IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 1 March 2008 (all times are UTC)
| 00:22 | cyberorg has quit IRC | |
| 01:13 | <vagrantc> warren: debian has had very little problem with bind-mounting /etc/X11/xorg.conf in the last 2 years ...
| |
| 01:14 | warren: also, all that code regading the handling of the X configuration file is duplicated from the init script
| |
| 01:14 | warren: no need to have the screen script messing with that at all.
| |
| 01:17 | warren: at one point, gadi had made some common code between all the screen scripts ... i.e. the ldm and startx screen scripts contain(ed) a fair amount of duplicate code ...
| |
| 01:19 | warren: i object to using /tmp/xorg.conf ... if we *must* use some writeable location, it needs to be something other than /tmp ... we've moved most of the code to /var/run ...
| |
| 01:20 | * vagrantc is also tempted to rename the script "xdmcp" and make a symlink for "startx" | |
| 01:21 | <vagrantc> much more descriptive of what it actually does.
| |
| 02:06 | cyberorg has joined #ltsp | |
| 02:13 | <warren> vagrantc, "xdmcp" is a very good idea
| |
| 02:14 | vagrantc, OK, I might be able to be convinced to bind mount it upon /etc/X11/xorg.conf
| |
| 02:14 | vagrantc, I want to do some testing first
| |
| 02:15 | <vagrantc> the only bug i encountered so far with bind-mounting /etc/X11/xorg.conf ... was i had to patch configure-x.sh to use cp instead of mv ...
| |
| 02:32 | * ogra *yawns* | |
| 02:33 | <ogra> vagrantc, the XFCFG stuff came from francis
| |
| 02:34 | warren, the startx scren script works just fine for me, please dont break it :)
| |
| 02:35 | <warren> ogra, if you seriously wanted the shite that was there, then we need to have plugins for it.
| |
| 02:35 | <ogra> (do whatever you want with it beyond that though :) )
| |
| 02:35 | <warren> ogra, we didn't remove XFCFG
| |
| 02:35 | <ogra> i just dont want it to break :) remove as you like ;)
| |
| 02:36 | <warren> I think the full plan is crystalizing in my brain now
| |
| 02:36 | <ogra> as lon as it goes on doing what it should (startingxdmcp in any way) i'm fine
| |
| 02:36 | <warren> Yes, and I think it will do it better, with simpler code.
| |
| 02:36 | <ogra> i never touched that script
| |
| 02:37 | it came originally from 4.2 scott and francis cheanged a lot in int last release cycle
| |
| 02:37 | <warren> gah. I seem to have REALLY GOOD IDEAS when I'm on this drug.
| |
| 02:37 | zolpidem tartrate (generic of ambien)
| |
| 02:38 | The drug lowers inhibitions in a similar way to alcohol, except without the inebriation.
| |
| 02:39 | <ogra> isnt that supposed to make you sleepy instead ?
| |
| 03:22 | vagrantc has quit IRC | |
| 03:25 | Q-FUNK has joined #ltsp | |
| 04:28 | sepski has joined #ltsp | |
| 04:32 | burnerx has quit IRC | |
| 04:32 | burnerx has joined #ltsp | |
| 04:51 | burnerx has quit IRC | |
| 04:53 | otavio_ has joined #ltsp | |
| 04:54 | otavio has quit IRC | |
| 05:01 | <laga> ogra: you here?
| |
| 05:02 | sepski has quit IRC | |
| 05:05 | <laga> ogra: i'd like to integrate mythbuntu-diskless into the alternate disk - we're getting our own alternate disks built now on cdimage. i've thought about just preseeding ltsp-client-builder/build-client-opts to add --mythbuntu. but that'll still mean it'll install ltsp-server-standalone instead of mythbuntu-diskless*. do you think we can make ltsp-client-builder.postinst more generic or can i just fork it?
| |
| 05:17 | otavio_ is now known as otavio | |
| 05:45 | <laga> ogra: never mind, i need to fork ltsp-client-builder because everything needs to be on one network interface :/
| |
| 05:45 | this is gonna be fun
| |
| 06:07 | mhterres has joined #ltsp | |
| 06:07 | mhterres has left #ltsp | |
| 06:09 | Guaraldo has joined #ltsp | |
| 06:09 | DonSilver has joined #ltsp | |
| 06:10 | DonSilver has quit IRC | |
| 06:15 | K_O-Gnom has joined #ltsp | |
| 06:52 | J45p3r has joined #ltsp | |
| 07:23 | spectra has joined #ltsp | |
| 07:40 | tux_440volt has joined #ltsp | |
| 07:48 | bobby_C has joined #ltsp | |
| 08:42 | tux_440volt has quit IRC | |
| 09:22 | mccann has quit IRC | |
| 09:38 | cliebow has joined #ltsp | |
| 10:10 | bobby_C has quit IRC | |
| 10:22 | tux_440volt has joined #ltsp | |
| 10:57 | Jorge_ has joined #ltsp | |
| 10:58 | <Jorge_> Is it possible to setup a ltsp system with one server on the router, the terminals on the router as well as the internet connection? Or does the server require two network cards?
| |
| 11:02 | <jammcq> Jorge_: yes, it's absolutely possible
| |
| 11:02 | I think there's some docs on the wiki about a single-nic setup
| |
| 11:03 | <Jorge_> jammcq: thanks! I'll continue my research then.
| |
| 11:04 | Is the server software in the Ubuntu repos?
| |
| 11:04 | <jammcq> yeah
| |
| 11:06 | <Jorge_> hmmmmm. I can't find it in the Add/Remove
| |
| 11:06 | <jammcq> what version of Ubuntu?
| |
| 11:06 | <Jorge_> 7.10
| |
| 11:06 | <jammcq> look for "ltsp-server" or "ltsp-server-standalone"
| |
| 11:06 | both of those should be there
| |
| 11:07 | <Jorge_> nope, not there
| |
| 11:08 | maybe I have a setting wrong?
| |
| 11:18 | subsume has joined #ltsp | |
| 11:19 | <Jorge_> Any thoughts on what I'm doing wrong so that I can't find the files?
| |
| 11:24 | Maybe Ubuntu removed ltsp support in the latest edition?
| |
| 11:25 | <subsume> What are you trying to do?
| |
| 11:25 | Sorry, just joined
| |
| 11:25 | Jorge_: ^
| |
| 11:25 | <Jorge_> I'm trying to install ltsp in Ubuntu
| |
| 11:25 | <subsume> same here.
| |
| 11:25 | I got a 30 computer lab I'm trying to get it rocking
| |
| 11:25 | you want thin or fat?
| |
| 11:25 | <Jorge_> thin
| |
| 11:25 | It seems to be missing from the current repos though
| |
| 11:26 | <subsume> 7.10?
| |
| 11:26 | <Jorge_> yup
| |
| 11:26 | <jammcq> it comes up just fine on my 7.10 box
| |
| 11:26 | <Jorge_> Add/Remove app?
| |
| 11:26 | <jammcq> apt-cache search ltsp
| |
| 11:26 | <Jorge_> ah, so it is not accessible via the add/remove tool
| |
| 11:27 | <jammcq> apparently. it's in the synaptic tool tho
| |
| 11:27 | <subsume> Jorge_: stuff like this you're going to need command line =)
| |
| 11:27 | <jammcq> ltsp is NOT an application
| |
| 11:27 | <Jorge_> no problem....time to learn
| |
| 11:29 | Ok, found it in the synaptic tool
| |
| 11:29 | edubuntuvsMSTS has joined #ltsp | |
| 11:30 | <edubuntuvsMSTS> hi
| |
| 11:30 | I need some advice
| |
| 11:31 | Today on /. as part of un-related post, the following was posted: "* Remove the crippling in Terminal Server, allow multiuser use over networks. If you can't afford to upgrade all your computers to Vista you can use the old ones as terminals to your Windows Home Server." in a reply
| |
| 11:31 | Can someone here explain what this means?
| |
| 11:32 | Reference: http://slashdot.org/article.pl?sid=08/03/01/1312233 (Search for "Microsoft could have done plenty..." post)
| |
| 11:36 | <Jorge_> which file to install from the list in synaptic?
| |
| 11:36 | edubuntuvsMSTS has quit IRC | |
| 11:50 | <Jorge_> anyone? There are quite a few to choose from.
| |
| 11:50 | gonzaloaf has quit IRC | |
| 11:50 | <laga> Jorge_: odepends what you want :)
| |
| 11:51 | <Jorge_> laga, I want a basic thin setup
| |
| 11:51 | <laga> on the server, you should install ltsp-server or ltsp-server-standalone. the standalone packages also comes with a dhcp server
| |
| 11:52 | <Jorge_> laga, my router does the dhcp
| |
| 11:53 | edubuntuvsMSTS has joined #ltsp | |
| 11:58 | <warren> ogra_cmpc, ogra: any idea why configure localdev and starting sound happens after you start the screen scripts?
| |
| 11:59 | <cliebow> Jorge_, wll your router pass filename for tftp?
| |
| 12:00 | <Jorge_> good question
| |
| 12:00 | <cliebow> better to pass dhcp with your ltsp server...
| |
| 12:00 | <Jorge_> cliebow, then I need two nics?
| |
| 12:00 | <edubuntuvsMSTS> > The following /. story had this in it: "* Remove the crippling in Terminal Server, allow multiuser use over networks. If you can't afford to upgrade all your computers to Vista you can use the old ones as terminals to your Windows Home Server." (Ref: http://slashdot.org/article.pl?sid=08/03/01/1312233 - search for "Microsoft could have done plenty..." post) - what does this mean / refer to...
| |
| 12:00 | ...please?
| |
| 12:01 | <cliebow> Jorge_, why is that?
| |
| 12:02 | <Jorge_> cliebow, does my router not have to manage the dhcp duties?
| |
| 12:02 | <cliebow> your server can usually provide dhcp for your netwotk...
| |
| 12:02 | anything fancy going on there?
| |
| 12:02 | <jammcq> edubuntuvsMSTS: dunno, but it's NOT LTSP
| |
| 12:03 | <Jorge_> cliebow, with the interet hooked to the router and all the computers wireless? Nothing unusual about my network here. Very stock.
| |
| 12:03 | <edubuntuvsMSTS> Yeah, I know - it's referring to Microsoft Terminal Services
| |
| 12:03 | <jammcq> ok, so why are you asking about it here? this isn't the MS terminal services channel
| |
| 12:03 | <cliebow> jammcq:found the prob in my vmware..of 2 nics winders picked one...linux the other..
| |
| 12:04 | <jammcq> cliebow: ah, that old problem
| |
| 12:04 | <cliebow> woke up at 4 am with the answer...
| |
| 12:05 | indradg has joined #ltsp | |
| 12:05 | <edubuntuvsMSTS> Because I'm trying hard to push an EduBuntu solution and if I can find out what this post about Microsoft Terminal Services being crippled means it gives me much more ammunition!
| |
| 12:06 | <laga> well
| |
| 12:06 | it's microsoft
| |
| 12:06 | <edubuntuvsMSTS> and?
| |
| 12:06 | <Jorge_> microsoft.....isn't that a banned word?
| |
| 12:06 | <cliebow> edubuntuvsMSTS, looks like a wish list to me..
| |
| 12:06 | Jorge_, you just need to slur when you say it
| |
| 12:06 | <Jorge_> cliebow....got you
| |
| 12:07 | <cliebow> the way my dad used to say.."thhhigarettes"
| |
| 12:07 | <edubuntuvsMSTS> meaning?
| |
| 12:07 | <Jorge_> microthoft?
| |
| 12:07 | <cliebow> he was a lung pathologist..didnt like smokers
| |
| 12:08 | ahh..that is good...microsloth
| |
| 12:08 | <edubuntuvsMSTS> OK. I have an opportunity here to plug in an EduBuntu / LTSP solution here.
| |
| 12:08 | <cliebow> let er rip!
| |
| 12:09 | <edubuntuvsMSTS> If, as the question I've asked paints the MSTS in a less than favourable light, I need to know so I can let the pointy haired bosses know how shit and expensive it (MSTS) is!
| |
| 12:10 | So, again: > The following /. story had this in it: "* Remove the crippling in Terminal Server, allow multiuser use over networks. If you can't afford to upgrade all your computers to Vista you can use the old ones as terminals to your Windows Home Server." (Ref: http://slashdot.org/article.pl?sid=08/03/01/1312233 - search for "Microsoft could have done plenty..." post) - what does this mean...
| |
| 12:10 | .../ refer to please?
| |
| 12:11 | <cliebow> msts wont do many msts sessions...ltsp is limited only by your imagination..
| |
| 12:11 | <edubuntuvsMSTS> thank you, please expand on that answer
| |
| 12:11 | <cliebow> there was someone here the other day with a 1000 terminal deployment..
| |
| 12:12 | <edubuntuvsMSTS> (preferably without being superior or glib - I'm already sold on LTSP - it's da management I have to convince!)
| |
| 12:12 | ok
| |
| 12:12 | <cliebow> jammcq should speak..he invented it..
| |
| 12:13 | <edubuntuvsMSTS> Invented what?
| |
| 12:13 | <cliebow> ltsp
| |
| 12:13 | <Jorge_> ok, I got past the build client now.
| |
| 12:13 | <edubuntuvsMSTS> Well, if so, then yes he should!
| |
| 12:13 | <Jorge_> but, the next step is just full of errors
| |
| 12:13 | <cliebow> he is prob wicked busy..
| |
| 12:14 | <Jorge_> could not get lock /ar/lib/dpkg/lock
| |
| 12:14 | <edubuntuvsMSTS> Too busy to quickly help answer a question that will decide whether we go MSTS or EduBuntu ( / LTSP)? Hmm!
| |
| 12:14 | <cliebow> edubuntuvsMSTS, is this install at all dependent on heavy audio/video needing to be pushed across the network
| |
| 12:15 | <jammcq> edubuntuvsMSTS: the thing you'll need to convince them of isn't LTSP vs Terminal servers. It's Linux vs Windows. THAT's the big thing
| |
| 12:15 | <edubuntuvsMSTS> Not really, though there will be a bit of that (not really mission-critical but people will want to do it a bit)
| |
| 12:15 | <cliebow> Jorge_:ps aux|grep apt should show an already running process that has locked apt
| |
| 12:15 | <edubuntuvsMSTS> jammcq - this is a stepping stone ... trust me on this one ok!
| |
| 12:15 | <jammcq> as for "crippling in terminal server", I know nothing of MSTS, so I can't comment on that
| |
| 12:15 | <cliebow> edubuntuvsMSTS, set up a test network..that should help a lot..nowadays it is pretty easy
| |
| 12:16 | <edubuntuvsMSTS> nor me, which is why I'm here asking you as I'd have imagined you'd have sized up the "competition"!
| |
| 12:16 | which is what I am doing ...
| |
| 12:16 | <Jorge_> cliebow....it asks the question "is another process using it"
| |
| 12:16 | <edubuntuvsMSTS> whilst keeping everything live ...
| |
| 12:17 | <cliebow> i showed My boss an old nubus mac running redhat via ltsp
| |
| 12:17 | Jorge_, right..you have an old process locking it..kill that and you should be golden
| |
| 12:17 | <Jorge_> cliebow....hmmm....maybe a reboot? I have no idea how to find that process
| |
| 12:18 | <cliebow> ps aux|grep apt
| |
| 12:18 | prepnd a sudo
| |
| 12:19 | <edubuntuvsMSTS> If, perhaps partially based on the suggestions mentioned in the slashdot post in my original question, I can show that MSTS is shite compared to LTSP, this will be a walk in the park for us ...
| |
| 12:19 | tux_440volt has quit IRC | |
| 12:19 | <Jorge_> cliebow...woohoo
| |
| 12:19 | <cliebow> Jorge_, got it?
| |
| 12:20 | <Jorge_> cliebow...getting there....synaptic was stopping the process
| |
| 12:20 | <edubuntuvsMSTS> try aptitude ...
| |
| 12:20 | :-)
| |
| 12:20 | <cliebow> edubuntuvsMSTS, you'll never find better support than you'll find here...awesome bunch of peole
| |
| 12:20 | moquist has joined #ltsp | |
| 12:20 | <cliebow> errr people
| |
| 12:20 | like that moquist dude
| |
| 12:20 | <edubuntuvsMSTS> but you're not actually answering my question ...
| |
| 12:21 | I am trying very, very hard to push an LTSP solution here
| |
| 12:21 | <Jorge_> cliebow....ok, as I understand it the last step failed......invoke....gives an error unknown initscript
| |
| 12:21 | * cliebow cliebow looks back to clarify thew question | |
| 12:21 | <edubuntuvsMSTS> I am facing a fierce fight from others saying "No, go Microsoft Terminal Services" ...
| |
| 12:22 | <cliebow> hmm..how to show that vista is crap
| |
| 12:22 | <edubuntuvsMSTS> I then come across a weird /. post ...
| |
| 12:22 | <cliebow> how to show that ltsp Rocks
| |
| 12:22 | <Jorge_> cliebow...let them try to install vista, they will learn
| |
| 12:22 | <edubuntuvsMSTS> And I ask here for clarification (ammunition) as to why MSTS is shite!
| |
| 12:23 | <cliebow> nonne of us use the durn thing..
| |
| 12:23 | <edubuntuvsMSTS> The point you're missing is that I am them in the sense that they will trust me to sort it out but I can't stand back and say "Install MSTS / Vista etc. then call me when it breaks"
| |
| 12:24 | No, but given that I am very new to LTSP etc. I was hoping you may have more understanding of your opposition and be able to answer my original question
| |
| 12:25 | and so doing would help us all
| |
| 12:25 | <cliebow> i dont think i can....
| |
| 12:25 | indradg_ has joined #ltsp | |
| 12:25 | <edubuntuvsMSTS> well, can you forward the question to someone who can?
| |
| 12:25 | <jammcq> sure, it would be nice if we had the time and resources to work with all the competition, but the simple truth is, we don't. we built and use LTSP. that's what we're good at.
| |
| 12:26 | <edubuntuvsMSTS> ... or perhaps my question is ill-phrased?
| |
| 12:26 | <cliebow> cant help you then
| |
| 12:26 | <jammcq> you could ask it any way you like. doesn't change the fact that we don't know msts
| |
| 12:26 | <Jorge_> cliebow....any thoughts on my last error?
| |
| 12:26 | <edubuntuvsMSTS> And, I am on the verge on deploying an LTSP solution ...
| |
| 12:27 | <jammcq> good, we applaud that. still, we don't know msts
| |
| 12:27 | <edubuntuvsMSTS> So, where should I look for MSTS criticism?
| |
| 12:27 | <jammcq> umm, google seems to know everything about everything
| |
| 12:27 | <cliebow> Jorge_, what distro you working with?
| |
| 12:27 | <Jorge_> Ubuntu 7.10
| |
| 12:28 | <cliebow> so a fresh install then sudo apt-get install ltsp-server-standalone/
| |
| 12:28 | <Jorge_> ltsp-server
| |
| 12:28 | <edubuntuvsMSTS> FFS! I am in the trenches under heavy fire for even mooting an LTSP solution with all the "Where's my Start Button?" monkeys
| |
| 12:28 | <Jorge_> using the instructions on ubuntu
| |
| 12:29 | <edubuntuvsMSTS> Yeah, it's a piece of piss to install but first you have to win over management (and everyone else!)
| |
| 12:32 | Gutsy!
| |
| 12:34 | <cliebow> first take half an hour and set up a testbed
| |
| 12:34 | Jorge_, i think you will need ogra's help..
| |
| 12:35 | <Jorge_> cliebow.....ok, how do I get that?
| |
| 12:35 | <cliebow> ogra...ohhh..ogra..comeout comeout..
| |
| 12:35 | <Jorge_> cliebow...so close here
| |
| 12:35 | <edubuntuvsMSTS> bbiab
| |
| 12:35 | edubuntuvsMSTS has left #ltsp | |
| 12:37 | <cliebow> Jorge_, i dont thnk i can help...butput dpown what you can see and pastebot.ltsp.org..use the dropdown to pick #ltsp
| |
| 12:39 | <Jorge_> cliebow: thanks for getting me this far, I appreciate the help
| |
| 12:39 | <cliebow> Jorge_, pastebot what oyu have..
| |
| 12:40 | hmmm..just for a guess..is it trying to invoke-rc.d?
| |
| 12:41 | <ltsppbot> "jorge_" pasted "installation woes" (2 lines) at http://pastebot.ltsp.org/455
| |
| 12:42 | <cliebow> ohhhh..ltsp-server does not install dhcp3-server..
| |
| 12:42 | was that just a warning??
| |
| 12:43 | indradg has quit IRC | |
| 12:44 | <Jorge_> cliebow.....no, error as seen
| |
| 12:44 | <cliebow> and then returns you to bash?
| |
| 12:44 | <Jorge_> yes
| |
| 12:45 | <cliebow> i think you are all set then..you need to provide a djcp server capable of passing the filename statement to a client..the rest will just work
| |
| 12:45 | dhcp that is..
| |
| 12:46 | they only keep me around here cause they admire my typing
| |
| 12:46 | <Jorge_> cliebow, I'm not sure I understand you on this step
| |
| 12:47 | <rjune> cliebow: I thought it was that manly beard
| |
| 12:47 | <subsume> Through what means does NIS update /etc/passwd files on client machines?
| |
| 12:47 | <rjune> it does not
| |
| 12:47 | <cliebow> ummm...
| |
| 12:47 | <rjune> the pam module queries NIS/YP directly
| |
| 12:47 | <cliebow> keeps my chin warm..
| |
| 12:47 | <rjune> much the same way PAM will query LDAP
| |
| 12:48 | <cliebow> Jorge_, you understand that a thin-client will query the dhcpserver for a kernel to download via tftp?
| |
| 12:48 | a kernel name?
| |
| 12:49 | <Jorge_> I understand that portion
| |
| 12:50 | <cliebow> so you must have a dhcp server capable of providing that info..it is doubtful a router can
| |
| 12:50 | a home style router..
| |
| 12:51 | <Jorge_> cliebow, so I should have d/l and installed the turnkey version?
| |
| 12:52 | <cliebow> well..all you need to do is sudo apt-get install dhcp3-server (i think) and one of us can give you a dhcpd.conf
| |
| 12:53 | maybe you are better off throwing away /opt/ltsp/i386 and then sudo apt-get install ltsp-server-standalone
| |
| 12:54 | * cliebow cliebow hopes ogra would approve. | |
| 12:55 | <cliebow> assuming you even have one
| |
| 12:58 | <Jorge_> cliebow....let me try the throw away method
| |
| 12:58 | <cliebow> you may not jave even built your chroot../opt/ltsp/i386
| |
| 12:59 | <Jorge_> yes, it did
| |
| 13:00 | so I just delete that dir?
| |
| 13:00 | <cliebow> yep..or set it aside
| |
| 13:00 | <Jorge_> no problem
| |
| 13:04 | foxxbuntu has joined #ltsp | |
| 13:05 | <Jorge_> working............................
| |
| 13:14 | wow, learning so much here
| |
| 13:14 | foxbuntu___ has joined #ltsp | |
| 13:15 | <Jorge_> so, this is linux within linux
| |
| 13:16 | Egyptian[Home1 has joined #ltsp | |
| 13:16 | Egyptian[Home] has quit IRC | |
| 13:17 | <Jorge_> Ok, a bit further, but....invoke-rc.d: initscript dhcp3-server, action "start" failed
| |
| 13:17 | any ideas?
| |
| 13:22 | anyone? I'm so close and I really don't want to give up
| |
| 13:22 | foxbuntu has quit IRC | |
| 13:22 | <laga> Jorge_: check /var/log/daemon.log
| |
| 13:23 | sepski has joined #ltsp | |
| 13:26 | <Jorge_> No subnet declaration for ath0
| |
| 13:26 | Not configured to listen to any interfaces
| |
| 13:28 | <laga> Jorge_: you need to edit /etc/ltsp/dhcp* to suit your network
| |
| 13:31 | <cliebow> Ath0?
| |
| 13:31 | foxxbuntu has quit IRC | |
| 13:31 | <cliebow> Jorge_, what network do you now have?
| |
| 13:34 | <Jorge_> cliebow......no, bad address
| |
| 13:34 | I have a simple network
| |
| 13:35 | 192.168.2.1 and 255.255.255.0
| |
| 13:35 | <cliebow> your dhcpd.conf must be configured in the same network..
| |
| 13:35 | so look at /etc/ltsp/dhcpd.conf
| |
| 13:35 | address of your server is?
| |
| 13:37 | <Jorge_> 192.168.2.5
| |
| 13:37 | <cliebow> k look at subnet declaration in dhcpd.conf
| |
| 13:38 | <Jorge_> ok
| |
| 13:38 | <cliebow> subnet 192.168.2.0 netmask 255.255.255.0 {
| |
| 13:39 | make sure the range also is 192.168.2.*
| |
| 13:39 | <Jorge_> ok, that worked
| |
| 13:39 | <cliebow> dhcp started?
| |
| 13:40 | <Jorge_> Yep!
| |
| 13:41 | Is that a good thing?
| |
| 13:41 | <cliebow> ya can check with netstat -anp|grep ":67"..see that it is listening
| |
| 13:41 | you bet it is a good thing
| |
| 13:41 | <Jorge_> ok, this is getting exciting now
| |
| 13:42 | <cliebow> if you want to be pedantic you can see that tftp is also listening with netstat -anp|grep ":69"
| |
| 13:42 | <Jorge_> Do I need to turn off the dhcp on the router?
| |
| 13:42 | <cliebow> yeas
| |
| 13:42 | as in..yes
| |
| 13:42 | subsume has quit IRC | |
| 13:43 | <Jorge_> ok, done
| |
| 13:43 | now the key is getting a terminal running?
| |
| 13:43 | <cliebow> using pxe or etherboot?
| |
| 13:44 | <Jorge_> either
| |
| 13:44 | <cliebow> try one..
| |
| 13:44 | <Jorge_> i am on a newer one.....compaq R4000 AMD 64 3500
| |
| 13:45 | does it matter?
| |
| 13:45 | <cliebow> no..just getting into the thythm..
| |
| 13:45 | no..just getting into the rhythm..
| |
| 13:45 | Jorge_ has quit IRC | |
| 13:45 | <cliebow> moving to the groovin
| |
| 13:45 | Jorge_ has joined #ltsp | |
| 13:46 | <Jorge_> ok, on the server now
| |
| 13:48 | <cliebow> get off it..you'll break it
| |
| 13:49 | <Jorge_> hmmm...boot from lan did nothing
| |
| 13:50 | <cliebow> did it say pxe..spit out a mac address?
| |
| 13:50 | <Jorge_> yes
| |
| 13:50 | <cliebow> are you using a switch or crossover?
| |
| 13:52 | ill be gone a bit..maybe rjune will keep an eye on you..
| |
| 13:52 | cliebow has quit IRC | |
| 13:54 | <Jorge_> anyone else that can be of kind assistance here?
| |
| 13:57 | bobby_C has joined #ltsp | |
| 14:01 | cliebow has joined #ltsp | |
| 14:05 | <Jorge_> ok, still no luck
| |
| 14:06 | from wired or wireless, it still does not work
| |
| 14:06 | <cliebow> hm...do you have wireshark on server?? forget ahhhhhhhhhhhh
| |
| 14:06 | <Jorge_> yes
| |
| 14:06 | <cliebow> your ath0 is a wireless card?
| |
| 14:07 | <Jorge_> yes
| |
| 14:07 | <cliebow> you must have a wired netowrk card...pxe wont work with wireless
| |
| 14:07 | subsume has joined #ltsp | |
| 14:07 | <laga> cliebow: my prism2 works with pxe ;))
| |
| 14:08 | <cliebow> ohh..cool!
| |
| 14:08 | <laga> cliebow: not very useful because there's no encryption then.
| |
| 14:08 | <Jorge_> I will make it work!
| |
| 14:08 | <cliebow> heh
| |
| 14:08 | <Jorge_> Let's see
| |
| 14:08 | <laga> Jorge_: no, pxe really won't work with wireless devices except for some very special cases
| |
| 14:09 | <cliebow> laga:8~)
| |
| 14:09 | <Jorge_> laga, why not?
| |
| 14:10 | <cliebow> suffice it to say you need to run dhcp on your wired network
| |
| 14:10 | <laga> Jorge_: because the operating system has to log into the network first. if there's not operating system, it's not logged in. you can't really PXE boot when it's not plugged in, right?
| |
| 14:11 | <Jorge_> hmmmmm
| |
| 14:11 | and wired is different because?
| |
| 14:12 | <cliebow> the netwrok card of your client is a wired interface right?
| |
| 14:12 | <laga> Jorge_: because it's already "plugged in".
| |
| 14:13 | <Jorge_> ok, so I have to move my machine.....
| |
| 14:15 | I'm going to make this work!
| |
| 14:15 | <cliebow> or get a longer cable 8~)
| |
| 14:15 | <Jorge_> hahahahhaha
| |
| 14:16 | <cliebow> i dont mean to make fun..
| |
| 14:16 | it is just that things pass thru my head
| |
| 14:16 | <Jorge_> Nope, all is good
| |
| 14:16 | I love levity
| |
| 14:17 | If I could get this working it would be awesome!
| |
| 14:17 | <cliebow> you will..
| |
| 14:17 | <laga> yes, ltsp is awesome
| |
| 14:17 | * laga hugs the plugin interface | |
| 14:17 | <Jorge_> I hope so
| |
| 14:18 | sepski has quit IRC | |
| 14:19 | <Jorge_> Ok, I have one wired machine....maybe I should try it there?
| |
| 14:20 | <cliebow> ffif you have server and a client connected with a crossover..should be ok..or a switch in between with standard patch cables
| |
| 14:20 | <Jorge_> Do you guys have the spare time to help me make this work?
| |
| 14:20 | If not, that is cool.
| |
| 14:20 | <cliebow> someone should be around..
| |
| 14:21 | <Jorge_> ok, I'm moving downstairs
| |
| 14:21 | Jorge_ has quit IRC | |
| 14:22 | Guaraldo has left #ltsp | |
| 14:36 | vagrantc has joined #ltsp | |
| 14:40 | klausade has quit IRC | |
| 14:40 | klausade has joined #ltsp | |
| 14:52 | mhterres has joined #ltsp | |
| 14:52 | mhterres has left #ltsp | |
| 14:58 | * vagrantc contemplates the proposed feature to implement http://bugs.debian.org/459369 | |
| 15:05 | <vagrantc> ok, i think i've got the 3 or so python thingies needed to implement ...
| |
| 15:08 | <warren> vagrantc, got a sec?
| |
| 15:08 | vagrantc, I'm confused about screen_sessions
| |
| 15:08 | <vagrantc> warren: sure
| |
| 15:08 | <warren> oh!
| |
| 15:08 | the thing that calls screen_sessions loops
| |
| 15:08 | this makes sense
| |
| 15:08 | nm
| |
| 15:09 | /etc/lts.conf reads:
| |
| 15:09 | [default]
| |
| 15:09 | SCREEN_02=shell
| |
| 15:09 | SCREEN_03=shell
| |
| 15:09 | Only 02 shell is being run
| |
| 15:10 | not sure what's goin gon
| |
| 15:10 | <vagrantc> hmm... and there's nothing running on tty3 ?
| |
| 15:11 | subsume has quit IRC | |
| 15:12 | <warren> nothing
| |
| 15:13 | hmm... shell might be better done with openvt
| |
| 15:13 | <vagrantc> but why on earth would tty2 work but tty3 not ... what is different?
| |
| 15:13 | <warren> huh....
| |
| 15:13 | screen.d/shell
| |
| 15:13 | echo "tty=" `tty`
| |
| 15:13 | exec /bin/bash --login
| |
| 15:14 | does this make any sense? =)
| |
| 15:14 | why does it echo tty?
| |
| 15:14 | <vagrantc> hm.
| |
| 15:14 | warren: comment it out and see what happens :)
| |
| 15:14 | <warren> oh
| |
| 15:14 | it is only informational
| |
| 15:14 | how does the screen script know to run it on tty2?
| |
| 15:15 | <johnny> hey warren did you see my mention of the dnsmasq post on the list?
| |
| 15:16 | <warren> johnny, yes
| |
| 15:16 | <johnny> would you like it if i somehow got that fixed?
| |
| 15:16 | <warren> johnny, I don't have time to do anything with dnsmasq now though
| |
| 15:16 | johnny, realistically, dnsmasq is not a priority for me
| |
| 15:16 | <johnny> i'll prolly try to get it fixed anyways
| |
| 15:16 | i'll let you know
| |
| 15:17 | <vagrantc> warren: screen_session figures out which tty it's run on ...
| |
| 15:17 | maticue has joined #ltsp | |
| 15:18 | <vagrantc> actually, i can't figure out how it works either
| |
| 15:18 | * vagrantc suspects black magic | |
| 15:18 | <warren> it might be racy
| |
| 15:18 | and it works for most people because they don't define many
| |
| 15:19 | <vagrantc> how many do you have defined?
| |
| 15:19 | <warren> seven
| |
| 15:19 | steph_ has quit IRC | |
| 15:19 | <warren> just to test it
| |
| 15:19 | <vagrantc> i've definitely used 2-3 without problem.
| |
| 15:19 | <warren> I'm beginning to consider not using the SCREEN_XX at all
| |
| 15:19 | <vagrantc> SCREEN_XX has been around for ages.
| |
| 15:19 | steph_ has joined #ltsp | |
| 15:19 | <warren> "been around for ages" doesn't necessarily make it a good thing
| |
| 15:20 | <vagrantc> no, but people will expect it to be there.
| |
| 15:20 | is it such a bad concept?
| |
| 15:20 | or is the implementation merely frustrating?
| |
| 15:20 | <warren> I think the implementation really sucks
| |
| 15:21 | and watching it in "set -x" it reads the lts.conf way too many times
| |
| 15:21 | maybe the concept is OK
| |
| 15:21 | <vagrantc> we could probably just read lts.conf once at boot and cache the results.
| |
| 15:21 | but that's a separate issue.
| |
| 15:21 | <warren> vagrantc, I discovered that it is seemingly impossible to get custom defined dhcp options on the client
| |
| 15:21 | vagrantc, without running dhclient again
| |
| 15:21 | <vagrantc> yeah, that's what i figured.
| |
| 15:22 | <warren> vagrantc, and I can't figure out how to get the IP address of the tftp server
| |
| 15:22 | either within or after initrd
| |
| 15:23 | <vagrantc> hm. i haven't ever had to do anything special with the dhcp client we use ... it just sets a variable and there it is.
| |
| 15:23 | warren: you're setting the next-server option in dhcpd.conf?
| |
| 15:23 | <warren> yes
| |
| 15:23 | it doesn't appear in the leases file though
| |
| 15:23 | <vagrantc> you might have to configure your client to explicitly ask for it
| |
| 15:23 | which client?
| |
| 15:24 | <warren> whatever is embedded in mkinitrd
| |
| 15:24 | it might be based on ISC dhcpv3
| |
| 15:24 | doh!
| |
| 15:24 | X doesn't like an empty xorg.conf
| |
| 15:25 | vagrantc, X doesn't like an empty xorg.conf file, it fails to autoconfigure in such cases
| |
| 15:26 | vagrantc, thus there is no way around putting the tftp grabbed xorg.conf file elsewhere
| |
| 15:30 | <vagrantc> warren: bah.
| |
| 15:30 | <warren> I know, this stinks.
| |
| 15:31 | <vagrantc> warren: just copy the xorg.conf to /etc/X11/xorg.conf ... ??
| |
| 15:31 | <warren> vagrantc, can't
| |
| 15:31 | <vagrantc> warren: because
| |
| 15:31 | <warren> vagrantc, if there is no file on the r/o filesystem you can't bind mount anything on top of it.
| |
| 15:32 | vagrantc, /etc/X11/xorg.conf (blank file), bind mount a read/write file on top, then you can copy on it.
| |
| 15:32 | vagrantc, /etc/X11/xorg.conf blank without anything bind mounted on top, X fails.
| |
| 15:32 | <vagrantc> warren: so always put something in /etc/X11/xorg.conf
| |
| 15:32 | either autogenerate it, or copy it if configured to do so
| |
| 15:33 | <warren> you can't do that.
| |
| 15:33 | <vagrantc> that's exactly what we do.
| |
| 15:33 | debian/ubuntu, that is.
| |
| 15:33 | <warren> vagrantc, X refuses to autogenerate if xorg.conf exists
| |
| 15:33 | <vagrantc> works GREAT for me.
| |
| 15:33 | xorg.conf ALWAYS contains something by the time X starts.
| |
| 15:34 | <warren> huh?
| |
| 15:34 | <vagrantc> either what we've autoconfigured, or the file specified in lts.conf
| |
| 15:34 | <warren> vagrantc, then you aren't using X autogeneration
| |
| 15:34 | <vagrantc> warren: we are using X's autoconfiguration to generate a configuration file.
| |
| 15:34 | warren: look at the configure-x.sh script
| |
| 15:35 | warren: we either run that, or copy the X_CONF file defined in lts.conf
| |
| 15:35 | <warren> hm
| |
| 15:35 | <vagrantc> warren: or some other command that updates /etc/X11/xorg.conf
| |
| 15:35 | <warren> So X needs to run twice in the case where you have no xorg.conf
| |
| 15:35 | <vagrantc> more or less, yes.
| |
| 15:35 | <warren> vagrantc, seriously, what is the problem with having a non-default xorg.conf NOT be called at /etc/X11/xorg.conf?
| |
| 15:36 | vagrantc, or even something like /etc/X11/xorg-external.conf is fine
| |
| 15:36 | as long as it isn't /etc/X11/xorg.conf
| |
| 15:36 | <vagrantc> warren: because that's where xorg.conf is expected to be, and i don't want to special-case LTSP
| |
| 15:36 | <warren> I am NOT going to run X twice just because you are emotional about a config location.
| |
| 15:36 | <vagrantc> pfft.
| |
| 15:37 | <warren> We worked very hard to get X to autoconfigure and just work the first time.
| |
| 15:37 | <vagrantc> warren: that's kind of borderline on insulting... let's drop that kind of talk, please.
| |
| 15:37 | <warren> most of our systems have no /etc/X11/xorg.conf at all
| |
| 15:37 | <vagrantc> warren: how hard would it be to fix xorg to behave like there is no xorg.conf if xorg.conf is empty?
| |
| 15:38 | <warren> vagrantc, it would require a new command line option or something, and we wouldn't be able to add it to old versions.
| |
| 15:38 | vagrantc, how about this compromise:
| |
| 15:38 | * laga suggests a three line shell script | |
| 15:39 | <warren> vagrantc, if /etc/X11/xorg.conf exists, and you have a X_CONFIG, copy it on top of that and use it. If it doesn't exist but you have X_CONFIG, then copy it to elsewhere and use it from there. If it doesn't exist and no X_CONFIG then just run X.
| |
| 15:39 | vagrantc, that should satisfy both of us.
| |
| 15:39 | vagrantc, the first case is really no different than the second case, but if it makes you feel better I'll make it that way.
| |
| 15:39 | <vagrantc> warren: well, then all the screen scripts that call X will need to support xorg.conf being somewhere else.
| |
| 15:40 | <warren> vagrantc, well, that is my problem, not your problem.
| |
| 15:40 | <vagrantc> warren: so it's basically just the way you want it without and compromise.
| |
| 15:40 | * vagrantc suspects ldm will need patching. | |
| 15:41 | <warren> vagrantc, and the X arg finding should be put into a common function, we shouldn't need 3+ instances of this
| |
| 15:41 | vagrantc, I can do that...
| |
| 15:41 | <vagrantc> while i don't have anything at the moment, i suspect there is a better way.
| |
| 15:41 | warren: this is something i would put to the list and give it some time.
| |
| 15:41 | bobby_C has quit IRC | |
| 15:41 | <warren> bottom line: People put LOTS of work into X's autoconfiguration working the first time. You are suggesting running it twice when it isn't necessary.
| |
| 15:42 | vagrantc, my suggested compromise satisfies both of us.
| |
| 15:42 | <vagrantc> warren: i don't like it, i see what your issues are and acknowledge that they are valid, but i don't want to commit to your proposal without some time to come up with something better.
| |
| 15:42 | <warren> vagrantc, the /etc/X11/xorg.conf always exists in your case so the other location codepath would never happen right?
| |
| 15:43 | <vagrantc> warren: yes ... my understanding was X would work with only a partial xorg.conf
| |
| 15:43 | <warren> vagrantc, sort of
| |
| 15:43 | <vagrantc> warren: so maybe there's just like two lines you'd need to add to get autoconfiguration to kick in
| |
| 15:43 | <warren> two lines of what?
| |
| 15:43 | I cannot edit the X server itself
| |
| 15:43 | especially not in older versions
| |
| 15:43 | <vagrantc> warren: so explain your "sort of" response.
| |
| 15:44 | <warren> vagrantc, I'll implement a proof of concept that behaves exactly how you expect.
| |
| 15:44 | vagrantc, I seriously don't have time to argue this, I need a working version in F-9 by the Tuesday freeze. =(
| |
| 15:44 | vagrantc, I understand the behavior you want and I will implement it.
| |
| 15:44 | <vagrantc> warren: i don't know the technical ins and outs of X, but it was my understanding that a partial xorg.conf would allow for undefined parts to be autodetected ... maybe there's just a a few lines you could add that would allow for most of it to be autodetected.
| |
| 15:44 | <warren> I would appreciate a review/critique of the implementation thereafter.
| |
| 15:45 | <vagrantc> warren: while you may be on a time crunch, i definitely don't want to see commits on this to trunk without review.
| |
| 15:45 | <warren> vagrantc, I read a bit of the xorg-server code and that doesn't seem to be the case, the entire autoconfig thing is bypassed if a config file exists.
| |
| 15:45 | <vagrantc> hrm.
| |
| 15:46 | <warren> vagrantc, I have an entire pile of changes already queued depending on this behavior.
| |
| 15:46 | <vagrantc> maybe that was one of those "and it shouldn't be hard to implement this shiny new feature" that someone was talking about
| |
| 15:46 | <warren> vagrantc, is this really a big deal if I don't change the behavior that you expect?
| |
| 15:46 | <vagrantc> warren: it's a fair amount of extra code that i would not want to maintain indefinitely.
| |
| 15:47 | <warren> vagrantc, no it isn't, it is like 2 more lines of shell
| |
| 15:47 | <vagrantc> warren: there's no way to add it to ldm's screen script ... you'll have to modify ldm itself.
| |
| 15:47 | <warren> vagrantc, ldm itself doesn't work for me at all anyway
| |
| 15:47 | vagrantc, not touch it yet
| |
| 15:47 | not touching it yet
| |
| 15:48 | <vagrantc> warren: i would like to review the changes before committing on trunk. if the diff is small and clean, i can't imagine objecting to it.
| |
| 15:48 | <warren> vagrantc, oh, and I already renamed it to xdmcp with a startx symlink
| |
| 15:48 | * vagrantc likes that | |
| 15:49 | <warren> if [ -n "${X_CONF}" ]; then
| |
| 15:49 | if [ -f ${X_CONF} ]; then
| |
| 15:49 | XFCFG=/tmp/xorg.conf
| |
| 15:49 | + [ -e /etc/X11/xorg.conf ] && XFCFG=/etc/X11/xorg.conf
| |
| 15:49 | cp ${X_CONF} ${XFCFG}
| |
| 15:49 | else
| |
| 15:49 | vagrantc, doesn't that satisfy both of us?
| |
| 15:50 | <vagrantc> warren: looks quite reasonable.
| |
| 15:50 | sometimes code is easier than english.
| |
| 15:50 | warren: i didn't realize you had completely abandoned LDM for the moment.
| |
| 15:51 | <warren> vagrantc, the X args finding that I have in screen.d/xdmcp is the new "modernized" version, I think I can split this out into a library so it can be shared with others easily.
| |
| 15:51 | No more two stage X
| |
| 15:51 | <vagrantc> good good.
| |
| 15:52 | <warren> Your choices are 1) use /etc/X11/xorg.conf already in the chroot 2) Use X_CONF 3) Autoconfigure
| |
| 15:52 | <cliebow> someone should be around..
| |
| 15:52 | phew!!
| |
| 15:52 | <vagrantc> warren: what about X_CONFIG_COMMAND ?
| |
| 15:53 | <warren> vagrantc, what is that? that wasn't in screen.d/startx before I began to rip it apart
| |
| 15:53 | <vagrantc> warren: allowing them to specify a command that configures for them ...
| |
| 15:53 | warren: i think it's in the init scripts you don't use
| |
| 15:53 | warren: see, i don't think X configuration belongs in the screen scripts at all.
| |
| 15:53 | <warren> vagrantc, [ -n $X_CONFIG_COMMAND ] && run it?
| |
| 15:54 | vagrantc, agreed, except in the case where you have no xorg.conf at all
| |
| 15:54 | <vagrantc> warren: yeah, something like that ... but i think that code should be in the init scripts ...
| |
| 15:54 | <warren> vagrantc, no xorg.conf is DEFAULT for Fedora
| |
| 15:55 | <vagrantc> warren: oh, one other thing ... instead of /tmp/xorg.conf ... should be something like /var/run/xorg.conf or something ... otherwise you have predictable filenames in /tmp issues
| |
| 15:55 | foxbuntu___ has quit IRC | |
| 15:55 | <warren> mm
| |
| 15:55 | ok. looks good.
| |
| 15:56 | <vagrantc> warren: then you could put the configuration code all in the init scripts ... and the screen script could just check for /var/run/xorg.conf ...
| |
| 15:56 | i think that's a little cleaner ...
| |
| 15:56 | or /var/run/ltsp-xorg.conf or something ... just to avoid potential namespace conflicts
| |
| 15:56 | <warren> ok
| |
| 15:57 | vagrantc, I'll eventually split this X arg finding into a library
| |
| 15:57 | vagrantc, I personally will run it in the service that launches all the screen sessions. I don't really see any need for initscripts.
| |
| 15:57 | <vagrantc> sorry if i seem so fussy about this, while you're in the middle of a time crush ... but, well ... i just like hashing things out properly as much as possible.
| |
| 15:58 | <warren> ah... found an even simpler way
| |
| 15:58 | if /etc/X11/xorg.conf exists, copy X_CONFIG on top of it, but you don't need XFCFG anymore
| |
| 15:58 | because X will use it automatically
| |
| 15:58 | <vagrantc> warren: ok... whatever is more the fedora way is probably best for fedora LTSP5 ... that was part of the point of ltsp5 ... that the ltsp implementation mirrored as much as possible the conventions of the distro it was running on
| |
| 15:59 | <warren> vagrantc, I am trying hard to find common functions and to split them out
| |
| 15:59 | <vagrantc> warren: you still need it if X_CONF is defined.
| |
| 15:59 | warren: and xorg.conf doesn't exist ... don't you?
| |
| 15:59 | <warren> vagrantc, after X_CONF is copied to /etc/X11/xorg.conf ?
| |
| 16:00 | <vagrantc> warren: but if you have no /etc/X11/xorg.conf, and /etc/X11 is read-only.
| |
| 16:00 | <warren> yeah, XFCFG is kept after copying only if /etc/X11/xorg.conf doesn't exist
| |
| 16:00 | <vagrantc> ah.
| |
| 16:01 | warren: i'm now remembering what happened with all of this code ... originally, ogra tried to use it without any xorg.conf ... but then it became difficult to support all the X configuration options in lts.conf... like setting default color depth, resolutions, etc.
| |
| 16:02 | so ogra generated an xorg.conf and then tweaked it with the lts.conf defined values
| |
| 16:02 | <warren> In Fedora if that stuff doesn't have a sane default without xorg.conf then it is a bug.
| |
| 16:02 | Yes I do realize that some people will need an xorg.conf
| |
| 16:02 | and this new way has two ways you can do it
| |
| 16:02 | 1) drop /etc/X11/xorg.conf into the chroot
| |
| 16:02 | <vagrantc> warren: well, it's not that it wasn't a sane default ... it's like you have a class of mis-matched hardware and you want it all to be the same resolution and color depth
| |
| 16:02 | <warren> 2) use X_CONFIG
| |
| 16:03 | * vagrantc notes that it's called X_CONF in ubuntu and debian, and would like to see consistancy across distros | |
| 16:03 | <warren> vagrantc, does your initramfs thing have the ability to pull a different lts.conf for a different MAC address?
| |
| 16:03 | oops
| |
| 16:03 | vagrantc, yeah, X_CONF, just brain thinko
| |
| 16:03 | <vagrantc> warren: no, lts.conf supports mac-address definitions
| |
| 16:04 | <warren> vagrantc, oh
| |
| 16:04 | hah
| |
| 16:04 | <vagrantc> i.e. [XX:11:ad:fo:ba:r0]
| |
| 16:04 | <warren> forgot about that
| |
| 16:04 | * warren rips out MAC-addressed based lts pulling | |
| 16:04 | <vagrantc> which, if you have a *huge* lts.conf is an issue
| |
| 16:05 | * vagrantc would like partial mac matching added ... | |
| 16:05 | <warren> because it is slow?
| |
| 16:05 | ahh
| |
| 16:05 | getltscfg right?
| |
| 16:06 | <jammcq> taking out mac address matching?
| |
| 16:06 | <vagrantc> well, say you've got a lab of 30 machines of one type, and 10 of another, and 5 of another ... you don't want to have 45 separate entries in lts.conf ... just a way so you only need three based on mac address patterns
| |
| 16:06 | <cliebow> yeah!!!!
| |
| 16:06 | <vagrantc> jammcq: no, being able to match against ... [XX:11:ad:fo:*]
| |
| 16:06 | <jammcq> ah
| |
| 16:06 | <warren> jammcq, I wrote initscript based tftp grabbing
| |
| 16:06 | <jammcq> ok
| |
| 16:06 | <warren> which is useless
| |
| 16:06 | <jammcq> had me worried there
| |
| 16:07 | <cliebow> me too
| |
| 16:07 | <vagrantc> well, warren *was* working on it, not realizing that it wasn't needed :)
| |
| 16:07 | <warren> getltscfg could be improved to do globs
| |
| 16:07 | <jammcq> I tried that
| |
| 16:07 | <vagrantc> and?
| |
| 16:07 | <jammcq> I couldn't find a glob library
| |
| 16:08 | <vagrantc> wow
| |
| 16:08 | <jammcq> and ended up running out of time to write my own
| |
| 16:08 | <warren> For something this small, probably best not to pull in an external library
| |
| 16:08 | oh
| |
| 16:08 | <vagrantc> matching against partial ip or hostname would be useful too
| |
| 16:08 | <warren> there's probably something in glibc, which would already be in the chroot
| |
| 16:08 | <johnny> i agree
| |
| 16:08 | i was just thinking the same thing
| |
| 16:08 | <jammcq> well, 6 or 7 years ago, I couldn't find anything
| |
| 16:09 | <vagrantc> once you support partial mac, hostname or ip, i'magine you get all three
| |
| 16:09 | <warren> glibc-headers /usr/include/glob.h
| |
| 16:09 | <jammcq> vagrantc: eharrison and I were sitting in the lucky lab in portland, trying to hack that functionality in
| |
| 16:09 | <vagrantc> jammcq: heh.
| |
| 16:09 | jammcq: i was probably all of 8 blocks from you at that very moment.
| |
| 16:09 | <warren> man 3 glob
| |
| 16:09 | part of glibc and POSIX
| |
| 16:10 | hm... glob is for pathnames =(
| |
| 16:10 | <jammcq> yeah, that's what I was finding
| |
| 16:10 | <vagrantc> C's string manipulation is ... weak.
| |
| 16:10 | <jammcq> but that's the type of pattern matching I'd like to do
| |
| 16:11 | using regex like perl is waaaaay overkill for the task
| |
| 16:11 | <vagrantc> yay!
| |
| 16:11 | <laga> shell?
| |
| 16:11 | <jammcq> simple * and ? would be great
| |
| 16:11 | <vagrantc> while all this discussion was going on, i implemented: /usr/share/ltspfs/mounter.d
| |
| 16:11 | J45p3r has quit IRC | |
| 16:11 | <jammcq> maybe [1-5] too
| |
| 16:12 | a range
| |
| 16:12 | <vagrantc> yeah, definitely.
| |
| 16:15 | <warren> oh
| |
| 16:15 | man 3 regex
| |
| 16:15 | regcomp, regexec, regerror, regfree - POSIX regex functions
| |
| 16:16 | another part of GLIBC
| |
| 16:16 | * jammcq wonders why he has no glibc man pages :( | |
| 16:16 | <warren> By this man page this should totally do the job
| |
| 16:16 | <vagrantc> warren: is this an urgently needed feature? :)
| |
| 16:16 | <warren> vagrantc, you're the one that first mentioned wanting it =)
| |
| 16:17 | vagrantc, my goal is to make auto-config work as much as possible and leave open the possibility of manual configs
| |
| 16:17 | <vagrantc> warren: well, sure. i didn't expect you to drop everything you're doing to make it happen. :)
| |
| 16:17 | warren: that goal is infinitely in tune with my desires.
| |
| 16:17 | <warren> vagrantc, I only went hunting for a glibc library that would do the job
| |
| 16:18 | Patina has quit IRC | |
| 16:18 | Patina has joined #ltsp | |
| 16:20 | <warren> vagrantc, hmm if ldm does the equivalent in python, then I might have to rewrite what I have in xdmcp in python, and have xdmcp call python.
| |
| 16:23 | steph_ has quit IRC | |
| 16:24 | steph_ has joined #ltsp | |
| 16:24 | <maticue> hi, i need help... i dont speak english... somebody speak spanish?
| |
| 16:34 | * warren tries new and improved screen.d/shell | |
| 16:34 | * laga tries out his hacked ltsp-client-builder udeb | |
| 16:37 | <warren> vagrantc, I suspect that non-X screen sessions landing on a certain tty was entirely by chance =)
| |
| 16:40 | vagrantc, screen_session:SCREEN_ARGS=`echo $SCREEN_CMD | cut -f2- -d" " -s`
| |
| 16:40 | vagrantc, I'm not sure this line is doing anything
| |
| 16:57 | <vagrantc> warren: if it's by chance, it's pretty consistantly by chance.
| |
| 16:58 | <warren> vagrantc, have you tried skipping lines? define SCREEN_02=shell SCREEN_04+shell etc.
| |
| 16:58 | <vagrantc> warren: i think it's the exec<$TTY>$TTY 2>&1 bit
| |
| 16:59 | warren: in screen_session
| |
| 17:00 | <warren> gotta go to dinner
| |
| 17:00 | bbl
| |
| 17:00 | vagrantc, uploaded a newer xdmcp script to trunk
| |
| 17:01 | <vagrantc> warren: great.
| |
| 17:01 | <warren> vagrantc, I think the shell script could be improved to use openvt in order to have job control, but I don't know how t make it line up with the SCREEN numbers
| |
| 17:02 | <vagrantc> warren: openvt -c
| |
| 17:02 | ??
| |
| 17:03 | -c vtnumber Use the given VT number and not the first available. Note you must have write access to the supplied VT for this to work.
| |
| 17:04 | <warren> vagrantc, yeah
| |
| 17:04 | vagrantc, but how do you get that # into the screen script?
| |
| 17:04 | <vagrantc> warren: ah ...
| |
| 17:04 | <warren> vagrantc, using it another way openvt works great with: openvt -c # -w -l /bin/bash
| |
| 17:04 | <vagrantc> warren: well, in our case it typically gets run with screen_screen NN
| |
| 17:05 | warren: so you can grab it from the first argument
| |
| 17:05 | at least i think
| |
| 17:06 | i haven't delved into this stuff for so long ... once we got it working, i haven't put a lot of energy into it ...
| |
| 17:06 | good to have new thoughts on everything
| |
| 17:07 | <warren> bbl dnner
| |
| 17:07 | <vagrantc> warren: enjoy
| |
| 17:20 | Q-FUNK has quit IRC | |
| 17:47 | foxxbuntu has joined #ltsp | |
| 18:00 | <vagrantc> need the python equvalent of "sort -u" for a list ...
| |
| 18:03 | <johnny> python set ?
| |
| 18:04 | f = [1, 2, 3, 4, 5, 5, 5, 6]
| |
| 18:04 | list(set(f))
| |
| 18:04 | <vagrantc> johnny: nice!
| |
| 18:05 | <johnny> pre python 2.4.. something like this ..
| |
| 18:05 | x_nodups = dict(map(lambda i: (i,1),x)).keys()
| |
| 18:05 | <vagrantc> "Build an unordered collection."
| |
| 18:06 | <johnny> then you can sort it i imagine..
| |
| 18:06 | i'm not python guru .. yet
| |
| 18:06 | <vagrantc> list.sort()
| |
| 18:06 | <maticue> sorry, can i do a question?. i have ltsp 5 in debian lenny (tarball debian_etch). It's work!!, but dont in all clients. Compaq Clients dont loading the the kernel (TFTP) in theboot. What can i do=
| |
| 18:06 | <johnny> i saw 3 ways in a quick web search
| |
| 18:06 | one was a long def ..
| |
| 18:07 | <vagrantc> maticue: where did you get the tarball from, and why are you using it?
| |
| 18:08 | <maticue> vagrantc, hi!, the tarball is from www.ltsp.org/download and i make a job for my university
| |
| 18:08 | <vagrantc> maticue: don't use that tarball ... it is from before etch was released.
| |
| 18:08 | <maticue> y make a class with compaq i
| |
| 18:08 | compa pentium i
| |
| 18:08 | what tarball is good?
| |
| 18:08 | <vagrantc> maticue: if you're using a lenny server, why not build a lenny chroot ?
| |
| 18:09 | maticue: no need to use a tarball, install ltsp-server and run ltsp-build-client
| |
| 18:09 | maticue: although lenny is slightly broken right now, you'd be better off with sid.
| |
| 18:09 | <maticue> becouse its more easy with tarball... but i wasnt?? know this problem
| |
| 18:11 | <vagrantc> maticue: the tarball is REALLY broken. don't use it.
| |
| 18:11 | <maticue> vagrantc, ok!, thanks.. the floopy boot? or the bios of the terminal?
| |
| 18:11 | <vagrantc> maticue: ??
| |
| 18:11 | <maticue> vagrantc, you speak spanish?
| |
| 18:11 | <vagrantc> maticue: un poco
| |
| 18:12 | <maticue> vagrantc, gracias,
| |
| 18:12 | vagrantc, voy a intentar con ltsp-build-client. pero el floppy para bootear el cliente no tiene algo que ver?
| |
| 18:14 | vagrantc, por que probe el tarball de etch y me funcionaba en unas pentium III. el problema estaba con el protocolo TFTP, por eso imagine que el problema podria estar en el cliente.
| |
| 18:14 | <laga> i think we had someone else in here the other day who couldn't boot his compaq client
| |
| 18:15 | <vagrantc> maticue: cual vercion de PXE en el cliente con problemas?
| |
| 18:17 | <maticue> vagrantc, ahi me fijo. el cliente bootea con un floppy que tiene un menu azul con varias opciones
| |
| 18:17 | <vagrantc> maticue: cual floppy ?
| |
| 18:17 | <maticue> vagrantc, la version exacta del floppy de booteo no la se, ahora estoy buscando
| |
| 18:17 | <vagrantc> maticue: etherboot ?
| |
| 18:17 | <maticue> vagrantc, para bootear yo uso el BootDisk522b
| |
| 18:17 | ebnet522.dsk
| |
| 18:19 | vagrantc, existe alguna otra? q pueda probar y que sea tan buena como esa?
| |
| 18:20 | <vagrantc> !bootfloppy
| |
| 18:20 | <ltspbot> vagrantc: "bootfloppy" is http://etherboot.anadex.de or ftp://k12linux.mesd.k12.or.us/pub/K12LTSP/Universal_boot_floppy
| |
| 18:20 | <vagrantc> maticue: uso etherboot.anadex.de ...
| |
| 18:21 | <maticue> si si!, tambien probe con esta
| |
| 18:21 | <vagrantc> entiendo ... 75% :)
| |
| 18:21 | <maticue> vagrantc, sorry.... yes, i try with this version too, but i have the some problem. the tftp protocol dont work
| |
| 18:22 | <vagrantc> maticue: hmmm...
| |
| 18:22 | maticue: pero otro clientes funciona?
| |
| 18:22 | <maticue> we have a 33 compaqs pentium i!! ;(
| |
| 18:23 | en en unas pentium III con 64mb ram funciona
| |
| 18:23 | y muy bien funciona!
| |
| 18:23 | <cliebow> laga:ive had weired responses from a lot of compaq..both in pxe and etherboot...
| |
| 18:23 | <maticue> am i the problem?? jaja
| |
| 18:23 | <vagrantc> maticue: cuantos ram tiene los compaqs?
| |
| 18:24 | <maticue> 32 mb
| |
| 18:25 | <vagrantc> es possible, pero necesita el nueva vercion de LTSP ... el vercion de etch necesita mucho ram.
| |
| 18:25 | <maticue> Compaq Presario 2240 Desktop PC
| |
| 18:26 | <vagrantc> pero, el ram no es la problema con TFTP ...
| |
| 18:26 | <maticue> pero tambien lo probe en una compaq con 512mb de ram y tampoco funciono
| |
| 18:27 | i don't know what to do... :(
| |
| 18:27 | because we have a lot of compaqs
| |
| 18:28 | i thinking in the client boot, but you know more than me
| |
| 18:30 | <vagrantc> maticue: where exactly is it breaking ? downloading the kernel ?
| |
| 18:32 | <maticue> | |
| 18:33 | vagrantc, el reloj de carga se mueve muy muy poco. la deje mas de 30 minutos esperando que cargue y no funciona
| |
| 18:45 | ok, thans any way. i try to find the solution and tell you. bye!
| |
| 18:45 | maticue has quit IRC | |
| 19:03 | <vagrantc> warren: you're setting a root shell for 02-06 on all your clients by default?
| |
| 19:04 | mcfloppy has quit IRC | |
| 19:07 | <vagrantc> warren: isn't that a bit of a security compromise?
| |
| 19:09 | <laga> that's odd. my client boots fine if i boot it over PXE. if i copy the kernel to a USB pen drive and pass nbdroot=10.0.10.1 ip=dhcp as boot options, it gets "connection refused" from nbdrootd
| |
| 19:10 | <vagrantc> laga: nbdport= ??
| |
| 19:11 | <laga> vagrantc: nbdport should be autodetected (eg it defaults to 2000). but you're right, that can be the problem
| |
| 19:13 | <vagrantc> laga: check /var/log/syslog on the server
| |
| 19:13 | <laga> um. it's connecting to port 10.
| |
| 19:13 | vagrantc: syslog is empty
| |
| 19:13 | um
| |
| 19:13 | sorry, i was in a chroot :)
| |
| 19:13 | <vagrantc> heh
| |
| 19:16 | <laga> works with nbdroot=10.0.10.1:2000 now
| |
| 19:42 | joebaker has joined #ltsp | |
| 19:44 | japie213 has joined #ltsp | |
| 19:44 | <japie213> Hi, can I lock/resume a session?
| |
| 19:45 | I would like to keep some X programs running overnight on the X server. Is that possible?
| |
| 20:05 | johnny_ has joined #ltsp | |
| 20:18 | cliebow has quit IRC | |
| 20:23 | K_O-Gnom has quit IRC | |
| 20:24 | <joebaker> japie213: For things like that I like to use NoMachine's NX technologies. They have an open source version called freenx.
| |
| 20:24 | <japie213> joebaker, I tried NX. Too many problems.
| |
| 20:25 | <joebaker> The other way I've accomplished something similar is to use vncserver spawned desktops.
| |
| 20:26 | Otherwise, I don't know of any other way to do what you're suggesting.
| |
| 20:26 | I'd love to learn of anything you come up with that meets your requirements... joebaker@dcresearch.com
| |
| 20:33 | cyberorg has quit IRC | |
| 20:35 | MacIver has quit IRC | |
| 20:36 | <rjune> You could run VNC as a service
| |
| 20:36 | the desktop always runs, you just connect to it
| |
| 20:37 | <joebaker> Hi rjune
| |
| 20:37 | japie213 has quit IRC | |
| 20:38 | <joebaker> You did a lot of work on LTSP sound as I recall.
| |
| 20:38 | <rjune> once upon a time
| |
| 20:50 | foxbuntu___ has joined #ltsp | |
| 20:53 | chupacabra has quit IRC | |
| 20:53 | chupacabra has joined #ltsp | |
| 20:54 | J45p3r has joined #ltsp | |
| 20:55 | steph_ has quit IRC | |
| 20:55 | johnny_ has quit IRC | |
| 20:59 | vagrantc has quit IRC | |
| 20:59 | spectra has quit IRC | |
| 21:04 | MacIver has joined #ltsp | |
| 21:08 | foxxbuntu has quit IRC | |
| 21:10 | foxxbuntu has joined #ltsp | |
| 21:13 | MacIver has quit IRC | |
| 21:16 | MacIver has joined #ltsp | |
| 21:27 | stgraber has quit IRC | |
| 21:29 | foxbuntu___ has quit IRC | |
| 21:40 | J45p3r has quit IRC | |
| 21:52 | ogra_cmpc has quit IRC | |
| 21:53 | ogra has quit IRC | |
| 21:53 | ogra_cmpc has joined #ltsp | |
| 21:53 | ogra has joined #ltsp | |
| 21:54 | MacIver has quit IRC | |
| 21:58 | MacIver has joined #ltsp | |
| 21:59 | stgraber has joined #ltsp | |
| 22:39 | foxbuntu___ has joined #ltsp | |
| 22:54 | foxxbuntu has quit IRC | |
| 23:38 | chup has joined #ltsp | |
| 23:40 | chupacabra has quit IRC | |