00:00 | <sahil> sbalneav:still x on the screen
| |
00:00 | <sbalneav> netstat -an | grep ":177"
| |
00:01 | Do you get anything?
| |
00:02 | <sahil> quite a bit
| |
00:02 | <sbalneav> More than one line?
| |
00:02 | <sahil> must've typed it wrong
| |
00:02 | just one line
| |
00:02 | <sbalneav> paste it please
| |
00:03 | <sahil> ha i am im working on laptop and server
| |
00:03 | so theres a little delay
| |
00:03 | udp6 0 0 :::177 :::*
| |
00:03 | doesn't say anything about xdmcp
| |
00:03 | <sbalneav> udp6?
| |
00:03 | <sahil> no idea
| |
00:04 | <sbalneav> That doesn't seem right. What does your /opt/ltsp/i386/etc/lts.conf look like?
| |
00:04 | <sahil> under xdmcp login window preferences it has listen on udp port: 177
| |
00:05 | i got it
| |
00:05 | we had an interface problem earlier
| |
00:05 | it put that ip all over the place
| |
00:06 | the server line had the wrong ip
| |
00:07 | <sbalneav> So does it boot now?
| |
00:07 | <sahil> waiting to find out
| |
00:07 | good to go
| |
00:07 | except mouse no go but ill get that
| |
00:11 | /dev/input/mice or /dev/psaux
| |
00:16 | <sbalneav> usually psaux
| |
00:16 | Late here, I'm heading to bed.
| |
00:16 | 'Night.
| |
00:16 | sbalneav has quit IRC | |
00:48 | kaos01 has quit IRC | |
00:48 | PMantis has left #LTSP | |
00:49 | kaos01 has joined #ltsp | |
00:49 | <kaos01> hi, whats the correct way to ensure that clients have the correct time ?
| |
00:50 | <f3ew> ntpd
| |
00:50 | <kaos01> mmm
| |
00:50 | are you sure
| |
00:50 | ntd clients are not instoled by default
| |
00:50 | and what about timezones
| |
00:51 | Burgundavia has quit IRC | |
00:52 | <kaos01> f3ew, have you implemented ntpd on your clients ?
| |
00:54 | * f3ew isn't running LTSP yet | |
00:54 | <f3ew> but I will be doing that
| |
00:55 | <kaos01> haha
| |
00:56 | woudnt one need to run something liek yum update tz-data but install it in /opt/ltsp/whatever
| |
00:58 | sahil has quit IRC | |
00:58 | McDutchie has joined #ltsp | |
01:05 | Burgundavia has joined #ltsp | |
01:14 | MoL0ToV has joined #ltsp | |
01:29 | Burgundavia has quit IRC | |
01:56 | daya has joined #ltsp | |
02:13 | <daya> hi to all,
| |
02:22 | Jenna has quit IRC | |
03:15 | McDutchie has quit IRC | |
03:16 | yopla has joined #ltsp | |
03:23 | Zenton has joined #ltsp | |
03:25 | gonzaloaf has quit IRC | |
03:32 | sepski has joined #ltsp | |
04:02 | lambda_ has joined #ltsp | |
04:03 | ogra-classmate has joined #ltsp | |
04:21 | f3ew has quit IRC | |
04:21 | f3ew has joined #ltsp | |
04:26 | envite has joined #ltsp | |
04:26 | <envite> Hello all from Spain
| |
04:26 | Hola a todos desde España
| |
04:36 | <Nikopoika> hi, i would like to use "ldm" displaymanager in my ltsp-server, but i'm not sure how to make it a default manager, my default manager atm is gdm
| |
04:39 | <ogra-classmate> Nikopoika: ldm isnt built for that, currently you can only use it on thin clients
| |
04:44 | bill_c has quit IRC | |
04:45 | bill_c has joined #ltsp | |
04:48 | jsgotangco has joined #ltsp | |
04:53 | <daya> do i have to run ltsp-build-client in debian etch for installing ltsp,
| |
04:58 | <ogra-classmate> !debian
| |
04:58 | <ltspbot`> ogra-classmate: "debian" is is a GNU/Linux based operating system that makes an excellent LTSP server. You can find it at http://www.debian.org. for information about LTSP on debian see http://wiki.debian.org/LTSP
| |
05:08 | <Nikopoika> ogra-classmate: ok
| |
05:11 | <Zenton> hi all,
| |
05:11 | nx is not right now integrated in ltsp, isnt it?
| |
05:21 | <highvoltage> nope, it isn't
| |
05:23 | <Zenton> thanks highvoltage.
| |
05:26 | <highvoltage> you're welcome.
| |
05:28 | jsgotangco has quit IRC | |
05:38 | daya has quit IRC | |
05:51 | Jenna has joined #ltsp | |
05:57 | J45p3r has joined #ltsp | |
06:03 | gonzaloaf has joined #ltsp | |
06:06 | nepali has joined #ltsp | |
06:06 | <kaos01> how does one set the correct timeon the clients ?
| |
06:10 | <envite> clients need correct time?
| |
06:11 | I think not
| |
06:12 | yopla has quit IRC | |
06:16 | <kaos01> well why is there a /etc/localtime than ?
| |
06:16 | i realise there are no benefits in it :)
| |
06:17 | how does one start up sshd on the client
| |
06:19 | <sepski> kaos01, for ltsp5 i just install ssh inside the chroot, and reboot client.
| |
06:20 | <kaos01> iam using 4.2
| |
06:25 | vagrantc has joined #ltsp | |
06:29 | <envite> kaos01: there is an option in lts.conf for ssh to start
| |
06:30 | but it is not named ssh nor sshd
| |
06:30 | let me see
| |
06:31 | http://wiki.ltsp.org/twiki/bin/view/Ltsp/SSH
| |
06:38 | nf1 has quit IRC | |
06:39 | <vagrantc> so ... who's familiar with ltspswapd ? was it basically a fork of nbd-server with some features added?
| |
06:44 | <kaos01> ok, so LOCAL_APPS = Y probably starts up sshd
| |
06:44 | <envite> drop the "probably" ;)
| |
06:44 | Jenna has quit IRC | |
07:02 | J45p3r has quit IRC | |
07:11 | damnation has joined #LTSP | |
07:11 | <damnation> hey here
| |
07:11 | is someone able to help me with a connection problem ?
| |
07:11 | <vagrantc> if you describe the problem it will be much easier to help with :)
| |
07:12 | usually, just go ahead an ask your question, and people will respond if they can
| |
07:13 | <damnation> :-) when i boot up the package after the nfs script i get connect : connection refused and then keeps saying read: connection refused
| |
07:13 | i lookd at host. host.allow
| |
07:13 | if gdm is enabled
| |
07:13 | i lookd at export file
| |
07:13 | but they all look good to me off course , i'm a newby :p
| |
07:13 | <vagrantc> what distro ?
| |
07:14 | <damnation> where do i find the distro
| |
07:14 | i mean how do i know wich one it is
| |
07:14 | <vagrantc> what linux distribution do you have installed?
| |
07:14 | <damnation> debian
| |
07:14 | edge
| |
07:15 | <vagrantc> lsb_release -a
| |
07:15 | <damnation> the one i posted the debian edge :-)
| |
07:15 | <cyberorg> "connection refused" sshd running on the server?
| |
07:16 | <vagrantc> well, that's particularly confiusing, in that there is debian etch, and ubuntu edgy ... so an answer that looks like a mix of both is very confusing :P
| |
07:16 | <damnation> it is now let me test :-)
| |
07:16 | <cyberorg> oh, if it is nfs try mounting the exported folder from another machine
| |
07:17 | check the path exported from /etc/exports exists and points to the right place
| |
07:17 | <vagrantc> damnation: how did you install ltsp ?
| |
07:18 | <damnation> i did apt-get install openssh-server ltsp-server-standalone
| |
07:18 | <vagrantc> and then?
| |
07:18 | <damnation> then i installed , the ltsp-utils to create the files
| |
07:18 | <vagrantc> ltsp-utils is incompatible with ltsp-server
| |
07:18 | <damnation> and i did a manual configure of the dhcp setting , host , export and host.allow
| |
07:19 | i did not install the package of the ltsp-utils
| |
07:19 | i only used it to see if all service's where running
| |
07:19 | <vagrantc> ok.
| |
07:20 | so, what did you do next?
| |
07:20 | <damnation> first i'am gonne check now if the export points to the right place
| |
07:20 | brb
| |
07:20 | <vagrantc> well ... i'm wondering if you ever actually built the LTSP environment ...
| |
07:21 | damnation: dpkg --root=/opt/ltsp/i386 -l ltsp-client
| |
07:21 | <damnation> i did build the client
| |
07:21 | the path is correct
| |
07:22 | doe's it hurt to rebuild the ltsp client just to make sure ?
| |
07:22 | <vagrantc> it can cause problems, yes.
| |
07:22 | i'd recommend: mv /opt/ltsp/i386 /opt/ltsp/i386.old before rebuilding
| |
07:23 | <cyberorg> also check the root path in /etc/dhcpd.conf if it matches the exports too
| |
07:24 | should be something like this: option root-path "/opt/ltsp/i386/";
| |
07:24 | <damnation> o.0 i dont have a root path in dhcp.conf
| |
07:25 | <vagrantc> that'll do it
| |
07:25 | damnation: take a look at /usr/share/doc/ltsp-server/examples/dhcpd.conf
| |
07:25 | f3ew has quit IRC | |
07:26 | f3ew has joined #ltsp | |
07:27 | <damnation> ok i now hava :
| |
07:27 | host ws001 {
| |
07:27 | ah damn
| |
07:27 | sec
| |
07:28 | http://rafb.net/p/Dg7Ome29.html
| |
07:28 | that is my dhcp.conf
| |
07:29 | still the same error
| |
07:31 | <cyberorg> try mount serverip:/opt/ltsp/i386 /mnt from another machine
| |
07:31 | move option along with other options on top, not sure if that makes any difference
| |
07:32 | and is this the nfs server too: 192.168.0.160
| |
07:32 | <damnation> yes it is
| |
07:32 | to wich other machine
| |
07:33 | i only have my laptop
| |
07:33 | its a skool project
| |
07:33 | <vagrantc> damnation: is the NFS server the same as the DHCP server?
| |
07:33 | <damnation> i suppose so yes there both running on the same machine
| |
07:33 | <vagrantc> yes.
| |
07:34 | <cyberorg> try mounting /opt/ltsp/i386 on your laptop
| |
07:34 | <vagrantc> damnation: so ... what happens when you: mount -o nolock,ro 192.168.0.1:/opt/ltsp/i386 /mnt
| |
07:35 | <cyberorg> vagrantc: 192.168.0.160:/opt/ltsp/i386 ?
| |
07:35 | <damnation> server is down
| |
07:35 | <vagrantc> cyberorg: ?
| |
07:36 | <damnation> i get the message mount to nfs server '192.1668.0.1" failed : server is down
| |
07:36 | <cyberorg> next-server is 192.168.0.160
| |
07:36 | <vagrantc> ah, yes.
| |
07:36 | sorry.
| |
07:36 | <cyberorg> so i assume nfsserver, dhcp and tftp all is running on that machine
| |
07:37 | * vagrantc goes back to working on the next release | |
07:37 | <cyberorg> :)
| |
07:37 | <vagrantc> cyberorg: i'll leave it to you :)
| |
07:37 | nepali has quit IRC | |
07:37 | <damnation> thanks vagrantc for the help
| |
07:37 | i did a mount to 160 no errors
| |
07:38 | <cyberorg> move the options with all the other options and restart dhcpd and boot up the client
| |
07:39 | <damnation> when i open the ltspadmin and see if all my service's are running there all in the green
| |
07:40 | restart of dhcp . boot up client
| |
07:40 | same error
| |
07:40 | connection refused
| |
07:41 | <cyberorg> what does tail -f /var/log/messages on the server say
| |
07:41 | most likely client is trying to mount wrong thing
| |
07:42 | <damnation> http://rafb.net/p/f0A15380.html
| |
07:42 | i pasted it to the html page
| |
07:42 | a part of the last
| |
07:43 | suppose it is indeed trying to mount the wrong package
| |
07:43 | where do i get a new package . mount something else ?
| |
07:45 | <cyberorg> i dont see any messages about dhcp or nfsserver, check those messages
| |
07:45 | <damnation> debian dhcdbd: message_handler: message handler not found under /com/redhat/dhcp/eth0 for sub-path eth0.dbus.get.reason
| |
07:45 | lambda_ has left #ltsp | |
07:47 | <cyberorg> gotto go too, check the logs for nfs requests
| |
07:47 | <damnation> lk
| |
07:47 | ok thanks for the help
| |
07:50 | nf1 has joined #ltsp | |
07:54 | <vagrantc> damnation: you might want to check /var/log/syslog instead of /var/log/messages
| |
07:55 | <damnation> :-)
| |
07:56 | all i see is the dhcp requests that the client did to the server
| |
07:57 | is it possible that the ssh server is denying the client ?
| |
07:58 | how do you ppl set up a ltsp server and client ?
| |
07:59 | is it with the lstp-utils or server ?
| |
08:00 | <jammcq> ltsp-utils is for the older ltsp-4.2
| |
08:00 | are you trying to setup ltsp-5 ?
| |
08:00 | <damnation> i dont know , it the ltsp-server-standalone ltsp 5 ?
| |
08:01 | <jammcq> yes
| |
08:01 | <damnation> then yes
| |
08:01 | <jammcq> so you shouldn't be using ltsp-utils at all
| |
08:01 | <damnation> nope , i just use it to see if all my service's are running
| |
08:01 | <jammcq> k
| |
08:01 | <damnation> i did not install the package
| |
08:01 | <jammcq> so what problem are you having?
| |
08:02 | <damnation> the client boots up but after the nfs script
| |
08:02 | i get connection refused
| |
08:02 | <jammcq> at what point do you get that error?
| |
08:02 | during NFS, or has NFS completed ?
| |
08:02 | <damnation> the nfs script is completed it says done
| |
08:03 | <jammcq> what's the next thing it says?
| |
08:03 | <damnation> then connection : connetion refused and then keeps repeating
| |
08:03 | read : connetion refused
| |
08:03 | <jammcq> "connection refused" keeps repeating?
| |
08:03 | <damnation> yes
| |
08:03 | <jammcq> or NFS keeps repeating?
| |
08:04 | <damnation> ill type it down
| |
08:04 | begin: running /scripts/nfs-premount ....
| |
08:04 | Done.
| |
08:05 | connect: Connection refused
| |
08:05 | and then it keeps repeating
| |
08:05 | <jammcq> ok, so that's just the nfs-premount
| |
08:05 | nfs hasn't finished yet
| |
08:05 | <damnation> read: connection reused
| |
08:05 | <jammcq> ok, do this:
| |
08:05 | showmount -e
| |
08:05 | J45p3r has joined #ltsp | |
08:06 | <damnation> opt/ltsp *
| |
08:06 | <jammcq> opt or /opt ?
| |
08:06 | <damnation> d slash opt
| |
08:06 | <jammcq> ok, do this: grep root-path /etc/ltsp/dhcpd.conf
| |
08:07 | <damnation> no such file
| |
08:07 | envite has quit IRC | |
08:07 | <damnation> wait
| |
08:07 | <jammcq> ok, try this: grep root-path /etc/dhcp3/dhcpd.conf
| |
08:08 | <damnation> option root-path ".opt/ltsp/i386
| |
08:08 | <jammcq> "." ?
| |
08:08 | <damnation> never mind that
| |
08:08 | typo
| |
08:08 | <jammcq> so that looks right
| |
08:08 | <damnation> :-)
| |
08:08 | <vagrantc> presuming there aren't more typos :P
| |
08:08 | <jammcq> heh
| |
08:09 | <damnation> | |
08:09 | <vagrantc> damnation: why not just cut-and-paste? :P
| |
08:09 | <damnation> its running in vmware
| |
08:09 | <jammcq> oh jeez
| |
08:09 | <damnation> lol , soz still a student
| |
08:10 | ill fire up putty
| |
08:11 | <jammcq> you running this on a windows host ?
| |
08:11 | <damnation> yes (blush)
| |
08:11 | <vagrantc> the plot thickens
| |
08:11 | * jammcq goes and washes his hands really really well | |
08:11 | <damnation> lol
| |
08:12 | <vagrantc> damnation: when people are helping you troubleshoot problems... it's usually good to explain anything weird you might be doing :P
| |
08:12 | <damnation> but understand i will change in the future lol , its all the fault of the skool
| |
08:13 | ok soz , ill do so in the future i just didn't think vmware whould be the problem
| |
08:14 | anything else i can do ?
| |
08:14 | <jammcq> do you have a real thin client? or is that virtual too?
| |
08:14 | <damnation> its a virtual to
| |
08:14 | <vagrantc> damnation: no, but it's worth mentioning
| |
08:15 | <damnation> its a dos client with a pxe
| |
08:18 | ccherret1 has joined #ltsp | |
08:18 | <damnation> so what happens now ? i get a gun and shoot myself for using the evil OS ?
| |
08:18 | :)
| |
08:19 | <jammcq> umm, that wouldn't be enough :)
| |
08:19 | can you paste your dhcpd.conf file to the ltsp pastebot ?
| |
08:19 | <damnation> lol
| |
08:19 | <jammcq> !pastebot
| |
08:19 | <ltspbot`> jammcq: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
| |
08:20 | <ogra-classmate> hmm, how can he have ltsp-server-standlone installed but there is no /etc/ltsp/dhcpd.conf ?
| |
08:20 | thats not possible
| |
08:21 | <jammcq> umm
| |
08:21 | <damnation> i do have it
| |
08:21 | its the same text
| |
08:21 | i forgot to type a d
| |
08:21 | <ltsppbot> "damnation" pasted "authoritative; subnet 192.168." (22 lines) at http://pastebot.ltsp.org/144
| |
08:22 | <ogra-classmate> (03:05:07 PM) jammcq: ok, do this: grep root-path /etc/ltsp/dhcpd.conf
| |
08:22 | (03:05:38 PM) damnation: no such file
| |
08:23 | <damnation> option root-path "opt/ltsp/i386";
| |
08:23 | <jammcq> that's not the typical ltsp-5 dhcpd conf file
| |
08:23 | <ogra-classmate> filename "pxelinux.0"; ??
| |
08:23 | <jammcq> but it should work
| |
08:23 | <ogra-classmate> thats surely wrong
| |
08:23 | <damnation> yes i boxed it together myself
| |
08:23 | with reading from the net
| |
08:24 | <jammcq> do this: grep tftp /etc/inetd.conf
| |
08:24 | <ogra-classmate> the default file shipped by ltsp-server-standalone is already preconfigured to run out of the box if your client interface has an ip from the 192.168.0.0 net
| |
08:24 | you shouldnt even touch it
| |
08:24 | <damnation> tftp dgram udp wait root /usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /var/lib/tftpboot
| |
08:24 | <jammcq> have you moved pxelinux.0 ?
| |
08:25 | cuz your current setup shouldn't be working
| |
08:25 | <ogra-classmate> so that paste was the /etc/ltsp/dhcpd.conf ?
| |
08:25 | <damnation> its my first setup , yes i moved the pxelinux to the tftp boot
| |
08:25 | <jammcq> oye
| |
08:25 | <ogra-classmate> thats set up automatically by ltsp-build-client, you shouldnt need to touch it either
| |
08:25 | <damnation> i also made the default with the mac file , and made the file
| |
08:25 | <jammcq> ok, lets check your network config. run this:
| |
08:26 | ifconfig -a
| |
08:26 | and paste the output to the pastebot
| |
08:26 | <ogra-classmate> so what you pasted to the pastebot was */etc/ltsp/dhcpd.conf* ??
| |
08:26 | or any other file ?
| |
08:26 | <ltsppbot> "damnation" pasted "eth0 Link encap:Ethernet HWadd" (25 lines) at http://pastebot.ltsp.org/145
| |
08:27 | <damnation> no i pasted the dhcp3 dhcp.conf
| |
08:27 | <jammcq> damnation: do you have a /etc/ltsp/dhcpd.conf file?
| |
08:27 | <ogra-classmate> that file is ignored if /etc/ltsp/dhcpd.conf exists
| |
08:27 | <jammcq> btw, your eth0 looks reasonable
| |
08:27 | <vagrantc> ogra-classmate: this is on debian
| |
08:27 | ogra-classmate: i don't think it's ignored.
| |
08:27 | <ogra-classmate> vagrantc: and you hacked that ?
| |
08:28 | ccherrett has quit IRC | |
08:28 | * jammcq wonders which nfs-server package was installed | |
08:28 | <ogra-classmate> vagrantc: afaik mdz made that change in debian
| |
08:28 | <vagrantc> ogra-classmate: maybe i misunderstood your "ignored" comment
| |
08:28 | <damnation> how do i look it up
| |
08:28 | <ogra-classmate> jammcq: ltsp-server-standalone should depend on the right one
| |
08:28 | <vagrantc> ogra-classmate: mdz never made any changes in debian
| |
08:29 | ogra-classmate: it's only been pere, myself, otavio and stratus who've actually worked on debian ltsp directly
| |
08:29 | <jammcq> ogra-classmate: is that true of debian too ?
| |
08:30 | <vagrantc> ltsp-server only depends on nfs-kernel-server
| |
08:30 | <ogra-classmate> vagrantc:
| |
08:30 | if [ -f /etc/ltsp/dhcpd.conf ]; then
| |
08:30 | CONFIG_FILE=/etc/ltsp/dhcpd.conf
| |
08:30 | fi
| |
08:30 | * vagrantc wonders where that code is | |
08:30 | <ogra-classmate> mdz was the dhcp3-server maintainer at that time in debian and i think he made the change there
| |
08:30 | the initscript
| |
08:30 | <damnation> the output in the ltsp/dhcpd.conf is diffrent
| |
08:31 | i know it does'nt use that one in the ltsp dir
| |
08:31 | <vagrantc> ah.
| |
08:31 | <ogra-classmate> thats badly broken
| |
08:31 | <damnation> becouse i doesn't work if i remove the dhcp3 /dhcp
| |
08:31 | <vagrantc> ogra-classmate: nope.
| |
08:31 | ogra-classmate: code's not in dhcp3-server init script
| |
08:31 | <jammcq> damnation: I know you are excited, and I know you are in a hurry, but you REALLY need to type correctly. 'ltsp/dhcpd.conf' doesn't tell us whether you mean '/etc/ltsp/dhcpd.conf' or something else. PLEASE be complete, and correct in your typing
| |
08:31 | <ogra-classmate> meh
| |
08:32 | <vagrantc> better than typing correctly, cut and paste
| |
08:32 | <ogra-classmate> vagrantc: you should document that difference
| |
08:32 | (or better fix it)
| |
08:32 | <vagrantc> ogra-classmate: that seems like insane behavior to me ...
| |
08:32 | <damnation> i will sorry
| |
08:33 | <ogra-classmate> vagrantc: thats default behavior since day one and mentioned in the docs
| |
08:33 | <vagrantc> ogra-classmate: i've never seen any of these docs.
| |
08:33 | <ogra-classmate> this way you dont br4eak existing configs and give the admin stikll the option to copy it over for his ltsp server
| |
08:34 | <vagrantc> ogra-classmate: does it still read /etc/dhcp3/dhcpd.conf ?
| |
08:34 | <jammcq> ah, but you do break existing config. install ltsp-standalone-server, and reboot your server, and presto..... your existing dhcpd behaviour is now broken
| |
08:34 | <vagrantc> ogra-classmate: because when people edit that file, they sure expect it to modify dhcp3-server's behavior.
| |
08:34 | <ogra-classmate> vagrantc: well
| |
08:34 | dhcp3 (3.0.2-1ubuntu2) breezy; urgency=low
| |
08:34 | * Allow ltsp to override the default dhcpd.conf
| |
08:34 | -- Matt Zimmerman <mdz@ubuntu.com> Wed, 10 Aug 2005 12:45:53 -0700
| |
08:35 | <vagrantc> s,Allow,Force,
| |
08:35 | <ogra-classmate> we use it since two years like that
| |
08:36 | and its mentioned on many wikipages on the edubuntu and ubuntu wikis
| |
08:36 | <vagrantc> no wonder i haven't seen it
| |
08:36 | <ogra-classmate> as well as on the ltsp wiki i think (NOT SURE ABOUT THAT)
| |
08:36 | OOPS
| |
08:36 | * jammcq doesn't see a point in arguing about it now. it is what it is | |
08:36 | <ogra-classmate> right
| |
08:37 | but that we have qa difference in behavior here is bad
| |
08:37 | <jammcq> i'm not crazy about how that works, but it's there
| |
08:37 | <vagrantc> and it *won't ever be there* in debian, i assure you.
| |
08:37 | <ogra-classmate> well, i'm very happy we dont break esxisting configs due to it
| |
08:37 | <vagrantc> that would be an RC bug in debian.
| |
08:37 | <jammcq> even if you do break them :)
| |
08:38 | <ogra-classmate> jammcq: i dont break the config
| |
08:38 | <jammcq> umm, yeah
| |
08:38 | <ogra-classmate> the old *config* persists
| |
08:38 | <vagrantc> if you ignore a working configuration file, it's totally broken from the user's perspective
| |
08:38 | <ogra-classmate> ;)
| |
08:38 | <vagrantc> oh, yes, you only break the server, not the configuration file :P
| |
08:38 | <jammcq> some poor slob had a perfectly working dhcpd, then installed ltsp-server-standalone, and now it's broken
| |
08:38 | <ogra-classmate> vagrantc: so i should sed through it and inject my values to get automation going?
| |
08:39 | <damnation> oh look a birdy :p
| |
08:39 | <jammcq> heh
| |
08:39 | <ogra-classmate> jammcq: sudo cp /etc/dhcp3-server/dhcpd.conf /etc/ltsp
| |
08:39 | :P
| |
08:39 | <vagrantc> which then breaks ltsp-server-standalone's configuration
| |
08:39 | <jammcq> ogra-classmate: sure, that works, after the guy got fired, for breaking his entire network
| |
08:39 | <ogra-classmate> vagrantc: why ?
| |
08:40 | ccherret1 has quit IRC | |
08:40 | J45p3r has quit IRC | |
08:40 | <jammcq> poor guy's on the unemployment line now
| |
08:40 | <vagrantc> ogra-classmate: because ltsp-server-standalone's dhcp feature no longer functions, and you'll have ugly upgrade questions about replacing the file
| |
08:40 | <damnation> and i did not even start working sig
| |
08:41 | <ogra-classmate> jammcq: sorry for him, really, bit there is no doc on the ubuntu wiki that is ltsp and dhcp related that doesnt talk about that file
| |
08:41 | <jammcq> damnation: ummmmmm, yours *should* be working
| |
08:41 | ccherrett has joined #ltsp | |
08:41 | <ogra-classmate> vagrantc: its no conffile
| |
08:42 | you wont get upgrade questions
| |
08:42 | <damnation> ok lol but it isn't , what todo .? fresh install then
| |
08:42 | <jammcq> dhcpd.conf is not a conf file?
| |
08:42 | <ogra-classmate> jammcq: not for dpkg
| |
08:42 | <vagrantc> jammcq: do not confuse "conf file" with "configuration file"
| |
08:42 | <jammcq> wow
| |
08:42 | my bad
| |
08:42 | <ogra-classmate> jammcq: you have to mnention files to dpkg while packaging so it monitors if they are changed
| |
08:43 | <jammcq> I guess any idiot with half a brain should know that a "conf" file isn't a "configuration" file :)
| |
08:43 | <vagrantc> ogra-classmate: i'm pretty sure it's a conf file on debian.
| |
08:43 | <ogra-classmate> we explicitly dont want that for this file
| |
08:43 | Guaraldo has joined #ltsp | |
08:43 | <jammcq> maybe "conf" stands for "Confusion"
| |
08:43 | <vagrantc> jammcq: it's a debian-ism. i don't look down on your for not knowing this obscure bit of policy
| |
08:43 | <ogra-classmate> heh
| |
08:44 | * jammcq is getting hungry, what's for breakfast? | |
08:44 | * jammcq is still wishing he had some of that wonderful Sevilla Orange Juice | |
08:44 | <damnation> | |
08:44 | <ogra-classmate> get the machine ;)
| |
08:44 | <jammcq> it's not the machine, it's the Oranges :)
| |
08:45 | <ogra-classmate> and some orange trees perhaps :)
| |
08:45 | <jammcq> yeah, those would last about 1 week, here in Michigan
| |
08:45 | <damnation> anyway's thank's for the help ppl i appriciate it , maybe ill come back laters
| |
08:45 | <ogra-classmate> well, wait one, two years, global warming will help :)
| |
08:45 | <jammcq> ogra-classmate: yeah, I guess someone will be a winner with global warming
| |
08:46 | maybe we should just "look at the bright side" of global warming :)
| |
08:46 | <ogra-classmate> our year is pretty exactly one month ahead here
| |
08:46 | <vagrantc> global warming, while technically correct, is a misleading term.
| |
08:46 | many places will likely get very cold.
| |
08:46 | <jammcq> our year seems behind. it's been rather cold
| |
08:46 | <ogra-classmate> yeah
| |
08:46 | <vagrantc> or unpredictable
| |
08:47 | <ogra-classmate> i had my first strawberries from the garden yesterday
| |
08:47 | and they were damned sweet
| |
08:48 | <vagrantc> ogra-classmate: ok, you second-guessed me on some debian-ness the other day. now's my turn. you *sure* /etc/ltsp/dhcpd.conf is not a "conf file" on ubuntu?
| |
08:48 | <ogra-classmate> it wasnt one ... if it got one thats a merge mistake on my side
| |
08:49 | <vagrantc> unless there are significant differences in the debhelper utilities, i suspect it is a conf file ...
| |
08:50 | i made some nice progress yesterday ... now i forget where i was going with it ...
| |
08:50 | think i'm about ready to split ltsp-client into two packages
| |
08:51 | <ogra-classmate> good
| |
08:52 | MoL0ToV^ has joined #ltsp | |
08:52 | <vagrantc> also dropped support for XFree86 yesterday, finally :)
| |
08:52 | <jammcq> wow, cool
| |
08:52 | <vagrantc> hopefully nobody still uses it
| |
08:52 | there's probably more places where we could drop support ...
| |
08:53 | <damnation> hey , the nfsroot=/opt/ltsp/i386 | is this correct to where the nfsroot is ?
| |
08:54 | <vagrantc> damnation: ?
| |
08:54 | damnation: could you re-phrase the question, perhaps with more context
| |
08:54 | <damnation> the default fine in pxelinux.cfg has a line append initrd=initrd.img root=/dev/nfs ip=dhcp nfsroot=/opt/ltsp/i386
| |
08:55 | and i was wondering if the root to the nfs was correct
| |
08:55 | fine = file
| |
08:55 | as it is a nfs problem
| |
08:55 | <vagrantc> DEFAULT vmlinuz ro initrd=initrd.img root=/dev/nfs ip=dhcp
| |
08:56 | is all i see in /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
| |
08:56 | damnation: which file are you talking about?
| |
08:56 | <damnation> under var/lib/tftboot i have the folder pxelinux.cfg with in it a default file
| |
08:57 | <vagrantc> damnation: grep filename /etc/*/dhcpd.conf
| |
08:58 | <ltsppbot> "damnation" pasted "/etc/dhcp3/dhcpd.conf: filenam" (4 lines) at http://pastebot.ltsp.org/146
| |
08:59 | <vagrantc> damnation: unless you have your tftp server configured very weird, it's probably not looking in /var/lib/tftpboot/pxelinux.cfg
| |
09:00 | MoL0ToV has quit IRC | |
09:00 | <damnation> now i get nfsmount: need a path
| |
09:00 | ill figure it out thanks
| |
09:00 | <vagrantc> damnation: grep root-path /etc/*/dhcpd.conf
| |
09:00 | <damnation> i know where to look now
| |
09:01 | :-)
| |
09:03 | nicoAMG has joined #ltsp | |
09:08 | MoL0ToV^ has quit IRC | |
09:17 | pablo__ has joined #ltsp | |
09:19 | <pablo__> hi! ive installed ltsp5 using ubuntu 7.04. when clients boot gdm apears but when users login screen goes black and after a while gdm appears again
| |
09:22 | jsgotangco has joined #ltsp | |
09:27 | C1ph3r has joined #ltsp | |
09:32 | sepski has quit IRC | |
09:36 | sbalneav has joined #ltsp | |
09:37 | <sbalneav> Morning all!
| |
09:39 | <ogra-classmate> !s
| |
09:39 | <ltspbot`> ogra-classmate: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
| |
09:39 | <sbalneav> wheee
| |
09:40 | yeah, last night I managed to get it to log in. So, now I have to write the support routines to detect the expire (a bunch of read() with select() etc etc stuff)
| |
09:40 | <jammcq> !s
| |
09:40 | <ltspbot`> jammcq: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
| |
09:41 | <sbalneav> We'll have to make mods to the greeters to handle a "your password has expired, please enter a new password, oh you mutton head you didn't enter the same password correctly etc etc etc.)
| |
09:41 | Morning jammcq
| |
09:42 | <ogra-classmate> sbalneav: i think i got a bit further with the kernel ... i still need to prove it, but i think we'll gain a lot by using a different scheduler in the kernel
| |
09:42 | <sbalneav> What scheduler is used now?
| |
09:42 | <ogra-classmate> on the classmate using elevator=anticipatory as kernel option nearly doubles the responsiveness of the whole thing
| |
09:43 | <sbalneav> I prefer elevator=musak :)
| |
09:43 | <ogra-classmate> cfq is currently used
| |
09:43 | <jammcq> I prefer escalator
| |
09:44 | <ogra-classmate> `i tried with the deadline scheduler, thats speeds up even more but hangs the system after a while for some seconds
| |
09:44 | <vagrantc> sbalneav: do you know if the expiry code is localized?
| |
09:44 | sbalneav: that could really mess things up
| |
09:45 | <sbalneav> Nope, I have no idea.
| |
09:45 | <ogra-classmate> we'll get it locdalized if it isnt
| |
09:46 | <sbalneav> I've got no clue as how to handle that on my end.
| |
09:46 | <vagrantc> ogra-classmate: yes, but if sbalneav is reading the output to get a response ...
| |
09:46 | <ogra-classmate> but regarding who maintains ssh i would assume it is
| |
09:46 | pablo_ has quit IRC | |
09:46 | <ogra-classmate> colin is very cautious about transaltions
| |
09:47 | <vagrantc> it will be very hard to handle the ssh login and expiry code in LDM if the responses are localized
| |
09:48 | <ogra-classmate> we could force a LANG=C worst case
| |
09:49 | i guess
| |
09:49 | <vagrantc> and then switch back somehow?
| |
09:49 | Gadi has joined #ltsp | |
09:49 | Gadi has left #ltsp | |
09:49 | Gadi has joined #ltsp | |
09:51 | <ogra-classmate> vagrantc: well, LANG=C ssh -X user@server 'LAQNG=$LDM_LANG Xsession'
| |
09:51 | like that
| |
09:51 | <vagrantc> yeah
| |
09:51 | <ogra-classmate> its ugly but works
| |
09:53 | <Gadi> hey, guys
| |
09:54 | <sbalneav> ergh.
| |
09:54 | ssh -l goober localhost "LANG=C /bin/bash --login"
| |
09:54 | goober@localhost's password:
| |
09:54 | Permission denied, please try again.
| |
09:54 | goober@localhost's password:
| |
09:54 | WARNING: Your password has expired.
| |
09:54 | <jammcq> Gadi: good morning
| |
09:54 | <sbalneav> Password change required but no TTY available.
| |
09:54 | Morning Gadi
| |
09:54 | <Gadi> seems we all had the same idea for this Monday ;)
| |
09:55 | <vagrantc> ?
| |
09:55 | <sbalneav> ah
| |
09:55 | ssh -t -l goober localhost "LANG=C /bin/bash --login"
| |
09:55 | goober@localhost's password:
| |
09:55 | WARNING: Your password has expired.
| |
09:55 | You must change your password now and login again!
| |
09:55 | Changing password for goober
| |
09:55 | (current) UNIX password:
| |
09:56 | ogra-classmate: can you set up an expired account on your box, and verify that works for another language?
| |
09:57 | jsgotangco has quit IRC | |
09:57 | <sbalneav> shadow file should look like:
| |
09:57 | goober:Eou/BHDN/g.Tw:13000:0:99:7:::
| |
09:57 | the 1300 and 99 are the important bits.
| |
09:58 | <ogra-classmate> sbalneav: not atm, my classmate only knows english
| |
09:59 | <sbalneav> hm
| |
09:59 | nf1 has quit IRC | |
10:00 | <sbalneav> need to set some box up with a lang other then eng
| |
10:00 | nf1 has joined #ltsp | |
10:03 | pablo__ has quit IRC | |
10:22 | irule has joined #ltsp | |
10:22 | <vagrantc> sbalneav: you can practice your french :)
| |
10:23 | <sbalneav> Mais oui!
| |
10:23 | <vagrantc> i could try and set up a spanish install from qemu
| |
10:24 | nf1 has quit IRC | |
10:24 | * vagrantc boots a virtual thin client with the new "ltsp-client-core" package installed | |
10:25 | <ogra-classmate> Filesystem Size Used Avail Use% Mounted on
| |
10:25 | unionfs 278M 215M 63M 78% /
| |
10:25 | no nchance for me
| |
10:30 | <vagrantc> sbalneav: working on testing it on an es_ES system
| |
10:31 | <sbalneav> cool
| |
10:31 | <vagrantc> the other project i work on, simple-cdd, makes it really easy to make custom installer CDs and test them in qemu
| |
10:32 | <sbalneav> Another one that would be nice to figure out would be ssh -t -l foo remote "LANG=C PS1=LTSPLOGINENDED /bin/bash --login"
| |
10:32 | <vagrantc> and ssh'ing to a reasonably fast machine in portland makes it possible for me to actually run it. :)
| |
10:32 | <sbalneav> the PS1 seems to get overriden
| |
10:32 | <vagrantc> LTSPLOGINENDED ?
| |
10:33 | <sbalneav> That way you could search for that string, and know that you're successfully logged in and sitting at a prompt.
| |
10:33 | hmm
| |
10:33 | echo LTSPLOGINENDED ; /bin/bash ?
| |
10:33 | yep
| |
10:33 | that'll work.
| |
10:34 | <vagrantc> what's wrong with LTSP_CLIENT=ip.ad.dre.ss
| |
10:34 | <sbalneav> oin$ ssh -t -l goober localhost "LANG=C echo FOOO ; /bin/bash"
| |
10:34 | goober@localhost's password:
| |
10:34 | FOOO
| |
10:34 | goober@oin:~$
| |
10:34 | Well, you're going to do that anyway.
| |
10:34 | what you're wanting is when you script the login, you have some definite way of knowing you've successfully logged in.
| |
10:34 | <vagrantc> so the password isn't being taken from stdin ?
| |
10:35 | right.
| |
10:35 | <sbalneav> Password's being taken from the greeter, and passeed to the ssh's stdin
| |
10:48 | damnation has quit IRC | |
10:52 | <ogra-classmate> heh, that mail to ltsp-developer is an ubuntu feature request list :)
| |
10:52 | <jammcq> yeah, I saw that too
| |
10:52 | not very LTSP specific
| |
10:53 | <ogra-classmate> yup
| |
10:53 | meduxa has joined #ltsp | |
11:01 | * vagrantc can't wait to see it | |
11:03 | C1ph3r has quit IRC | |
11:21 | spectra has joined #ltsp | |
11:29 | <vagrantc> sbalneav: well, my first try suggests that ssh and password expiry don't localize ... or only partially so
| |
11:31 | Retype new UNIX password:
| |
11:31 | passwd: contraseña actualizada correctamente
| |
11:31 | Connection to localhost closed.
| |
11:32 | * vagrantc notices that debian etch doesn't have python-gnomecanvas | |
11:32 | <vagrantc> nor sid
| |
11:32 | introduced in ubuntu edgy
| |
11:34 | <Gadi> stupid question: are you trying to gather all of the input first and then "expect" it to the CLI?
| |
11:34 | <vagrantc> Gadi: that seems to be the basic jist
| |
11:35 | <Gadi> perhaps you could just gui-fy the entire interaction and pass-through the prompts untouched
| |
11:35 | * vagrantc doesn't like the idea of password handling code being poorly tested | |
11:35 | <Gadi> then, its just a matter of the GUI supporting the fonts
| |
11:35 | <vagrantc> like, put it all in an xterm ?
| |
11:35 | <Gadi> no
| |
11:36 | just dont gather the entire interaction for a "preseeding"
| |
11:36 | <ogra-classmate> vagrantc: its in python-gnome
| |
11:36 | <Gadi> but, do the whole process interactively
| |
11:36 | <ogra-classmate> ubuntu has split it out for xubuntu
| |
11:36 | <vagrantc> ogra-classmate: python-gnome2
| |
11:36 | <ogra-classmate> yeah
| |
11:36 | we have split it because you can use gnomecanvas without the rest of gnome
| |
11:37 | and xubuntu makes some use of it
| |
11:37 | <vagrantc> ogra-classmate: i think we used to do python-gnomecanvas | python-gnome2 ... but i think i dropped it thinking it was a sarge-ism
| |
11:37 | <ogra-classmate> i think debian discussed the same split though
| |
11:37 | * vagrantc would like to see it | |
11:37 | <ogra-classmate> its a whille ago
| |
11:38 | but hey, we'll get rid of gnomecanvas anyway :)
| |
11:38 | <vagrantc> sounds like it
| |
11:38 | i'm just trying as fast as i can to get some new packages into debian
| |
11:38 | <ogra-classmate> backports
| |
11:38 | or sid
| |
11:38 | ?
| |
11:38 | <vagrantc> sid
| |
11:38 | <ogra-classmate> ah
| |
11:39 | right, i need a gutsy matching one as well
| |
11:39 | <vagrantc> trying to bring debian and ubuntu closer in sync
| |
11:39 | <ogra-classmate> else the first cd will break
| |
11:39 | <vagrantc> and implement some of these features i've long been putting off
| |
11:39 | <ogra-classmate> yeah
| |
11:40 | <sbalneav> Hm, yeah, passing the prompts through.
| |
11:40 | <ogra-classmate> i will hold back from ltsp apart from ldm and concentrate on the gui parts and speed issues for now
| |
11:40 | <sbalneav> We'll have to re-think the greeter a bit.
| |
11:40 | <ogra-classmate> feel free
| |
11:41 | <sbalneav> I always do! :)
| |
11:41 | <Gadi> actually, we already have the GUI done
| |
11:41 | <ogra-classmate> so the feedback comes directly from ssh to teh greeter instead of ansynchronous as before ?
| |
11:41 | <Gadi> just need to modify the ldm script a bit
| |
11:42 | right
| |
11:42 | <vagrantc> Gadi: this is part of the C re-write
| |
11:42 | <ogra-classmate> Gadi: there is no script anymore :)
| |
11:42 | <Gadi> what?
| |
11:42 | man
| |
11:42 | why?
| |
11:42 | <ogra-classmate> all C
| |
11:42 | speed
| |
11:42 | <sbalneav> Gadi: I've already started on a C rewrite :)
| |
11:42 | <vagrantc> the complaint that trumps all other complaints
| |
11:42 | <Gadi> there's really no speed hit because of the ldm script is there?
| |
11:42 | <ogra-classmate> sbalneav: so you need additional user input ?
| |
11:43 | <sbalneav> Gadi: Apparently, 20 seconds.
| |
11:43 | <ogra-classmate> which the current greeter doesnt provide?
| |
11:43 | <Gadi> what?!
| |
11:43 | <sbalneav> On slow hardware
| |
11:43 | <Gadi> because of /usr/sbin/ldm?
| |
11:43 | and not the greeter?
| |
11:43 | <sbalneav> yep.
| |
11:43 | <Gadi> huh
| |
11:43 | <sbalneav> Hard to beleive.
| |
11:43 | <ogra-classmate> the greeter was never reqally slow, just way to big
| |
11:44 | <Gadi> well, I have a slow greeter, but I thought it was all the greeter part
| |
11:44 | <sbalneav> But I guess python's a busy little bee, and uses lots of cucles.
| |
11:44 | err cycles.
| |
11:44 | <Gadi> huh
| |
11:44 | <sbalneav> so, since all the interested parties are here in the room... I think...
| |
11:44 | jammcq: ping
| |
11:44 | <ogra-classmate> sbalneav: so what do you want to change in the greeter?
| |
11:44 | <vagrantc> i think it's more related to how multiple instances of the same code get duplicated
| |
11:44 | <sbalneav> Well, lets hash this out, see what we need.
| |
11:45 | ok, here's the situation, as I see it.
| |
11:45 | <Gadi> well, at least in the py script, I would just pass through the prompts from ssh to change the "Username" text
| |
11:45 | <ogra-classmate> if its only additionalinput you only need in case the password is expired i'd prefer a separate gtk app that just pops in front
| |
11:45 | <sbalneav> I like gadi's idea of the setting the display=workstationip to get around the ssh, and what with some of the loadbalancing stuff we want in, I think ldm's the way to go.
| |
11:46 | * vagrantc waits for the situation as sbalneav sees it | |
11:46 | damnation has joined #LTSP | |
11:46 | <damnation> hi there
| |
11:46 | <sbalneav> However, we have a problem with password expiry.
| |
11:46 | <ogra-classmate> well,we have lots more
| |
11:46 | like smartcards
| |
11:46 | and other rather rare pam setups
| |
11:47 | <sbalneav> More specifically, an expired password isn't handled by the LOCAL ssh, it just passes through text from the REMOTE end, allowing the user to change the password.
| |
11:47 | <ogra-classmate> there are many cases where you need user interaction
| |
11:47 | <damnation> hey how do i fix this , port mapper failure -rpc: unable to receive
| |
11:47 | <sbalneav> ogra-classmate: yeah, but to fix those, we need local auth.
| |
11:47 | <ogra-classmate> shudder
| |
11:47 | <Gadi> not if it is all a matter of ssh auth
| |
11:47 | <sbalneav> if we had local auth NOW, we could fix this easy. Just have LDM do a pam-session.
| |
11:47 | <ogra-classmate> Gadi: its rather pam and ssh together
| |
11:47 | <sbalneav> however, in the short term, we need to do this.
| |
11:48 | <ogra-classmate> the combo makes it hard
| |
11:48 | <sbalneav> Here's an example of the problem.
| |
11:48 | <ogra-classmate> sbalneav: if we can do pam sessions we can use gdm
| |
11:48 | then ldm could go
| |
11:49 | <sbalneav> ogra-classmate: Well, there will still be some benefit to LDM, such as load balancing, but you're right.
| |
11:49 | <ogra-classmate> not having pam auth on the thin clients was 6the first reason to write ldm
| |
11:49 | <sbalneav> anyway, just to let people try, and check my math, make sure I'm not smoking crack, try this:
| |
11:49 | 1) Set up a user account locally, call it "goober". Make it expired.
| |
11:50 | 2) ssh -t -X -l goober localhost "xterm", will ask for a password, but then fail out with a "need to change password, but no controlling tty" message.
| |
11:51 | So, we can't kick off the xsession in the initial ssh
| |
11:51 | My solution to this is:
| |
11:51 | 3) ssh -t -M -S /tmp/foosocket -l goober localhos "/bin/bash --login"
| |
11:51 | <ogra-classmate> ++
| |
11:52 | <sbalneav> this gets us a bash shell, and allows us to script the password exchange.
| |
11:52 | Then we've got a command socket set up, which we can use to launch the Xsession, etc/
| |
11:52 | <vagrantc> what about xterm -e ssh -t -M -S /tmp/foosocket -l goober localhost
| |
11:53 | <ogra-classmate> and then ?
| |
11:53 | <Gadi> yuck
| |
11:53 | <sbalneav> Then, from there, we're plumbed and ready to go.
| |
11:53 | <vagrantc> well, to handle the password expiry
| |
11:54 | <ogra-classmate> shudder
| |
11:54 | nah
| |
11:54 | lets have a proper gui
| |
11:54 | <sbalneav> Now, the 99% use case is, we simply have to pass the password, and we're in.
| |
11:54 | <Gadi> all we need to do is Guify the exchange
| |
11:54 | <ogra-classmate> right
| |
11:54 | <vagrantc> that's what i'm thinking, too ... a quick and dirty way to do so.
| |
11:54 | <Gadi> doesnt matter how many prompts
| |
11:54 | <sbalneav> However, some small percentage of the time, we need to say, "oops, this didn;t work, we need a new password"
| |
11:54 | <ogra-classmate> and since its a rather rare occurence i'd plumb in an external gtk app
| |
11:54 | <sbalneav> Now, we can handle this a couuple of ways.
| |
11:55 | <ogra-classmate> so you dont need all the stuff in ram even if the password is fine
| |
11:56 | <sbalneav> 1) icky way: simple greeter, and if we need another new password, launch something that grabs it, checks it twice, naughty or nice, etc. And passes it to the ssh.
| |
11:56 | <Gadi> (side note: can you even expire a password in Users and Groups? )
| |
11:56 | <ogra-classmate> yeah that would just be another popup win in the code
| |
11:56 | <sbalneav> 2) second would be to have the greeter have some kind of command socket that we could pass it commands on the fly, like "ask login", "ask password" "print this error", "ask new password" etc.
| |
11:56 | <ogra-classmate> Gadi: you can set a lifetime iirc
| |
11:57 | sbalneav: ++
| |
11:57 | <sbalneav> So the greeter itself has a stdin/stdout, that you interact with it via.
| |
11:57 | <ogra-classmate> that sounds sane
| |
11:57 | we can need it for other stuff as well
| |
11:58 | <Gadi> right, but try not to rely on parsing the ssh prompts
| |
11:58 | thats what gets you into localization hell
| |
11:58 | <sbalneav> Well, we might not be able to get away from doing some of that
| |
11:58 | So the solution we came up with this morning is:
| |
11:58 | ssh -t....... "LANG=C /bin/bash --login"
| |
11:59 | <Gadi> just treat everything as an input box
| |
11:59 | <ogra-classmate> Gadi: new ldm is supposed to be 100% localized ion all aspects
| |
11:59 | <Gadi> right
| |
11:59 | <sbalneav> Force the language to english. Gross, but workable.
| |
11:59 | <Gadi> thats what im trying to achieve for you
| |
11:59 | <ogra-classmate> under the hood we should use english and have proper .po files for teh translators
| |
11:59 | <Gadi> if ssh expects input, put up an input box
| |
11:59 | <ogra-classmate> right
| |
11:59 | <sbalneav> Right, but here's the problem:
| |
12:00 | <ogra-classmate> but with our own transklation
| |
12:00 | <sbalneav> Launch ssh
| |
12:00 | read input "should be 'Password:'"
| |
12:00 | Now, we either get a shell, OR
| |
12:00 | something else happens.
| |
12:00 | I.e. "Your password has expired...."
| |
12:01 | If we're to know that we need to prompt for further input, we have to read what ssh has given us, and parse it.
| |
12:01 | <ogra-classmate> right
| |
12:02 | <sbalneav> So, I think we can't realistically get away from doing a bit of parsing.
| |
12:02 | <Gadi> well, it seems to me like ssh will do one of two things:
| |
12:02 | * ogra-classmate doesnt see the problem | |
12:02 | <sbalneav> However, we should be able to keep it to one or two words.
| |
12:02 | like "expired"
| |
12:02 | or "incorrect"
| |
12:02 | <Gadi> wait for stdin or end process
| |
12:02 | <ogra-classmate> just dio a regex patternmatch
| |
12:02 | whats your problem ?
| |
12:02 | <Gadi> can we not detect if it is waiting for input or the process has ended?
| |
12:03 | <sbalneav> The problem is, the process won't end.
| |
12:03 | <ogra-classmate> which is fine
| |
12:03 | <Gadi> ok ok
| |
12:03 | no problems
| |
12:03 | <sbalneav> We want to keep that command socket open, so once we log in we want it to sit there.
| |
12:03 | Gadi: We're not beating up on you :)
| |
12:03 | it's a tough problem is all.
| |
12:03 | <ogra-classmate> and start subsequent processes like gnome on it
| |
12:03 | <Gadi> just wondering if we could get away from maintaining translation tables
| |
12:04 | <ogra-classmate> sbalneav: whats tough about it ?
| |
12:04 | <sbalneav> Well, conceptionally it's not tough. Coding, it's a lot of crank-turning :)
| |
12:04 | <ogra-classmate> look if we get anything back from ssh ... if so, make a patternmatch ... if it con6tains "expired" leunch the matching tool
| |
12:04 | <sbalneav> right.
| |
12:05 | So, does everyone seem to think my approach seems, if not perfect, sane?
| |
12:05 | <ogra-classmate> if the tool has ended run the session scripts through the socket ....
| |
12:05 | its sexy imho
| |
12:06 | <sbalneav> ogra-classmate: right.
| |
12:06 | <Gadi> yeah - its not a bad solution at all
| |
12:06 | the most work will be the translating ;)
| |
12:06 | <ogra-classmate> which we dont have to do :)
| |
12:06 | * ogra-classmate cheers for launchpad | |
12:06 | <Gadi> it would prolly be wise to peek into the ssh source and look for responses
| |
12:07 | <ogra-classmate> right
| |
12:07 | <Gadi> its always nice to have a complete tool
| |
12:07 | :)
| |
12:07 | <sbalneav> What we really SHOULD do is pull the .po's from ssh :)
| |
12:07 | <Gadi> heh, that too
| |
12:07 | lns has quit IRC | |
12:07 | <sbalneav> Because we don't want to use one word for "expired" in one place, and another in another.
| |
12:08 | i.e. it's translated as "expired" in one place and "too old" in the other.
| |
12:08 | <Gadi> oh, and I wouldnt worry about smartcards until we address them directly - my hunch is no local pam required
| |
12:08 | <sbalneav> ok, I need a few more days "rough" coding to get SOMETHING going, at that point (I'm hoping by wed or thur) I'll be able to bzr push something for people to look at.
| |
12:09 | <ogra-classmate> Gadi: well, we need to handle them through ssh which wasnt possible with the old ldm design
| |
12:09 | <Gadi> why not?
| |
12:09 | <ogra-classmate> you need pam handling on the server side but there is communication involved we couldnt do yet
| |
12:10 | <sbalneav> ok, I gotta go for my lunchtime walk.
| |
12:10 | be back later
| |
12:10 | <ogra-classmate> with the socket we now have an open communication socket to attach to
| |
12:10 | * ogra-classmate has to go as well for 1h | |
12:10 | <Gadi> k
| |
12:19 | <damnation> someone able to help
| |
12:19 | my portmapper isn't allowing outgoing connections
| |
12:20 | or rather its unable to receive
| |
12:26 | damnation has quit IRC | |
12:30 | maaarek has joined #ltsp | |
12:31 | cyberorg has quit IRC | |
12:39 | eurialo has joined #ltsp | |
12:49 | <vagrantc> so what i hear is maintaining our own copy of the set of translations we're going to interact with? that's crazy.
| |
12:50 | J45p3r has joined #ltsp | |
12:50 | <Gadi> vagrantc: I think we can take from ssh
| |
12:50 | <sbalneav> vagrantc: I'm open to suggestions.
| |
12:50 | Back, BTW
| |
12:51 | <Gadi> hey, guys: how come we don't use unionfs on the client (but ltsp-client-setup makes /dev/root unionfs) ?
| |
12:51 | what am I missing?
| |
12:52 | <sbalneav> Where would we use unionfs?
| |
12:52 | In localdevs?
| |
12:52 | <Gadi> in general
| |
12:52 | <vagrantc> unionfs didn't work over NFS during the initial implementation. we should revisit it
| |
12:52 | <Gadi> ah, ok
| |
12:52 | it just seemed fishy that it was being set and not used
| |
12:53 | <vagrantc> i know it works, because i've used it with debian-live recently
| |
12:53 | <Gadi> that would/could speed up the boot some (since we wouldn't have all the copying from nfs to ram)
| |
12:53 | <vagrantc> Gadi: though, it would also probably waste some ram when booting. right now much of the filesystem is unwriteable
| |
12:53 | <sbalneav> vagrantc: Is there a way we could use the same strings from ssh? I don't know anything about i18n, so I'd need some help there.
| |
12:54 | <vagrantc> sbalneav: maybe. i don't know it real well myself.
| |
12:54 | <Gadi> vagrantc: do we not use RAM for all the tmpfs's already?
| |
12:55 | <vagrantc> Gadi: yes, but if more parts of the filesystem were writeable, more things would be written to ram.
| |
12:55 | <Gadi> i was thinking instead of the tmpfs's
| |
12:55 | <vagrantc> Gadi: like /lib/modules/*/modules.dep files might get updated
| |
12:56 | Gadi: what are you going to add to the union?
| |
12:56 | Gadi: i was thinking of layering tmpfs on top of NFS ... what were you proposing?
| |
12:56 | <Gadi> just make a writable tmpfs mount and add it to the union (but dont copy data to it)
| |
12:56 | right
| |
12:56 | <vagrantc> so ... you're still using a tmpfs
| |
12:57 | <Gadi> so, instead of many tmpfs's
| |
12:57 | <vagrantc> there's only one tmpfs now
| |
12:57 | <Gadi> the key is no copying
| |
12:57 | I see a million tmpfs's
| |
12:57 | ;)
| |
12:57 | <vagrantc> except that during the boot process things will get modified and copied into the tmpfs that won't on a read-only filesystem
| |
12:57 | <Gadi> all bind-mounted
| |
12:58 | and then the contents populated
| |
12:58 | <vagrantc> it looks like a billion tmpfs ... but it's only one, with bind mounts
| |
12:59 | <Gadi> right, ok - but if we layer a tmpfs on top of the nfs, we wouldn't need to copy files
| |
12:59 | all the mods would be written to the tmpfs
| |
12:59 | its prolly splitting hairs
| |
12:59 | not much of a speed drag in those file-copies
| |
13:00 | <vagrantc> and what i'm saying is that it ends up copying even more into the tmpfs in the end
| |
13:00 | <Gadi> tho, I imagine the more dependency on the network, the worse off we are
| |
13:00 | <vagrantc> several megabytes
| |
13:00 | <Gadi> really?
| |
13:00 | <vagrantc> yes.
| |
13:01 | <Gadi> how come?
| |
13:01 | it doesnt need to copy the contents for the union to work, does it?
| |
13:01 | <vagrantc> there are several things in the boot process that will update files trivially, which will then copy them into the initramfs
| |
13:02 | and the biggest thing will be the debconf databases
| |
13:02 | which will get backed up
| |
13:02 | eats up about 3MB unless we "clean" it out at the end.
| |
13:03 | though how to properly clean it is dubious
| |
13:03 | <Gadi> I see - what about whiting out those dirs?
| |
13:03 | <vagrantc> we need them
| |
13:03 | currently ... when X autoconfigures itself we maybe won't need them
| |
13:03 | which would be a huge speed improvement
| |
13:03 | <Gadi> how about excluding the dirs from the union?
| |
13:03 | so you still get the readonly files
| |
13:04 | <vagrantc> well, we still need them writeable
| |
13:04 | it's a mess, i tell you :)
| |
13:04 | <Gadi> so how is it any different with unionfs than what we do now?
| |
13:04 | <vagrantc> we do some ugly hacks that would be harder to do with unionfs
| |
13:05 | * Gadi shrugs - according to ltsp-client-setup it looks like you don't do anything too hacky | |
13:05 | <Gadi> other than only bind-mount-copy certain dirs
| |
13:06 | which (looks) akin to just a unionfs with a copy
| |
13:06 | pablo_ has joined #ltsp | |
13:08 | <pablo_> my clients wont show gnome desktop after loging in gdm :( instead nothing appears and after a while gdm login screen is displayed again
| |
13:09 | <vagrantc> again, we just need to try unionfs again and see what happens.
| |
13:09 | <pablo_> im using ubuntu 7.04 and ltsp5
| |
13:09 | <vagrantc> pablo_: did you configure it to use GDM, or is it really using LDM ?
| |
13:11 | <pablo_> mm i didnt configure it to use gdm (i guess its using it) but maybe its using ldm
| |
13:11 | ltsp 4.2 used gdm right?
| |
13:11 | ltsp5 uses ldm?
| |
13:11 | <vagrantc> by default, yes. ltsp5 uses LDM.
| |
13:12 | <pablo_> anyway why cant i see the gnome desktop after login? :(
| |
13:12 | should i use gdm instead?
| |
13:12 | <vagrantc> no
| |
13:12 | we just need to troubleshoot the problem
| |
13:13 | pablo_: is there anything in your user's ~/.xsession-errors file suggesting something wrong?
| |
13:13 | pablo_: can you log in as that user just using ssh ?
| |
13:14 | <mistik1> good day guys
| |
13:14 | <pablo_> let me see..
| |
13:14 | <sbalneav> pablo_: You probably need to run a sudo ltsp-update-sshkeys
| |
13:15 | Hello mistik1
| |
13:15 | <vagrantc> sbalneav: does that problem still exist in feisty ?
| |
13:15 | <Gadi> mistik1 !!!
| |
13:15 | <mistik1> ;-)
| |
13:17 | <pablo_> vagrantc: yes im loged in via ssh with the problematic user. i lookes at .xsessionerrors and..
| |
13:18 | lots of errors apper related to sabayon (profile editor)
| |
13:18 | but no sure if any of those are really the cause :( i tried loogin with a normal user (in the terminals)
| |
13:19 | and same happend (normal user means im not using sabayon profile editor to customize the profile)
| |
13:19 | <vagrantc> no.
| |
13:19 | it's probably some problem with the ssh connection
| |
13:19 | pablo_: try what sbalneav said: sudo ltsp-update-sshkeys
| |
13:19 | <pablo_> sbalneav: sshkeys are already updated
| |
13:20 | i did it as soon as i installed ltsp
| |
13:20 | <vagrantc> pablo_: is there any chance your ip address has changed?
| |
13:20 | <pablo_> no.
| |
13:21 | wait you mean that x travels via ssh?
| |
13:21 | <vagrantc> yes
| |
13:21 | <pablo_> ok
| |
13:21 | <vagrantc> that's what LDM does
| |
13:21 | <Gadi> why not reboot with SCREEN_07=ldm and SCREEN_02=shell and then you can flip to the shell and check /var/log/ldm.log
| |
13:21 | <vagrantc> so look for ssh errors in ~/.xsession-errors
| |
13:22 | Gadi: does tty2 not contain anything on ubuntu?
| |
13:22 | <Gadi> vagrantc: the above works
| |
13:22 | that will get you a prompt on tty2
| |
13:22 | <vagrantc> ok.
| |
13:23 | <Gadi> hey, vagrantc: what sets $SERVER on the client?
| |
13:23 | <vagrantc> /usr/lib/ltsp/ltsp_config i think
| |
13:24 | <Gadi> ah, indeed - thx
| |
13:28 | <pablo_> Gadi: i canot log in the terminals via console
| |
13:28 | i do ctrl+alt+f1
| |
13:29 | <Gadi> pablo_: create a: /opt/ltsp/i386/etc/lts.conf
| |
13:29 | and add:
| |
13:29 | SCREEN_07 = ldm
| |
13:29 | SCREEN_02 = shell
| |
13:29 | then, reboot the client
| |
13:29 | then, you chould be able to hit: ctrl-alt-f2 for shell on the client
| |
13:29 | and ctrl-alt-f7 for LDM
| |
13:30 | try logging in through ldm
| |
13:30 | when it is unsuccessful, go to the shell and cat /var/log/lgdm.log
| |
13:30 | er, ldm.log
| |
13:30 | it should tell you why it failed
| |
13:31 | --> sbalneav: perhaps this can be added to the greeter, as well, in some fashion: bring up the reason for failed login
| |
13:31 | <sbalneav> Right
| |
13:32 | That was my idea with haveing stdin/out interaction with the greeter, to feed it errors/modify it's action.
| |
13:32 | <jammcq> wow, wouldn't that be nice, to show the user what's wrong, rather than just blanking the screen and coming back to the login prompt :)
| |
13:33 | <Gadi> yeah, lets make it cryptic
| |
13:33 | like "Error 54673"
| |
13:33 | ;)
| |
13:33 | <jammcq> even that would be better than what we have now :)
| |
13:33 | * Gadi imagines Windows developers sitting around a table | |
13:34 | <jammcq> and what... ordering lunch ?
| |
13:34 | <ogra-classmate> Gadi: perfect, we could sell an error code book
| |
13:34 | <Gadi> heh
| |
13:34 | <ogra-classmate> several even ...
| |
13:34 | edition one, error 0001 to error 2000
| |
13:34 | etc :)
| |
13:34 | <Gadi> love it
| |
13:34 | <jammcq> we could call it a "resource kit"
| |
13:35 | <sbalneav> ogra-classmate: When's the next community council meeting?
| |
13:35 | <ogra-classmate> sbalneav: nothing on the fridge ?
| |
13:35 | <sbalneav> What's the fridge?
| |
13:35 | <jammcq> heh
| |
13:35 | <ogra-classmate> fride.ubuntu.com/events
| |
13:35 | <jammcq> the thing where you post tings
| |
13:35 | <ogra-classmate> err
| |
13:35 | <sbalneav> ah
| |
13:35 | <jammcq> things
| |
13:35 | <ogra-classmate> fridge
| |
13:36 | <jammcq> a place for sticky magnets
| |
13:36 | <ogra-classmate> the thing you put magnets on ... with little notes
| |
13:36 | heh
| |
13:37 | http://fridge.ubuntu.com/event btw
| |
13:37 | <sbalneav> Can anyone voice an opinion in the meeting, or is it open only to CC members?
| |
13:37 | eurialo_ has joined #ltsp | |
13:37 | <jammcq> or just http://fridge.ubuntu.com
| |
13:37 | <ogra-classmate> there is also an ical tou can import into evo
| |
13:37 | that keeps itself updated on changes :)
| |
13:38 | http://fridge.ubuntu.com/event/ical
| |
13:38 | <Gadi> wow, evo finally supports that? ;)
| |
13:38 | <ogra-classmate> since ages
| |
13:39 | <Gadi> and it doesnt make it crash?
| |
13:39 | <ogra-classmate> i use it since i work for canonical, so its older than two years
| |
13:39 | <Gadi> :)
| |
13:39 | <ogra-classmate> my evo doesnt crash
| |
13:39 | intrestingly i didnt notice that fact until you just mentioned it
| |
13:40 | mhterres has joined #ltsp | |
13:40 | <ogra-classmate> it didnt crash on me since months
| |
13:40 | <Gadi> wow
| |
13:40 | <mhterres> afternoon
| |
13:40 | <Gadi> ogra-classmate: quick - make a wish
| |
13:40 | ;)
| |
13:41 | <ogra-classmate> hehe
| |
13:41 | i wish that evo only ate 10% of ram it does ...
| |
13:41 | that would suit that kind of app
| |
13:42 | but thunderbird isnt much better so i wont complain
| |
13:43 | sbalneav: http://fridge.ubuntu.com/node/931
| |
13:43 | <sbalneav> Use mutt :)
| |
13:43 | <ogra-classmate> tomorrow evening
| |
13:43 | <sbalneav> 21:00 UTC, I assume?
| |
13:43 | * Gadi uses gmail now | |
13:43 | <Gadi> ;)
| |
13:44 | <ogra-classmate> sbalneav: well, i like to read my mail in a gui client, but every client we have is either bulky or has horrible usability issues
| |
13:44 | <sbalneav> I found the same.
| |
13:44 | Long long ago...
| |
13:44 | in a galaxy, far, far away
| |
13:45 | * ogra-classmate remembers a time where mailclients used to be small, fast apps that did one thing | |
13:45 | <sbalneav> I used to just use "mailx"
| |
13:45 | <ogra-classmate> ... display mail
| |
13:45 | <sbalneav> Then, I used elm
| |
13:45 | Then I used pine
| |
13:45 | <pablo_> Gadi: i did what you say and. i cant be sure of what is the error that causes gnome desktop not to display in the terminal
| |
13:45 | gepatino has joined #ltsp | |
13:45 | <sbalneav> Now I use mutt.
| |
13:45 | <ogra-classmate> i was a long time elm user as well
| |
13:45 | then i switched to netscape/mozilla mail
| |
13:45 | <sbalneav> pablo_: Paste the output to the pastebot, and we'll look
| |
13:46 | !pastebot
| |
13:46 | <ltspbot`> sbalneav: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
| |
13:46 | <ogra-classmate> and when it stopped wiping mailboxes i switched to evo
| |
13:46 | <Gadi> gmail is lightweight and always accessible
| |
13:46 | and fast
| |
13:46 | <ogra-classmate> well, it actually didnt have stopped when i switched ...
| |
13:46 | <Gadi> it works beautifully for me
| |
13:47 | <ogra-classmate> weell
| |
13:47 | <Gadi> heh
| |
13:47 | <vagrantc> alright folks: http://llama.freegeek.org/~vagrant/bzr/ltsp/features/vagrant-ltsp-client-core
| |
13:47 | <ogra-classmate> i have about 200000 mails ... i work for a software producing company ....
| |
13:47 | so indeed i use my horrible mailbox to stresstest our software :)
| |
13:48 | <Gadi> ah, internal spam :)
| |
13:48 | <ogra-classmate> but often it rather stresstests me
| |
13:49 | <pablo_> ok but i have a silly problem (im new in linux) where is the file /var/log/ldm.log so i can pastebot it? i mean i go to /var/log/ldm.log from the client ok. but forn the server? where do i find it? sorry for the silly questions :)
| |
13:49 | gepatino has quit IRC | |
13:50 | <ogra-classmate> there is none for the server side
| |
13:50 | you can look in /home/$USER/.xsession-errors for all stuff that happened after the session started and in /var/log/auth.log for everything login related
| |
13:52 | <Gadi> pablo_: also, try this: edit /opt/ltsp/i386/etc/lts.conf and turn off local devices
| |
13:52 | and sound
| |
13:52 | and then try logging in again
| |
13:52 | eurialo has quit IRC | |
13:52 | <ogra-classmate> Gadi: how should that help ?
| |
13:53 | <Gadi> it could be as simple as him not having ltspfs installed server side, while having local devices enabled
| |
13:53 | <ogra-classmate> running ltsp-update-sshkeys on the server would be a better advice ;)
| |
13:53 | <Gadi> that would produce the same outcome
| |
13:53 | he has done that already, wise guy
| |
13:53 | <pablo_> iv already done ltsp.update-sshkeys
| |
13:53 | <vagrantc> sure doesn't hurt to run lts-update-sshkeys again
| |
13:54 | <ogra-classmate> whoops sorry
| |
13:54 | <pablo_> ive done it again but problem still there
| |
13:54 | <vagrantc> ok. good.
| |
13:54 | that at least rules out doubt, then. :)
| |
13:54 | <Gadi> my money's on no ltspfs installed server side
| |
13:55 | pablo_: do this on the server; dpkg -l ltspfs
| |
13:55 | <vagrantc> Gadi: does that actually cause login to fail, though ?
| |
13:55 | <Gadi> indeed
| |
13:55 | * vagrantc hasn't seen that | |
13:55 | <Gadi> ssh will call ltspfsmounter
| |
13:55 | <ogra-classmate> Gadi: are you sure he doesnt have sabayon installed?
| |
13:55 | <Gadi> which does not exist
| |
13:56 | <vagrantc> ogra-classmate: we are sure he does have sabayon installed ... why ?
| |
13:56 | <Gadi> he does have sabayon installed
| |
13:56 | <ogra-classmate> that wouldnt let anyone in until you add every user to a profile
| |
13:56 | its a well known design bug of sabayon
| |
13:56 | * ogra-classmate just got dinner | |
13:56 | <ogra-classmate> bbl
| |
13:57 | * Gadi worries about his bet... | |
13:57 | <Gadi> :)
| |
13:57 | * vagrantc needs to get dinner | |
13:57 | <vagrantc> the cleaning folks threw out some of my food :(
| |
13:58 | <pablo_> Gadi: i did dpkg -l ltsps and seems not to be installed. i opened synaptic searched for ltspfs and. i see its not installed
| |
13:59 | <Gadi> ah, try installing it and then log in on the client
| |
13:59 | <pablo_> but this calls my atention cause i followed the ltsp doc (ubuntu quick start) to install ltsp5 (maybe that was missing there?)
| |
13:59 | * Gadi may win his bet after all | |
13:59 | <Gadi> pablo_: could be - I think the change to enable local devices by default was a last minute thing
| |
14:02 | vagrantc has quit IRC | |
14:07 | Q-FUNK has joined #ltsp | |
14:13 | <pablo_> Gadi: well it didnt work :(
| |
14:14 | i installed ltspfs rebooted a client and again same problem
| |
14:14 | <Gadi> in sabayon, are all of your users added to a profile?
| |
14:14 | <pablo_> yes they are
| |
14:14 | <Gadi> hmm... try editing lts.conf and disabling sound and local devices
| |
14:15 | <pablo_> wait wait..!! not all users where added to the sabayon profile
| |
14:15 | let me test again
| |
14:19 | eurialo__ has joined #ltsp | |
14:22 | <pablo_> well all users in belonging to the sabayon profile but the problema ramains
| |
14:22 | the only errors i see in /var/log/ldm.log fron the clients
| |
14:23 | are some font paths not found
| |
14:23 | Q-FUNK has quit IRC | |
14:23 | <pablo_> and a gtk waring tahs says no support for gnome
| |
14:23 | then continues
| |
14:24 | Q-FUNK has joined #ltsp | |
14:24 | <pablo_> and last 2 errors are related with fonts but i dont see a critical error that causes the problem. maybe im missing something
| |
14:24 | ill tri disabling suond and local devs from lts.conf
| |
14:25 | apart from that.. why does ltsp5 takes so longo to load in the terminals? 4.2 was much much faster in terms of speed
| |
14:26 | <Gadi> pablo_: indeed - that is the source of current development efforts
| |
14:26 | ltsp4 was built from ground up
| |
14:26 | ltsp5 is built from sky down, as it were
| |
14:26 | ;)
| |
14:27 | <pablo_> Gadi: ok tanks for the explanation :)
| |
14:30 | muh2000 has quit IRC | |
14:33 | <pablo_> well now i seted localdev=false and sound=false in lts.conf but problem remains there :( anythign else should i try?
| |
14:35 | im seriouly thinking of installing ltsp 4.2 (it wqas so much faster!)
| |
14:36 | <Gadi> and you have no errors in ~/.xsession-errors?
| |
14:36 | or /var/log/auth.log on the server?
| |
14:37 | eurialo_ has quit IRC | |
14:37 | * jammcq has to head to a family party. Only problem is, no beef and no beer at this party. kinda hard to call it a party, eh? | |
14:38 | meduxa has quit IRC | |
14:47 | <ltsppbot> "pablo_" pasted ".xsession-errors" (12 lines) at http://pastebot.ltsp.org/147
| |
14:47 | <pablo_> could you look at it to see if there is something really wrong?
| |
14:50 | eurialo_ has joined #ltsp | |
14:51 | <ltsppbot> "pablo_" pasted "tail -f /var/log/auth.log" (9 lines) at http://pastebot.ltsp.org/148
| |
14:53 | <Gadi> OSError: [Errno 2] No existe el fichero ó directorio: '/home/draco2/.gnome2/share'
| |
14:53 | that looks like the error
| |
14:54 | seems like a sabayon error
| |
14:55 | like it didn't properly create the user's .gnome2 directory
| |
14:55 | try either: not using sabayon OR repush out the sabayon profiles for the users
| |
14:56 | <pablo_> Gadi: ok maybe it has to do wiht sabayon but 2 questions the fact that ive never loged in as these problematic usershas anything to do?
| |
14:57 | <Gadi> not sure
| |
14:57 | <pablo_> now i tried to log in the terminals wjth a user that has nothing to do wiht sabayon profiles
| |
14:57 | and i have the same problem so it doesn t appears to be sabayon related no?
| |
14:58 | <Gadi> according to ogra, if you have sabayon installed, all users must belong to a sabayon profile
| |
14:58 | that it is a bug with sabayon
| |
14:58 | so, you should associate all users with a profile
| |
14:58 | then, push the profiles out again
| |
14:59 | <pablo_> but ALL system users? including this im using now? or only the ones that are going to be used in the terminal?
| |
14:59 | terminals?
| |
14:59 | <Gadi> all users
| |
15:00 | <pablo_> ok ill try it now
| |
15:00 | <Gadi> (not system users like "gdm", obviously)
| |
15:00 | but, all regular users on the system
| |
15:00 | <pablo_> sure :)
| |
15:01 | eurialo has joined #ltsp | |
15:06 | vagrantc has joined #ltsp | |
15:07 | eurialo__ has quit IRC | |
15:09 | eurialo has quit IRC | |
15:09 | <pablo_> ok i tried it but didnt work. how can i uninstall ltsp5 and install ltsp4.2? or they can live together?
| |
15:10 | <vagrantc> pablo_: well, you could just try to configure XDMCP, since that's what you'd get with ltsp4.2 anyways
| |
15:11 | pablo_: set SCREEN_07=startx in lts.conf, and install gdm or kdm or xdm on your server ...
| |
15:11 | !grayscreen
| |
15:11 | <ltspbot`> vagrantc: Error: "grayscreen" is not a valid command.
| |
15:11 | <pablo_> vagrantc: dpkg-reconfigure XDMCP?
| |
15:11 | <vagrantc> !greyscreen
| |
15:11 | <ltspbot`> vagrantc: "greyscreen" is No login, and only a grey screen? See http://wiki.ltsp.org/twiki/bin/view/Ltsp/Troubleshooting-GrayScreen
| |
15:11 | <vagrantc> pablo_: there's instructions there
| |
15:11 | <Gadi> pablo_: if it is a problem with sabayon, prolly ltsp 4.2 would fail, too
| |
15:12 | <vagrantc> oh.
| |
15:12 | <Gadi> perhaps try first without sabayon?
| |
15:12 | <vagrantc> right.
| |
15:12 | <Gadi> vagrantc: find dinner?
| |
15:12 | fernando1 has quit IRC | |
15:13 | <vagrantc> Gadi: yeah. though half of my food stash disappeared on me.
| |
15:13 | <pablo_> i did it. i tried to login in a terminal with a user not related to sabayon profiles (before setting all users to use sabayon profiel) but didnt work
| |
15:14 | <Gadi> pablo_: sabayon messes with the users' accounts, as well as /etc/skel, I think
| |
15:14 | which means it affects all your users
| |
15:14 | <vagrantc> awesome.
| |
15:15 | <Gadi> try blowing away the user's .gnome, .gnome2, .gconf, and .gconfd directories
| |
15:15 | and log in
| |
15:15 | fernando1 has joined #ltsp | |
15:16 | eurialo_ has quit IRC | |
15:16 | <Gadi> I use to many idioms - what I mean is: rm -rf ~/.gconf ~/.gcnofd ~/.gnome ~/.gnome2
| |
15:17 | * vagrantc speaks shell semi-fluently | |
15:21 | <sbalneav> rm -rf ${ALL_GNOMES_CONF_FILES}
| |
15:22 | eurialo has joined #ltsp | |
15:22 | <sbalneav> Just make sure you don't have export ALL_GNOMES_CONF_FILES="*" defined anywhere :)
| |
15:23 | <eurialo> hi, i follow this howto http://wiki.debian.org/LTSP/Howto, i add the users to fuse group.
| |
15:24 | How can i make local device work?
| |
15:24 | <vagrantc> eurialo: dpkg -l ltspfs ; dpkg --root=/opt/ltsp/i386 ltspfsd
| |
15:25 | eurialo:
| |
15:25 | actually
| |
15:25 | eurialo: dpkg -l ltsp* ; dpkg --root=/opt/ltsp/i386 ltsp*
| |
15:27 | <eurialo> the output on server
| |
15:27 | pn ltsp-client-bu <non definita> (descrizione non disponibile)
| |
15:27 | ii ltsp-server 0.99debian12 Basic LTSP server environment
| |
15:27 | ii ltsp-server-st 0.99debian12 Complete LTSP server environment
| |
15:27 | un ltsp-utils <non definita> (descrizione non disponibile)
| |
15:27 | ii ltspfs 0.4.3+debian2 Fuse based remote filesystem for LTSP thin c
| |
15:28 | on client chroot
| |
15:28 | ii ltsp-client 0.99debian12 LTSP client environment
| |
15:28 | un ltsp-utils <non definita> (descrizione non disponibile)
| |
15:28 | ii ltspfsd 0.4.3+debian2 Fuse based remote filesystem daemon for LTSP
| |
15:28 | <vagrantc> ok
| |
15:28 | grep LOCALDEV /opt/ltsp/i386/etc/lts.conf
| |
15:28 | <eurialo> i have localdev=true
| |
15:28 | <vagrantc> LOCALDEV=true ?
| |
15:29 | case matters.
| |
15:29 | well, for the variable name.
| |
15:29 | not for the "true" bit.
| |
15:33 | eurialo: so ... ?
| |
15:34 | <eurialo> i'm going to reboot the client
| |
15:36 | thi is my lts.conf
| |
15:36 | [Default]
| |
15:36 | LOCALDEV=True
| |
15:36 | SOUND=True
| |
15:36 | but nothing change
| |
15:37 | localdevice not work
| |
15:38 | <vagrantc> eurialo: getent group fuse
| |
15:39 | <eurialo> the client is in fuse group
| |
15:40 | <vagrantc> "client" ?
| |
15:40 | dudleif has joined #ltsp | |
15:40 | <eurialo> the user sorry
| |
15:40 | :)
| |
15:40 | <vagrantc> ok.
| |
15:41 | nicoAMG has quit IRC | |
15:42 | <vagrantc> eurialo: what kind of device is it?
| |
15:43 | <eurialo> i try with floppy and cd too
| |
15:43 | <vagrantc> built-in floppy? built-in cd ?
| |
15:43 | i.e. not USB or something
| |
15:44 | having never tested the floppy/cdrom support, i'm a little weak at troubleshooting it ...
| |
15:44 | sbalneav: ... ?
| |
15:45 | <sbalneav> boo
| |
15:45 | nf1 has joined #ltsp | |
15:45 | <eurialo> yes, they are built-in
| |
15:45 | <vagrantc> sbalneav: static devices with ltspfs ... i'm a little unfamiliar with how it's supposed to work
| |
15:46 | <eurialo> i can try with a pendrive
| |
15:46 | <sbalneav> There's supposed to be a daemon that re-adds them all the time.
| |
15:46 | hmmm
| |
15:46 | <vagrantc> eurialo: that would be good.
| |
15:46 | J45p3r_ has joined #ltsp | |
15:47 | J45p3r has quit IRC | |
15:48 | <sbalneav> Try a pendrive for now, see if that works.
| |
15:49 | <eurialo> yes, it works
| |
15:49 | <vagrantc> well, at least we know it's not completely broken.
| |
15:50 | Q-FUNK has quit IRC | |
15:50 | <Gadi> have you rebooted the client since you fixed things?
| |
15:51 | <eurialo> yes, i do it again
| |
15:51 | sound also not work
| |
15:51 | <vagrantc> sound is anoher issue
| |
15:52 | <eurialo> ok
| |
15:52 | <vagrantc> eurialo: as the logged in user, echo $ESPEAKER
| |
15:53 | <sbalneav> What I need to do is come up with a localdev troubleshooting page for ltsp5
| |
15:53 | <vagrantc> sbalneav: yes, please. :)
| |
15:55 | <Gadi> so, who was the genius who decided to make a linux command called ipconfig? ;)
| |
15:56 | <eurialo> echo $ESPEAKER:
| |
15:56 | 10.0.0.101:16001
| |
15:56 | <str4nd> lo.o.o.lol:l6ool
| |
15:56 | nvm ->
| |
15:58 | <vagrantc> eurialo: so ... try running a sound program from the commandline with esddsp or something
| |
15:58 | i.e. esddsp xmms
| |
15:58 | see if that works
| |
16:04 | Guaraldo has left #ltsp | |
16:08 | sbalneav has quit IRC | |
16:09 | dudleif has quit IRC | |
16:12 | <eurialo> it says failed to open mixer device (/dev/mixer)
| |
16:13 | sorry for long time but client is in an other room
| |
16:20 | vagrantc has quit IRC | |
16:27 | <eurialo> nothing to do, sound doesen't work
| |
16:29 | <Gadi> eurialo: does the client have a PCI soundcard or an ISA?
| |
16:30 | <eurialo> on board sound card
| |
16:30 | <Gadi> what type of client is it?
| |
16:30 | <eurialo> but it work with edubuntu 7.4
| |
16:31 | <Gadi> edubuntu LTSP?
| |
16:31 | <eurialo> (what type of client is it?)
| |
16:31 | hardware?
| |
16:32 | <Gadi> yes
| |
16:33 | listen, here are the things to check:
| |
16:33 | <eurialo> athlon xp 2400, 512mb, abit motherboard
| |
16:34 | <Gadi> 1. The driver was actually loaded on the client (if it worked with edubuntu, chances are that's fine)
| |
16:34 | 2. You can go to System > Preferences > Sound and have output set to "Auto" and do a test
| |
16:35 | (Sound Playback, that is)
| |
16:35 | if you can do #2, then sound works fine in general, and may not work for the particular application you tried
| |
16:36 | if you cannot do #2, then there is some incorrect setting that we need to find
| |
16:36 | note: for sound to work, the application needs to be able to use "esd" or "PulseAduio"
| |
16:37 | some applications are not capable of this (like Java) because they write directly to the sound device
| |
16:37 | most are fine
| |
16:45 | <eurialo> i can do #2
| |
16:45 | <Gadi> then, your sound is fine
| |
16:46 | what application was not playing sound?
| |
16:46 | <eurialo> but the sound icon say me there is no device
| |
16:46 | :?
| |
16:46 | <Gadi> that used to mean that you had no soundcard in the *server* ;)
| |
16:47 | i think ogra fixed that in Ubuntu, tho
| |
16:47 | to actually control the client's audio
| |
16:47 | not sure
| |
16:47 | <eurialo> no problem, i remove the icon
| |
16:47 | :)
| |
16:47 | <Gadi> :)
| |
16:48 | <eurialo> but xmms, totem, rhythmbox dont work
| |
16:48 | i will try to reboot server
| |
16:48 | <Gadi> did you have Sound Playback set to auto?
| |
16:48 | or did you have to change it from something else?
| |
16:49 | <eurialo> i have no auto option
| |
16:49 | <Gadi> if you had to change it, try logging out and logging back in
| |
16:49 | what?
| |
16:49 | Autodetect
| |
16:49 | sorry
| |
16:49 | <eurialo> no
| |
16:49 | <Gadi> what? no autodetect?
| |
16:49 | what distribution?
| |
16:50 | <eurialo> debian lenny
| |
16:50 | <Gadi> (I thought you said you did #2 and it worked)
| |
16:50 | PMantis has joined #LTSP | |
16:50 | <Gadi> whats debian lenny?
| |
16:51 | <eurialo> debian etch+1
| |
16:51 | <Gadi> ah
| |
16:51 | <eurialo> debian testing
| |
16:51 | <Gadi> hmm
| |
16:51 | what did you set sound playback to?
| |
16:51 | <eurialo> wait, i send you a screenshot
| |
16:57 | did you give it?
| |
16:58 | highvoltage has quit IRC | |
16:58 | highvoltage has joined #ltsp | |
16:58 | <Gadi> no
| |
16:59 | just tell me
| |
16:59 | is it set to esd or pulseaudio?
| |
17:05 | eurialo has quit IRC | |
17:05 | eurialo has joined #ltsp | |
17:06 | <eurialo> http://img512.imageshack.us/my.php?image=schermataoi7.png
| |
17:08 | @Gadi
| |
17:09 | <Gadi> ah, open a terminal window and type: gstreamer-properties
| |
17:09 | see what that is set to
| |
17:11 | * Gadi has to head home | |
17:11 | <Gadi> good luck, eurialo
| |
17:11 | Gadi has left #ltsp | |
17:27 | eurialo has quit IRC | |
17:29 | bill_c has quit IRC | |
17:39 | eurialo has joined #ltsp | |
17:56 | mhterres has quit IRC | |
18:02 | Legenda has quit IRC | |
18:09 | maaarek has quit IRC | |
18:10 | bill_c has joined #ltsp | |
18:19 | eurialo has quit IRC | |
18:40 | muh2000 has joined #ltsp | |
19:29 | muh2000 has quit IRC | |
19:56 | a5benwillis has quit IRC | |
20:08 | jsgotangco has joined #ltsp | |
20:28 | irule has quit IRC | |
20:48 | J45p3r_ has quit IRC | |
21:16 | <kaos01> hi, i get sshd[1061]: error: Couldn't create pid file "/etc/ssh/sshd.pid": Read-only file system in /var/log/messages
| |
21:16 | shouldnt it be putting pid in tmp or somethign the like
| |
21:16 | ogra-classmate has quit IRC | |
21:23 | ogra has quit IRC | |
21:24 | ogra has joined #ltsp | |
21:59 | pablo_ has quit IRC | |
22:11 | pablo_ has joined #ltsp | |
22:31 | ctyler has joined #ltsp | |
22:38 | ctyler has quit IRC | |
22:39 | muh2000 has joined #ltsp | |
23:08 | jsgotangco has quit IRC | |
23:13 | siji has joined #ltsp | |
23:28 | siji has quit IRC | |
23:41 | Jenna has joined #ltsp | |