| |
12:49 | <davidj> SDuensin: You don't need pxe for remote access to ltsp
| |
12:49 | <NeonLicht> Lns, his intention is to inform me about hes results, since I'm being too lazy to try gPXE myself XDDDD
| |
12:49 | <SDuensin> davidj - Then tell me a good way to do it!
| |
12:50 | <Lns> SDuensin: vnc over ssh?
| |
12:50 | <davidj> NX, VNC, or a simple "X -query"
| |
12:50 | VNC over ssh is simplest.
| |
12:50 | <Lns> or do you want a true thin-client over a wan. if so, why?
| |
12:50 | <davidj> run "vncclient --via MyLtspServer :1
| |
12:50 | <NeonLicht> davidj, from a local thin client with no OS nstalled locally?
| |
12:51 | <SDuensin> I don't care how it works, I just want to get remotely the same thing my users attached to terminals see.
| |
12:51 | <davidj> A true thin client over a wan is likely to be too slow to be fun.
| |
12:51 | <Lns> or look at the vnc/inetd/ssh/gdm howto i wrote :) https://help.ubuntu.com/community/UbuntuLTSP/GDMVNCInetdssh
| |
12:51 | <davidj> SDuensin: Does the remote machine have linux on it?
| |
12:51 | <NeonLicht> davidj: maybe not, once the image and X are running locally on RAM
| |
12:51 | <SDuensin> Lns - Looked at, read, and implemented. :-)
| |
12:52 | davidj - It'll likely be in a VM, so sure, Linux. :-)
| |
12:52 | Appiah has joined #ltsp | |
12:52 | <Lns> SDuensin: so you're wanting a TRUE thin-client session..
| |
12:52 | <davidj> NeonLicht: But who wants to wait 20 minutes for the client to boot?
| |
12:52 | <SDuensin> Lns - Yes.
| |
12:52 | davidj - I can keep the boot image. It doesn't have to drag it across the net every time.
| |
12:53 | <davidj> SDuensin: The simplest way to do it is to ssh $server -l $youraccount, then run vncserver :3.
| |
12:53 | <SDuensin> I also saw an NX proxy hack to put NX between the terms and the server. That'd be good for this.
| |
12:53 | <davidj> Then run "vncclient -via $server :3" locally.
| |
12:54 | <SDuensin> davidj - That's fine for me. Heck, the howto Lns wrote is ok for me. I want road warriors to get the same thing on a laptop they get at their desk.
| |
12:54 | <davidj> SDuensin: Ah.
| |
12:54 | That's a bit different.
| |
12:55 | <Lns> yeah..well let us know how it performs. I'd have to think that it'd be pretty darn slow myself, but i'm probably wrong =)
| |
12:55 | <alkisg> SDuensin: kernel /ltsp/i386/vmlinuz ro ip=${ip}:${next-server}:${gateway}:${netmask}:${hostname}:eth0:none filename=${filename} can get you past the ipconfig dhcp request - if you're using ubuntu or debian.
| |
12:55 | <davidj> Doing it for a single user is one thing. Doing it for a group is likely to be a problem.
| |
12:55 | <SDuensin> davidj - Yes, it is. :-)
| |
12:55 | <davidj> I'd suggest you use something like openvpn to create a secure tunnel from each laptop back to the office.
| |
12:56 | <SDuensin> alkisg - Toss that in grub or what?
| |
12:56 | <alkisg> In the gpxe command line
| |
12:56 | <Lns> or you could ask Gadi about their LTSP bootstick
| |
12:56 | <davidj> Give each laptop a different openvpn key so when one gets lost you can "change the locks"
| |
12:56 | Gadi's bootstick is one option.
| |
12:56 | SDuensin: A few things to consider:
| |
12:56 | <alkisg> SDuensin: of course you should set all the variables first (ip, gateway, dns etc)
| |
12:57 | <SDuensin> alkisg - ah! thanks.
| |
12:57 | <Lns> I'd say that's probably best. You're booting the chroot locally, so it won't take so long
| |
12:57 | <davidj> If your road warriors are using LTSP, everything's going to run over the network.
| |
12:57 | Appiah has quit IRC | |
12:57 | Appiah has joined #ltsp | |
12:57 | <SDuensin> davidj - Yea, but with nx proxy, it's zippy.
| |
12:58 | <davidj> Zippy or not isn't the biggest problem.
| |
12:58 | <alkisg> SDuensin: and here is a patched gpxe versions that can load gpxe scripts from files: https://help.ubuntu.com/community/UbuntuLTSP/grubgpxe
| |
12:58 | <davidj> You need to consider security
| |
12:58 | also consider printing (if a user prints, it comes out at the office)
| |
12:58 | <SDuensin> Don't need printing so much.
| |
12:59 | <davidj> Can you install linux on each laptop, then configure it to automatically mount /home from your LTSP server via openvpn?
| |
12:59 | Then you get security, speed, and access to the office.
| |
13:00 | <SDuensin> that eliminates the nice central management though
| |
13:00 | <NeonLicht> SDuensin, already got the right server/image?
| |
13:00 | <SDuensin> NeonLicht - For?
| |
13:01 | * SDuensin has an LTSP running. | |
13:01 | <NeonLicht> SDuensin, you said you got the wrong one b4
| |
13:01 | <SDuensin> Oh! Off my PXE stuff. Still fartin' with it.
| |
13:01 | <NeonLicht> SDuensin, you mean a gXPE booted over the iNet one?
| |
13:02 | <davidj> SDuensin: What central management do you loose?
| |
13:02 | SDuensin: Do the laptops come to the office on a schedule, or are they always in the field?
| |
13:02 | <SDuensin> davidj - Maybe I'm misunderstanding. If I just mount /home for them, they don't get the apps installed on the server. Just their local ones.
| |
13:02 | I may never see the person. Remote people helping on a project could be anywhere.
| |
13:03 | That's why I was looking at a VM with something in it. "Here, download this. Double-click it."
| |
13:06 | Appiah has quit IRC | |
13:06 | Appiah has joined #ltsp | |
13:07 | <SDuensin> See? This is not so easy! :-P
| |
13:07 | Remaille has joined #ltsp | |
13:07 | <SDuensin> What is this mythical magical Gadi boot stick?
| |
13:11 | Appiah has quit IRC | |
13:12 | Appiah has joined #ltsp | |
13:14 | The_Code has joined #ltsp | |
13:14 | din_os has left #ltsp | |
13:14 | twinprism has quit IRC | |
13:16 | ogra has quit IRC | |
13:19 | ogra has joined #ltsp | |
13:20 | <Lns> ogra !
| |
13:20 | rjune has quit IRC | |
13:21 | rjune has joined #ltsp | |
13:23 | <Gadi> SDuensin: http://www.thesymbiont.com/index.php?option=com_content&task=blogcategory&id=137&Itemid=145
| |
13:23 | <SDuensin> Gadi - Interesting.
| |
13:25 | <davidj> Hello, Gadi. How are you?
| |
13:25 | rjune: You here?
| |
13:25 | <Gadi> davidj: cant complain
| |
13:25 | wrestling with usplash of all things
| |
13:25 | :P
| |
13:26 | <ogra> do you win ?
| |
13:26 | <davidj> heh
| |
13:26 | <Gadi> rarely
| |
13:26 | I always forget the magic to make the stars align so it actually produces an image
| |
13:26 | :)
| |
13:27 | <ogra> heh
| |
13:27 | well, KK will likely have KMS
| |
13:27 | Subhodip has quit IRC | |
13:27 | <Gadi> great! by then, I'll have moved from gutsy!
| |
13:27 | ;)
| |
13:27 | <ogra> heh, you will have to
| |
13:27 | <Gadi> ogra, you're not my real mother!
| |
13:28 | <ogra> how do you know? did you even ask your father ? :)
| |
13:28 | <NeonLicht> is it possible to indicate on lts.cnf to run a vncviewer proccess on a VT as you do with rdesktop?
| |
13:28 | <ogra> no
| |
13:28 | but we accept patches
| |
13:28 | <davidj> You could do it with ltsp 4
| |
13:29 | but you don't want to go down that road...
| |
13:29 | <ogra> you can surely script something for ltsp5 as well
| |
13:29 | <NeonLicht> I think I did it at some point, but I couldn't remember how
| |
13:29 | <ogra> by just grabbing the redesktop scripts
| |
13:29 | <Gadi> NeonLicht: you can write a screen script
| |
13:29 | its trivial
| |
13:29 | <NeonLicht> Gadi, I'm not very good at coding, I'm afraid :(
| |
13:30 | <Gadi> NeonLicht: copy rdesktop, as ogra says
| |
13:30 | /usr/share/ltsp/screen.d/rdesktop
| |
13:30 | <NeonLicht> I just do some PHP for databases web interfaces :-)
| |
13:30 | <ogra> i think it only has two or three lines
| |
13:30 | <NeonLicht> Gadi, ok, I try and see what I get
| |
13:30 | <Gadi> cp /usr/share/ltsp/screen.d/rdesktop /usr/share/ltsp/screen.d/vncviewer
| |
13:30 | <ogra> just replace them with the call for vnc
| |
13:31 | <Gadi> then: SCREEN_07=vncviewer
| |
13:31 | cliebow has quit IRC | |
13:32 | <NeonLicht> mmhhhh... there is a commented line on "rdesktop" which goes "#/usr/bin/x11vnc -display ${DISP} -passwd password -httpport 5800"
| |
13:33 | <Gadi> yeah, the rdesktop script has some artifacts from days gone by
| |
13:34 | <ogra> bah, usplash on ARM doesnt work :(
| |
13:34 | <Gadi> ogra: do you know of a way to test why a usplash theme .so file would fail?
| |
13:34 | or do I need to resort to trial and error
| |
13:34 | :)
| |
13:35 | <ogra> i think the latter, did you resort to 256 colors ?
| |
13:35 | <Gadi> 16 colors
| |
13:36 | file usplash_primary.png
| |
13:36 | usplash_primary.png: PNG image data, 1600 x 1200, 4-bit colormap, non-interlaced
| |
13:36 | <ogra> indexed ?
| |
13:36 | <Gadi> yeah
| |
13:36 | should it be RGB?
| |
13:36 | <ogra> no, indexed
| |
13:37 | strange
| |
13:37 | should actually work then
| |
13:37 | <Gadi> hmm...
| |
13:37 | <ogra> do you build it properly ?
| |
13:37 | <Gadi> I use the usplash-theme deb
| |
13:37 | so, I think I do
| |
13:40 | <ogra> did you follow the wiki ?
| |
13:40 | The_Code has quit IRC | |
13:40 | <Gadi> which one?
| |
13:40 | :P
| |
13:41 | <ogra> https://help.ubuntu.com/community/USplashCustomizationHowto
| |
13:42 | <Gadi> yeah
| |
13:42 | but the steps in that are handled by dpkg-buildpackage
| |
13:42 | :)
| |
13:42 | and these days, it uses pngtousplash
| |
13:53 | <rjune> davidj, sup!
| |
13:53 | !g
| |
13:53 | <ltspbot> rjune: "g" is Gadi!!!!!!!!!!!!!!!!!!!!!!!!
| |
13:54 | <davidj> rjune: Hey!
| |
13:54 | How are you?
| |
13:54 | <rjune> meh, same old same old.
| |
13:55 | looking to hang out the shingle again
| |
13:56 | <Lns> Can you specify params in lts.conf RCFILE_NN - i.e. "RCFILE_01 = crontab /etc/ltsp/root.crontab" ?
| |
13:59 | <vagrantc> Lns: RCFILE_01 = /full/path/to/script
| |
13:59 | Lns: relative to the chroot
| |
13:59 | <Lns> vagrantc: right...but can i specify parms to the script?
| |
13:59 | <vagrantc> Lns: i.e. /opt/ltsp/i386/etc/foobar would be RCFILE_NN = /etc/foobar
| |
13:59 | Lns: no.
| |
13:59 | <davidj> rjune: Got a sec?
| |
13:59 | <Lns> vagrantc: ok..so i'll have to write a wrapper script
| |
13:59 | <ogra> write a wrapper script
| |
13:59 | <vagrantc> Lns: write a custom script.
| |
13:59 | <Lns> thx =)
| |
14:00 | * Lns is revamping automated shutdown to allow admins to disable / modify times if desired | |
14:17 | <Lns> Any special considerations when installing openssh-server in client chroot?
| |
14:17 | <ogra> apart from no security ?
| |
14:17 | <Lns> ogra: ?
| |
14:17 | <ogra> (all your clients have the same key)
| |
14:18 | <vagrantc> and regenerating the keys on the fly wouldn't give you any way to verify them
| |
14:18 | <Lns> ogra: they do by default, for LDM ssh stuff right?
| |
14:18 | <vagrantc> Lns: that's going the other direction
| |
14:18 | <ogra> right
| |
14:18 | <vagrantc> although the key verification is still suspect
| |
14:18 | <ogra> indeed
| |
14:19 | but not as insecure as running a server on the clienzt
| |
14:19 | <vagrantc> i.e. possible to do a man-in-the-middle attack by hacking NBD/NFS to return an invalid known_hosts
| |
14:19 | ogra: agreed.
| |
14:19 | <Lns> hmm.
| |
14:19 | <vagrantc> well, it depends on what you're doing, really.
| |
14:19 | <Lns> well i need a way to execute commands on all clients at once...
| |
14:19 | <vagrantc> i mean, if you just need a way to trigger certain commands, and you lock the keys down to run only those commands...
| |
14:19 | <ogra> restrict to a single user key
| |
14:19 | <Lns> kinda like what 4.2 did i guess. but i'd like to make it as secure as possible
| |
14:19 | <ogra> copy that into the chroot as well
| |
14:20 | <Lns> ogra: sounds like a plan
| |
14:20 | <ogra> that gives you at least some security
| |
14:20 | <Lns> yeah i was going to create a root-based ssh key and use auth for that
| |
14:21 | <vagrantc> for security, it's basically worthless ... for testing purposes or experimentation, no big deal
| |
14:21 | Lns: i'd experiment with locking the keys down to particular commands, if you can.
| |
14:21 | <Lns> vagrantc: ok, never thought you could do that - i'll google it =)
| |
14:22 | <vagrantc> in ~/.ssh/authorized_keys: command="exact_command --with --exact arguments" SSH_KEY
| |
14:22 | <Lns> vagrantc: noooice. =) thanks!
| |
14:22 | <vagrantc> Lns: but you can't securely pass it arguments that way ...
| |
14:23 | i.e. any command you try to run will run the command specified in authorized_keys
| |
14:23 | with the arguments specified there
| |
14:23 | <Lns> vagrantc: well it'll be to do a single command (disable root's one active cronjon on the client) so there wont' be any custom params
| |
14:23 | <Gadi> whose key are you worried is spoofed?
| |
14:23 | <Lns> cronjob*
| |
14:23 | <Gadi> authorized_keys has public keys
| |
14:23 | <vagrantc> the client's host keys
| |
14:24 | <Gadi> and say someone spoofs the clients keys
| |
14:24 | then, what?
| |
14:24 | they can have a command happen to them?
| |
14:25 | <ogra> you still need the root key from the initiating side
| |
14:25 | <vagrantc> that's why i say use authorized_keys
| |
14:25 | so you're not running arbitrary commands from the commandline on a spoofed client.
| |
14:25 | feel free to call me paranoid.
| |
14:26 | <Lns> better paranoid than microsoft security practice =p
| |
14:26 | * Gadi just wants to understand the security hole | |
14:26 | <Lns> ba-da-pssshhhh
| |
14:26 | <Gadi> :)
| |
14:27 | I suppose I could make a computer that I have root access to look like a thin client so that the arbitrary command runs on me
| |
14:27 | <vagrantc> i think you might run into issues where ssh defaults will complain a lot, but not positive ... i.e. you're trying to log into thinclient1 with ssh key foo, but known_hosts says thinclient2 has that key
| |
14:27 | <Gadi> of course, that requires root access
| |
14:27 | <vagrantc> well, you can mount the filesystem and emulate a thin client pretty easily...
| |
14:28 | RobertLaptop has quit IRC | |
14:28 | <Gadi> right, in which case the arbitrary command intended for any thin client runs on me
| |
14:28 | which does not seem like a security hole
| |
14:29 | <vagrantc> if that arbitrary command is soemthing like "ssh root@someserver" and it logs your keystrokes ...
| |
14:29 | that's a stupid thing to do, which is why i don't want to give anybody the opportunity to do something like that.
| |
14:29 | <Lns> vagrantc: well that'd be a pretty bad thing to do in the first place, regardless of what facilities you use
| |
14:30 | <Gadi> I see
| |
14:30 | RobertLaptop has joined #ltsp | |
14:30 | <vagrantc> so forced commands prevent you from running stupid things from an interactive commandline
| |
14:30 | <Gadi> not really
| |
14:30 | <Lns> lemme clarify a bit - I want to log into the thin-client from the ltsp server via ssh by using priv/pub keys and no password
| |
14:30 | <Gadi> because the same person that sets the command sets what commands can be set
| |
14:30 | <vagrantc> right
| |
14:31 | <Lns> i install openssh-server in chroot, create a new priv/pub keypair
| |
14:31 | set it up so only that key can be authed, to run a very specific command only
| |
14:31 | <vagrantc> Gadi: well, sure, they could fire up a shell or some other interactive command to do something stupid
| |
14:32 | if you use very specific commands, designed for particular tasks, you limit the risk significantly.
| |
14:32 | <Lns> vagrantc: how could someone get access to the chroot in the first place without a password?
| |
14:32 | <Gadi> no, I mean, the same admin that sets up the ssh key to run command foo, is the same one that runs command foo at regular intervals
| |
14:32 | regardless of the authenticity of the client
| |
14:32 | we are not worried about the root key being compromised
| |
14:32 | <vagrantc> well, there's thousands of root exploits for users with shell access, and localapps now gives everyone shell access to the thin client.
| |
14:32 | <Gadi> only the host keys
| |
14:33 | <Lns> vagrantc: well good thing i'm running hardy =p~~ ~
| |
14:33 | Gadi: ah
| |
14:33 | <Gadi> I dont see how that exploit impacts the root ssh-key being exploited
| |
14:33 | but, I must be missing something obvious
| |
14:34 | <vagrantc> Gadi: yes, but running foo from authorized_keys doesn't give you the ability to log in to an interactive shell, run foo, and then think "oh, while i'm here i may was well do ..."
| |
14:34 | you have to think about what commands you're going to allow in advance
| |
14:34 | <Gadi> but, only root on the server cn run foo
| |
14:34 | <vagrantc> which hopefully will prevent you from putting stupid things in authorized_keys
| |
14:36 | Gadi: by running arbitrary commands as root on the interactive commandline, on a machine which is fairly easy to man-in-the-middle, they can log the commands and keystrokes you type ...
| |
14:37 | which may be something stupid
| |
14:37 | <Gadi> but that is an exploit of local apps
| |
14:37 | not of a root ssh key exchange
| |
14:37 | <vagrantc> not necessarily... anyone can mount NBD/NFS and pretend to be a thin client.
| |
14:37 | <Gadi> because it is an exploit of the local root user
| |
14:37 | not the server root user
| |
14:37 | <vagrantc> anyone has root access to the thin client.
| |
14:38 | <Gadi> right
| |
14:38 | <vagrantc> and the thin client's host keys
| |
14:38 | <Gadi> but the NFS root only has a public root ssh key
| |
14:38 | not the private one
| |
14:38 | ByPasS has joined #ltsp | |
14:38 | <Gadi> only the server would have the private one
| |
14:38 | <vagrantc> and anything you do on that spoofed thin client could be logged.
| |
14:39 | <Gadi> ok...
| |
14:39 | <vagrantc> which could include logging into other machines on the network...
| |
14:39 | or any number of other things best to avoid
| |
14:40 | could even be done accidentally
| |
14:40 | <Gadi> indeed
| |
14:40 | we need more skulls and crossbones over LTSP
| |
14:40 | :)
| |
14:40 | <vagrantc> heh
| |
14:40 | hanthana_ has joined #ltsp | |
14:41 | <vagrantc> granted, all of this is fairly corner-case
| |
14:41 | <ogra> Gadi, you didnt submit that logo :P
| |
14:41 | <vagrantc> but running ssh/sshd on a thin-client isn't nearly as secure as people normally expect ssh connections to be. that's the basic point.
| |
14:42 | * ogra wonders what came out of the contest ... i voted looong ago | |
14:42 | <ogra> CAN-o-SPAM, ^^^ was there any outcome yet ?
| |
14:44 | <Lns> me too! can't find results.
| |
14:46 | <vagrantc> given the number of logos compared to the number of votes, i wouldn't be surprised if there was no clear winner.
| |
14:48 | * Gadi just voted today *blush* | |
14:52 | <Lns> lol
| |
14:54 | <ByPasS> I'm trying to get an HP Laserjet 1000(USB) to work on a thinclient, I had no problems so far with parallel printers this is my first attempt on a USB one
| |
14:55 | I've added in lts.conf PRINTER_0_TYPE = U & PRINTER_0_DEVICE = /dev/usblp0
| |
14:55 | added the printer in cups as hp laserjet 1000 but each time I send something or simply try to print test page
| |
14:55 | <vagrantc> i had a clear favorite logo, and a handful of 2nds and 3rds ... i had a hard time making up my mind.
| |
14:56 | dirigeant has joined #ltsp | |
14:56 | <ByPasS> I get an error saying unable to write print data: broken pipe
| |
14:56 | <Gadi> ByPasS: HPLJ1000 is not supported
| |
14:56 | by jetdirect
| |
14:56 | <ByPasS> maybe worth mentioning that jetpipe dies on the client
| |
14:56 | * jammcq is sitting on the edge of his seat, waiting for the outcome of the logo contest | |
14:56 | <ByPasS> ohhhh :(
| |
14:56 | <Gadi> it is not a real printer
| |
14:57 | <ByPasS> good thing Ive been on to ask I've been fooling around quite a while on it eh
| |
14:57 | <Gadi> http://h10025.www1.hp.com/ewfrf/wc/fastFaqLiteDocument?lc=en&cc=uk&dlc=en&docname=bpj06101
| |
14:59 | <ByPasS> thanks alot gadi :)
| |
14:59 | <Lns> I can't believe anyone (well, ok, it's HP after all) thought it would be a good idea to have a firmware-less printer. Like Winmodems...screams vendor lock-in
| |
15:00 | it's too bad they were so popular. I had to face one not working in ltsp myself
| |
15:02 | hanthana has quit IRC | |
15:02 | <Gadi> ByPasS: np
| |
15:02 | I think the 1000 was a best seller
| |
15:02 | comes up a lot in ltsp
| |
15:02 | :)
| |
15:02 | din_os has joined #ltsp | |
15:03 | <Lns> so it looks like with using command="string" in authorized keys, you're effectively limiting yourself to a single command when logging in (i.e. can't specify *which* command="string" to use i'm presuming)
| |
15:03 | so I could craft my deal to do that, but then i can't expand it at all in the future without some hackery
| |
15:04 | swapping out different authorized_keys with different commands on the fly or something dumb like that
| |
15:05 | <vagrantc> Lns: use a separate key for each command...
| |
15:05 | * Gadi curses usplash | |
15:06 | <Lns> vagrantc: ah =) *slaps head*
| |
15:06 | * CAN-o-SPAM lets everyone know outcome of the logo contest coming tomorrow | |
15:06 | * CAN-o-SPAM thinks get developers to cast votes sometimes isn't easy ;) | |
15:09 | <vagrantc> CAN-o-SPAM: well, as far as i know, this is the first time LTSP has really had a vote, at least since i've been involved.
| |
15:09 | much more used to interactive discussion and dialogue
| |
15:10 | <epsas> where are the contestants?
| |
15:10 | <jammcq> we've got all the contestants trapped on a cruise liner, floating off of the bahamas
| |
15:14 | <CAN-o-SPAM> http://www.disklessworkstations.com/cgi-bin/web/contestlogos.html
| |
15:14 | <rjune> jammcq!
| |
15:14 | <CAN-o-SPAM> if you want to see the logos ^
| |
15:14 | <jammcq> rjune!
| |
15:14 | <rjune> how goes it?
| |
15:14 | <jammcq> going well.
| |
15:14 | how's by you?
| |
15:15 | * rjune likes http://www.disklessworkstations.com/web/images/logos/sBalneaves1.png best | |
15:15 | <rjune> been alright
| |
15:17 | <jammcq> rjune: yeah, that was my 1st choice too
| |
15:20 | <rjune> what do we know, eh?
| |
15:23 | hanthana_ has quit IRC | |
15:23 | twinprism has joined #ltsp | |
15:27 | makghosh has joined #ltsp | |
15:30 | mahdi_ja has joined #ltsp | |
15:30 | <mahdi_ja> hi all.
| |
15:30 | can i use my pc instead of a thin client.
| |
15:32 | <CAN-o-SPAM> mahdi_ja: sure, as long as your PC can PXE boot
| |
15:35 | CAN-o-SPAM has quit IRC | |
15:48 | etyack has left #ltsp | |
15:51 | ByPasS has quit IRC | |
15:52 | evilx has quit IRC | |
15:53 | mahdi_ja has quit IRC | |
16:05 | Nubae has joined #ltsp | |
16:05 | <Lns> If I were to want IP addresses of connected clients, would this be the best way to do it (particularly regarding the sed cmds)? :
| |
16:05 | arp -an | cut -d " " -f 2 | sed s/"("//g | sed s/")"//g
| |
16:05 | (I just want to strip out the parentheses)
| |
16:05 | Nubae1 has quit IRC | |
16:06 | <Lns> I'm sure there's a cleaner way to do it...
| |
16:07 | <vagrantc> you could do a single call to sed by using sed -e s/"("//g -e s/")"//g
| |
16:07 | <Lns> vagrantc: ah thx
| |
16:15 | ccherret1 has joined #ltsp | |
16:15 | spectra has quit IRC | |
16:16 | ccherrett has quit IRC | |
16:19 | alkisg has quit IRC | |
16:20 | <din_os> vagrantc: thx from me too, I was looking for that one, I'll try it tomorrow
| |
16:24 | alkisg has joined #ltsp | |
16:25 | alkisg has joined #ltsp | |
16:30 | Remaille has quit IRC | |
16:46 | alkisg has quit IRC | |
16:57 | din_os has left #ltsp | |
17:02 | nicros is now known as nicros_away | |
17:08 | <Lns> vagrantc: i'm really sorry if i'm bugging you and/or being stupid - but can you tell me again, why restricting "authorized_keys" to certain commands, given the ASSUMPTION that whoever is sitting at a root shell is ALREADY authorized since they've got access to said shell, would be beneficial?
| |
17:09 | I'm trying to justify adding keys && rebuilding a chroot every time i want to add a command to my script would be worth it, instead of just using ssh commands in my script alone.
| |
17:09 | <vagrantc> Lns: because the ssh host keys (private and public) of the thin client are available to anyone on the network via NBD
| |
17:10 | Lns: so someone could "pretend" to be the thin client you're trying to log into by stealing it's ip address and ssh host key
| |
17:10 | jammcq has quit IRC | |
17:11 | <Lns> vagrantc: ok...but host key security aside, i'm talking strictly about restricting authorized_keys to specified commands
| |
17:11 | <vagrantc> Lns: and then they can log any commands or keystrokes on any sessions you run.
| |
17:11 | depends on what your threat level is.
| |
17:11 | <Lns> in all reality isn't what you're talking about a wider security issue, given anyone can boot a computer via pxe on the local net?
| |
17:12 | (at least in my situation :) )
| |
17:12 | <vagrantc> sure, it's a wider issue.
| |
17:12 | <Lns> ok... i do see your point
| |
17:12 | <vagrantc> the issue is really that people expect ssh to be secure, and you need to know that this is a compromised ssh connection.
| |
17:12 | <Lns> vagrantc: but just in the sense that you might run a command on an 'evil' client
| |
17:13 | <vagrantc> and if you understand all the implications of that, and don't make mistakes, you'll probably be fine.
| |
17:13 | Lns: yes. in particular, don't ever do anything on one client that could give you access to other machines on the network...
| |
17:14 | but it really depends on your risk assessment...
| |
17:14 | <Lns> vagrantc: cool. Ok, so i'm going to rewrite my little doodad to just use a single userkey instead of making one for each command to be executed, since what your'e saying is outside that scope of security
| |
17:14 | <vagrantc> the worst case scenario is someone can use it to gain root on your whole network :)
| |
17:14 | <Lns> vagrantc: right. this is for simple commands like shutting a system down
| |
17:15 | vagrantc: well if they have root on the ltsp server, i'd have to say that's the case anyway :)
| |
17:15 | <vagrantc> Lns: well, you're opening some additional possibilities to gain root on the ltsp server, even if they may be highly unlikely.
| |
17:16 | <Lns> vagrantc: such as?
| |
17:16 | <vagrantc> basically, all security issues come down to creating additional possibilities to gain root.
| |
17:16 | <Lns> ah
| |
17:16 | twinprism has quit IRC | |
17:17 | <Lns> well thank you vagrantc that makes it more understandable
| |
17:17 | <vagrantc> Lns: person boots their evil thin client, you log into their evil thin client and run a command as root, you accidentally type in your user or root password, and they use it to log into the server...
| |
17:17 | or you log into the thin client and ssh to another machine from it, type in your password ...
| |
17:18 | i'm not even realy savvy with this stuff :)
| |
17:18 | <Lns> vagrantc: ok so you're saying basically anything that you do remotely on a TC is vulnerable since it could be potentially comprimised/spoofed by an evil client
| |
17:19 | <vagrantc> Lns: basically
| |
17:19 | <Lns> ok cool.
| |
17:19 | that's totally sane
| |
17:19 | <vagrantc> so be careful.
| |
17:19 | <Lns> i will - in a nutshell i'm making a simple menu-driven script that has predefined commands on the ssh commandline already, the techs that will use it won't even have the ability to change it (from the script)
| |
17:20 | or at least the knowledge.. and if they did have the knowledge, they'd have the knowledge to compromise a lot more than that alone
| |
17:20 | given they have 'sudo' rights
| |
17:20 | mikkel has quit IRC | |
17:21 | <Lns> i hate leaving it up to "they won't know how to do that" but an admin acct. is an admin acct...
| |
17:21 | i can't take that away from them ;)
| |
17:22 | <vagrantc> there's enough user-level exploits that user access isn't far from root access
| |
17:23 | <Lns> very true
| |
17:24 | <Lumiere> Lns: give limited sudo
| |
17:25 | sudo is inredibly flexible ;)
| |
17:25 | Lns: what commands do you want them to access
| |
17:25 | <Lns> Lumiere: sudo is limited
| |
17:26 | only the on-site techs and I have sudo/root
| |
17:26 | <Lumiere> I am saying for the techs
| |
17:26 | if they only need to run a few commands
| |
17:26 | <Lns> Lumiere: i'm creating a script to automate shutdown of all thin clients, as well as other things i can come up with
| |
17:26 | <Lumiere> limit their sudo access to those commands
| |
17:27 | <Lns> Lumiere: that's a very good idea, though the techs need unrestricted access...i'm a subcontractor, but they're the techs that are learning to care for the network in the long term
| |
17:27 | <Lumiere> ah
| |
17:27 | <Lns> i don't really have the right to say "you can only do this"
| |
17:27 | i'm just trying to make these processes as foolproof and simple as possible for them
| |
17:27 | <Lumiere> yea
| |
17:28 | <Lns> having an ssh server in the chroot really is so valuable imho. that opens the door to all sorts of administration goodies :)
| |
17:28 | as long as you restrict that access as much as possible
| |
17:28 | <vagrantc> sure opens doors, yes. :)
| |
17:28 | <Lns> vagrantc: lol, yes...
| |
17:29 | but you can always restrict access to a certain admin workstation, a single user account
| |
17:29 | of course you can only do so much
| |
17:31 | * Lns wonders how iTalc does all it does on the clientside, and if that's any more secure | |
17:36 | <Lumiere> if it were me
| |
17:37 | I'd do the opposite
| |
17:37 | I would do something where clients that are logged in
| |
17:37 | check some sort of server side
| |
17:37 | for administration goodies
| |
17:38 | CAN-o-SPAM has joined #ltsp | |
17:38 | nicoAMG has quit IRC | |
17:41 | <Lns> Lumiere: i don't follow
| |
17:46 | Gadi has left #ltsp | |
17:47 | <Lumiere> Lns: something like cfengine does
| |
17:47 | where the client polls a server app for things to do
| |
17:47 | every x minutes
| |
17:47 | sshd on clients has private key issues
| |
17:48 | <Lns> ah, i see..i was thinking about that too, though ianap.. some sort of ltsp API for client commands, etc.
| |
17:48 | right?
| |
17:48 | <Lumiere> yea
| |
17:48 | <Lns> gotcha
| |
17:48 | that would be nifty
| |
17:48 | Faithful has quit IRC | |
18:19 | vvinet has joined #ltsp | |
18:23 | <Lns> LOL - http://www.youtube.com/watch?v=UjRdA25yauU
| |
18:39 | Lns has quit IRC | |
18:48 | Ahmuck has joined #ltsp | |
18:52 | laprag has joined #ltsp | |
18:53 | Faithful has joined #ltsp | |
18:59 | <Faithful> nComputing boxes work with LTSP ???
| |
19:02 | bobby_C has quit IRC | |
19:04 | vagrantc has quit IRC | |
19:13 | craver has joined #ltsp | |
19:13 | craver is now known as nicros | |
19:19 | ccherret1 is now known as ccherrett | |
19:20 | CAN-o-SPAM has quit IRC | |
19:24 | nicros has quit IRC | |
19:27 | nicros has joined #ltsp | |
19:44 | Faithful has quit IRC | |
20:13 | japerry has quit IRC | |
20:27 | laprag has left #ltsp | |
20:56 | dirigeant has quit IRC | |
20:59 | japerry has joined #ltsp | |
21:28 | vagrantc has joined #ltsp | |
21:32 | pmatulis has quit IRC | |
21:41 | michelboaventura has joined #ltsp | |
21:49 | makghosh has quit IRC | |
23:02 | CaScAdE^FarAway has joined #ltsp | |
23:05 | try2free has joined #ltsp | |
23:13 | try2free has left #ltsp | |
23:17 | vagrantc has quit IRC | |
23:18 | CaScAdE^1arAway has quit IRC | |
23:19 | faustino3331 has quit IRC | |
23:20 | faustino333 has joined #ltsp | |
23:23 | nicros has quit IRC | |
23:24 | nicros has joined #ltsp | |
23:36 | Egyptian[Home]1 has joined #ltsp | |
23:48 | safqvbss has joined #ltsp | |
23:53 | safqvbss has quit IRC | |