00:16 | sbalneav has left #ltsp | |
01:00 | PMantis has quit IRC | |
03:57 | ruede has joined #ltsp | |
04:06 | muh2000 has quit IRC | |
04:54 | Arsenick-TC2L has quit IRC | |
05:56 | gigabytes has joined #ltsp | |
06:16 | gigabytes has quit IRC | |
07:14 | shogunx has quit IRC | |
07:34 | ruede has quit IRC | |
07:35 | muh2000 has joined #ltsp | |
08:56 | chupacabra has quit IRC | |
10:14 | cliebow_ has joined #ltsp | |
10:18 | <cliebow_> http://pastebin.com/822574..jammcq..if you get a chance..see if anything is wonly doing like this?
| |
10:22 | apple appears to need a fixed address
| |
10:28 | Skarmeth has joined #ltsp | |
11:45 | <jammcq> http://hardware.slashdot.org/hardware/06/11/12/1535227.shtml
| |
11:45 | cliebow_: read that article I just linked
| |
11:58 | <cliebow_> Cool!
| |
12:02 | jammcq:mind taking a peek?anything wonky? http://pastebin.com/822574
| |
12:02 | <jammcq> yeah, it looks ok to me
| |
12:02 | I have no idea what the apple is expecting tho
| |
12:03 | <cliebow_> mac boots fine if i hard code and puthe same info in
| |
12:03 | soit seems to need a bootp like thingie
| |
12:05 | re slashdot..point taken..
| |
12:22 | EversonCaratti has joined #ltsp | |
12:34 | ogra has joined #ltsp | |
12:41 | <jammcq> ogra: hey
| |
12:42 | <ogra> yo
| |
12:43 | <cliebow_> ogra: X_MODE_0 supported now in ltsp?
| |
12:43 | i have a wonky lcd
| |
12:43 | <ogra> cliebow_, not yet ... but it should be in feusty
| |
12:43 | *feisty
| |
12:43 | <jammcq> ogra: how's fishermans wharf?
| |
12:43 | <cliebow_> any way to specfy a modeline?
| |
12:43 | <ogra> nope
| |
12:44 | <cliebow_> k
| |
12:44 | <ogra> use X_HORZSYNC and X_VERTREFRESH
| |
12:44 | if you supply the right values for your monitor it should work ok
| |
12:44 | <cliebow_> i try it tomorrow
| |
12:44 | <ogra> alternatively use a static xorg.conf
| |
12:45 | <cliebow_> wife gets agitated when i fill the kving room with my crap
| |
12:45 | <ogra> jammcq, well ... i saw everything already ... and my wallet got stolen in the club near the restaurant on friday
| |
12:45 | <jammcq> NO
| |
12:45 | <cliebow_> goog grief!!
| |
12:45 | <jammcq> that sux
| |
12:46 | <ogra> luckily i took out my credit card before and my passport was in the hotel ... so i'm still fine to survive ....
| |
12:46 | <jammcq> wow, that really sux
| |
12:46 | <ogra> went to SFPD yesterday to report it ... they laugehd at me and told me i'd have to go back to MV to report it there
| |
12:46 | <jammcq> heh
| |
12:48 | arc_ has joined #ltsp | |
13:01 | vagrantc has joined #ltsp | |
13:03 | J45p3r_ has joined #ltsp | |
13:03 | <ogra> vagrantc, !
| |
13:03 | <vagrantc> ogra: heya! how was the conference?
| |
13:04 | <ogra> great !
| |
13:04 | we had 10 specs, 8 are already approved
| |
13:04 | <vagrantc> cool.
| |
13:04 | did the encrypted nbd stuff get specced?
| |
13:05 | was there any talk about getting an upstream tarball for ltsp 5 ?
| |
13:06 | <jammcq> vagrantc: we've already got the tarball, I just need to make it available on the LTSP website
| |
13:06 | <ogra> nope, the encrypted nbd stuff is just a simple enhancement, no need for a spec
| |
13:06 | <jammcq> and... it needs some work, to de-brand it
| |
13:06 | <ogra> http://people.ubuntu.com/~ogra/ltsp-tarballs/ has the recent edgy version
| |
13:06 | <vagrantc> ogra: it seems like it would be good to get encrypted nbd in the nbd packages directly...
| |
13:06 | <ogra> (only i386 yet until i get home)
| |
13:06 | then lets patch it ;)
| |
13:07 | <jammcq> vagrantc: how would you do encrypted nbd ?
| |
13:08 | <ogra> he already does it ;)
| |
13:08 | <vagrantc> jammcq: how i implemented it in debian is that you set up the nbd device client-side, and then set up an encrypted device with cryptsetup that uses nbd as the backend.
| |
13:08 | * jammcq wonders how | |
13:08 | <ogra> i was planning to include the cryptosetup patches
| |
13:08 | <jammcq> hmmm
| |
13:08 | cool
| |
13:08 | <vagrantc> though i think it has a weakness- the encryption process may go over the wire at the initial setup, which would compromise it's security.
| |
13:08 | <ogra> its very straigthforward ...
| |
13:09 | right
| |
13:09 | <vagrantc> but it's still probably better than absolutely no encryption
| |
13:09 | probably would be better to set up a vpn or something :)
| |
13:09 | <ogra> i'm not sure i like the idea of patching it into nbd directly .... sunds a bit like bloeting it
| |
13:09 | *bloating
| |
13:10 | <vagrantc> well, it's a feature that anyone using nbd would probably want, and i don't think it will significantly change it
| |
13:11 | <ogra> its a good amount of code i guess ...
| |
13:11 | <vagrantc> you could just implment the nbd code i have for ltsp in the nbd init scripts
| |
13:11 | <ogra> ndb is tiny ...
| |
13:11 | *nbd
| |
13:11 | <jammcq> what does encryption do to performance?
| |
13:11 | <ogra> and still use the cryptsetup package ?
| |
13:11 | <vagrantc> ogra: yeah.
| |
13:11 | <jammcq> most thin clients that need swap, probably also have very limited cpu speed
| |
13:11 | <vagrantc> jammcq: minimal increase.
| |
13:11 | <ogra> jammcq, what does nbd swapping to performance ? does it even matter ?
| |
13:12 | <vagrantc> besides, as long as we make it an optional feature... :)
| |
13:12 | <jammcq> swapping over the network already sucks pretty bad. I'm just wondering how much worse it sucks when encrypting the stream
| |
13:13 | <vagrantc> well, if you're swapping regularly, it sucks. but much of the time, you only swap a chunk of data at the beginning, and then it sits in swap... no?
| |
13:13 | <ogra> yep
| |
13:13 | <vagrantc> if you're constantly swapping, it's bad news no matter if it's network or disk.
| |
13:13 | <ogra> until you really run out of main memory ...
| |
13:13 | right
| |
13:14 | * ogra goes to walk SF streets a bit as long as the weather is good ... bbl | |
13:14 | <vagrantc> ogra: it looks like the inputattach code is no longer needed in debian- dexconf properly supports the mouse configuration options.
| |
13:15 | <ogra> then move it out of the debian code ... we'll keep it in ubuntu
| |
13:16 | <vagrantc> ogra: i have a few patches for inputattach code you'll probably want, anyways.
| |
13:16 | <ogra> oki
| |
13:16 | <vagrantc> ogra: but go walk :)
| |
13:16 | <ogra> i'll start merging and setting up a feisty repo during the week ...
| |
13:17 | we're supposed to do regular work this week here
| |
13:17 | <vagrantc> we've managed to diverge a bunch again :)
| |
13:17 | <ogra> yep, as usual :)
| |
13:18 | see you ...
| |
13:19 | * vagrantc is getting really annoyed at http://bugs.debian.org/354669 | |
13:20 | Kodex has joined #ltsp | |
13:23 | <Kodex> Ready to blow my head off trying to build a bootsplash against a 2.6.17 through a pxe boot. Anyone tryed yet? Help? Pleeeease?
| |
13:24 | J45p3r_ has quit IRC | |
13:28 | <Kodex> Can't use LBE on my devel machine. I'm trying to patch a kernel initrd made off the kernel kit. Has anyone succeeded in this?
| |
13:35 | diego_ has joined #ltsp | |
13:35 | <Kodex> Ok. I've already tried the lists, forums, google and the all mighty lord of the cyber-universe and stuff. Sadly enough, I have to say that irc is my last and only resource...
| |
13:35 | <diego_> hi Kodex
| |
13:35 | <Kodex> Hi, diego_
| |
13:36 | <jammcq> Kodex: what's the problem you are having?
| |
13:37 | chupacabra has joined #ltsp | |
13:38 | flavioveras has joined #ltsp | |
13:39 | <Kodex> My kernel panics with: [...] Unable to mount root fs on unknown-block[0,1]
| |
13:39 | <jammcq> Kodex: did you start with one of my .config files for the kernel?
| |
13:39 | <Kodex> Sorry -- Hi, jammcq, and thanks.
| |
13:39 | Yes, and made oldconfig also.
| |
13:39 | <flavioveras> Hi All. What should I do if I want to change from inetd to xinetd ?
| |
13:40 | <Kodex> I'm using 2.6.17.8
| |
13:41 | Already tryed with 2.6.15, with no luck at all.
| |
13:42 | <jammcq> Kodex: what do you have in your pxelinux.cfg/default file?
| |
13:43 | <Kodex> For the splash option: append init=/linuxrc rw root=/dev/ram0 initrd=initramfs-splash.gz video=vesafb:mttr vga=791 splash=silent
| |
13:45 | <jammcq> 'init=/linuxrc' is wrong
| |
13:45 | <Kodex> jammcq: My kernel also says -> No filesystem could mount root, tryed: ext2.
| |
13:45 | <jammcq> take that out
| |
13:45 | linuxrc is from the older initrd, we're using initramfs now
| |
13:45 | <Kodex> But this only happens after I run splash on the initrd...
| |
13:46 | Already tryed with and without, still will not boot
| |
13:49 | <diego_> hi jammcq
| |
13:49 | <jammcq> first thing i'd do is try building a standard kernel, without bootsplash support,and just make sure you can get that to boot properly. Then, once you get booting working, add in the bootsplash
| |
13:50 | diego_: hello
| |
13:50 | <diego_> jammcq: I am a bit disappointed with the sound capabilities
| |
13:51 | I tought we were further
| |
13:51 | (by the way I am acesuares)
| |
13:51 | <Kodex> jammcq: The kernel is booting great! All the machines are working properly with it. This will only happens after I run 'splash' on the iniramfs.gz
| |
13:51 | <jammcq> diego_: sound is a tough issue that needs love from the distros. we're now getting there with Ubuntu
| |
13:51 | Kodex: I don't know what "run 'splash'" does
| |
13:52 | does it muck up the ramdisk ?
| |
13:52 | <diego_> jammcq: okay so I guess that new thing (portaudio or something?) is going to do it ?
| |
13:52 | <jammcq> pulse audio
| |
13:52 | we worked with pulse audio and gstreamer people last week in mountain view
| |
13:52 | <diego_> ah ja. I am trying to use it in schools but most softw is using lots of sounds (spelling and stuff).
| |
13:53 | good to hear that. At the moment it's not very usable. In gnome I can't get anything to work except system sounds
| |
13:53 | one nice thingt though: win98 under win4lin, all sounds in win98 goes through the same 'channel'
| |
13:54 | <Kodex> It adds a bootsplash to a gziped initrd. You can reference: http://www.bootsplash.org/Userspace
| |
13:54 | <diego_> so if you have sound working for one, then for all.
| |
13:54 | <jammcq> Kodex: is it expecting an initrd or initramfs ?
| |
13:54 | <diego_> I am very happy with the possibility of localdev in the new edubuntu,
| |
13:54 | <jammcq> they are VERY different
| |
13:54 | <diego_> so I hope in .04 we will see good sound stuff.
| |
13:56 | <Kodex> jammcq: initrd, I guess, from the patch in: http://wiki.ltsp.org/twiki/pub/Ltsp/BootsplashPatch/
| |
13:57 | I tryed to apply it to a iniramfs. That was the wrong, right?
| |
13:57 | <jammcq> yeah
| |
13:57 | initramfs is a cpio archive
| |
13:57 | initrd is a filesystem image
| |
14:00 | <Kodex> Yes, I was already trying to extract with cpio to check for differences. Is there anyway to patch (with splash or any other utility) the new initramfs. It seems the splash utility uses the gziped part of the file to store the image.
| |
14:00 | <jammcq> no clue, i'm not a fan of splash screens
| |
14:00 | I prefer to see the boot messages
| |
14:02 | <Kodex> Yes, I do also. And It can be done just by pressing a key. Thanks for your help anyway, jammcq. I'll keep trying with the bootsplash guys.
| |
14:13 | <flavioveras> hello jammcq
| |
14:14 | jammcq: DHCP is working
| |
14:14 | TFTP is working
| |
14:14 | NFS is working
| |
14:14 | But I am still getting from the clients an error like that:
| |
14:14 | Me: 10.1.1.133 :/tptpboot/lts/2.6.16.1-ltsp-2/pxelinux.0 . TFTP error 1 (File not found)
| |
14:14 | Unable to load file
| |
14:14 | Any help will be great
| |
14:16 | <jammcq> flavioveras: what distro?
| |
14:17 | <flavioveras> it is a debian (kurumin)
| |
14:18 | <jammcq> do this: dpkg -l | grep tftp
| |
14:22 | <flavioveras> I've got ii tftp-hpa 0.43-1 HPA's tftp client
| |
14:22 | ii tftpd-hpa 0.43-1
| |
14:24 | <jammcq> ok, do this:
| |
14:24 | grep tftp /etc/inetd.conf
| |
14:25 | <Kodex> I've done it, jammcq. Couldn't done without your reminder about the cpio archive. As soon as I can, I'll post the instructions to the list. Thank you so much!
| |
14:25 | <jammcq> Kodex: cool. Now you don't have to blow your head off. I've done good today :)
| |
14:25 | <Kodex> :D
| |
14:26 | <flavioveras> jammc: I've got: #tftp dgram udp wait nobody /usr/sbin/tcpd /usr/sbin/in.tftpd /boot
| |
14:26 | # Esta linha ativa o tftpd "normal":
| |
14:26 | #tftp dgram udp wait nobody /usr/sbin/tcpd /usr/sbin/in.tftpd /tftpboot
| |
14:26 | #tftp dgram udp wait root /usr/sbin/in.tftpd -s /tftpboot -r blksize
| |
14:26 | tftp dgram udp wait root /usr/sbin/in.tftpd /tftpboot -r blksize
| |
14:28 | Kodex has quit IRC | |
14:28 | Skarmeth has quit IRC | |
14:28 | <jammcq> flavioveras: check the spelling of 'tftpboot'. In your error you posted, you have 10.1.1.133 :/tptpboot
| |
14:35 | <flavioveras> it is 10.1.1.254:/tftpboot
| |
14:38 | <patrickb> how can i capture the output from a thin client in shell mode to a file on the server?
| |
14:44 | EversonCaratti has quit IRC | |
14:48 | <flavioveras> Me: 10.1.1.133, DHCP: 10.1.1.254, TFTP: 10.1.1.254, GATEWAY 10.1.1.1 Loadin 10,
| |
14:48 | dhcp.conf
| |
14:48 | # terminal 1:
| |
14:48 | host ws001 {
| |
14:48 | hardware ethernet 00:40:63:C3:5A:39;
| |
14:48 | fixed-address 10.1.1.133;
| |
14:48 | filename "/lts/2.6.16.1-ltsp-2/pxelinux.0";
| |
14:48 | I try to use this on the inedt.conf
| |
14:48 | tftp dgram udp wait root /usr/sbin/in.tftpd -s /tftpboot -r blksize
| |
14:48 | tftp dgram udp wait root /usr/sbin/in.tftpd /tftpboot -r blksize
| |
14:48 | one at a time
| |
14:58 | jammcq:
| |
14:58 | When I use on inetd.conf
| |
14:58 | tftp dgram udp wait root /usr/sbin/in.tftpd -s /tftpboot -r blksize #(tftpd-hpa)
| |
14:58 | I've got:
| |
14:58 | Me 10.1.1.133, DHCP: 10.1.1.254, TFTP: 10.1.1.254, GATEWAY 10.1.1.1
| |
14:58 | Loading 10.1.1.254:/lts/2.6.16.1-ltsp-2/pxelinux.0 . TFTP error 1 (File not found)
| |
15:00 | <McPolo> It's always easy to test that tftp is working with tftp client.
| |
15:01 | <cliebow_> i wonder whether your -s /tftpboot is being respected
| |
15:02 | <flavioveras> Hi McPolo: I did it with netstat -anp | grep ":69"
| |
15:02 | <McPolo> flavioveras: Ofcourse, but that means you have tftpd working, not that your filenames etc. are right.
| |
15:03 | Necessarily right, we had problems with those so somebody else might also :)
| |
15:04 | <cliebow_> you "could give the comlete pathname to your kernel in dhcpd.conf..and comment out the -s argument..and restart inet
| |
15:04 | <flavioveras> clidebow: I did that too
| |
15:07 | jammcq: files name are correct i have allready checked it out
| |
15:09 | McPolo: sorry, files name are correct i have allready checked it out
| |
15:10 | <cliebow_> flavio->do you have next-server statement in dhcpd.conf?
| |
15:10 | <flavioveras> clidebow_: yes I do
| |
15:12 | cliebow:
| |
15:12 | option root-path "10.1.1.254:/opt/ltsp/i386";
| |
15:12 | next-server 10.1.1.254;
| |
15:13 | g333k has quit IRC | |
15:14 | <cliebow_> my line from edgy ..tftp dgram udp wait root tftp dgram udp wait root /usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /var/lib/tftpbootpasti/usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /var/lib/tftpboot
| |
15:14 | cripes
| |
15:14 | tftp dgram udp wait root /usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /var/lib/tftpboot
| |
15:15 | <flavioveras> cliebow:my line tftp dgram udp wait root /usr/sbin/in.tftpd -s /tftpboot -r blksize
| |
15:16 | <cliebow_> seems just fine
| |
15:19 | <flavioveras> cliebow: yuo said i wonder whether your -s /tftpboot is being respected | i've been trying with s and without s
| |
15:22 | diego_ has quit IRC | |
15:22 | <flavioveras> what should i do if i want to use xinetd instead inetd?
| |
15:24 | <cliebow_> i had troublewith this too when i moved to ubuntu...i'm sure you are looking at http://wiki.ltsp.org/twiki/bin/view/Ltsp/TFTP
| |
15:26 | ogra has quit IRC | |
15:26 | <flavioveras> I don't know, but i gess the problem could be that! I gonna try the wiki
| |
15:27 | <cliebow_> for the hell of it try reversing order of options..so -s is last
| |
15:29 | <flavioveras> I did it :( allready
| |
15:52 | <cliebow> flavio: nothing wonky in your file perms on the tftpboot flder?
| |
16:16 | <flavioveras> cliebbow: unfortunately, yes :(
| |
16:17 | i guess is better take a shot using opensuse
| |
16:18 | bobby_C has joined #ltsp | |
16:28 | vagrantc has quit IRC | |
16:28 | muh2000 has quit IRC | |
16:28 | BadMagic has quit IRC | |
16:28 | floydwil1e has quit IRC | |
16:28 | bryn has quit IRC | |
16:28 | oh207 has quit IRC | |
16:28 | robbie has quit IRC | |
16:28 | yanu has quit IRC | |
16:28 | teun has quit IRC | |
16:28 | arc_ has quit IRC | |
16:28 | pimpministerp has quit IRC | |
16:28 | jammcq has quit IRC | |
16:28 | pscheie has quit IRC | |
16:28 | kwak has quit IRC | |
16:28 | joebaker has quit IRC | |
16:28 | str4nd has quit IRC | |
16:28 | Jeeves_ has quit IRC | |
16:28 | vagrantc has joined #ltsp | |
16:28 | arc_ has joined #ltsp | |
16:28 | muh2000 has joined #ltsp | |
16:28 | pimpministerp has joined #ltsp | |
16:28 | jammcq has joined #ltsp | |
16:28 | pscheie has joined #ltsp | |
16:28 | robbie has joined #ltsp | |
16:28 | teun has joined #ltsp | |
16:28 | joebaker has joined #ltsp | |
16:28 | kwak has joined #ltsp | |
16:28 | yanu has joined #ltsp | |
16:28 | bryn has joined #ltsp | |
16:28 | oh207 has joined #ltsp | |
16:28 | BadMagic has joined #ltsp | |
16:28 | floydwil1e has joined #ltsp | |
16:28 | Jeeves_ has joined #ltsp | |
16:28 | str4nd has joined #ltsp | |
16:40 | flavioveras has quit IRC | |
17:06 | flavioveras has joined #ltsp | |
17:32 | rodarvus has joined #ltsp | |
17:32 | rodarvus has left #ltsp | |
17:33 | bobby_C has quit IRC | |
17:50 | <robbie> mornings
| |
18:19 | <cliebow_> mornings
| |
18:30 | moose5435 has quit IRC | |
19:20 | MrMoo has quit IRC | |
19:27 | ogra has joined #ltsp | |
19:38 | cliebow_ has quit IRC | |
19:51 | jcbcotta has joined #ltsp | |
19:51 | <jcbcotta> hello people
| |
19:52 | I alredy come here searching to this question ..but didn't find nothing to my
| |
19:52 | maybe now somebody can help me
| |
19:53 | my ltsp is in function ..but the keyboard of the clients is very fast
| |
19:54 | the repetition is crazy
| |
19:54 | someone can help?
| |
19:54 | jcbcotta is now known as jcbcotta_br | |
19:55 | <jcbcotta_br> no one?
| |
19:55 | jcbcotta_br has quit IRC | |
19:57 | J45p3r has joined #ltsp | |
20:11 | moose5435 has joined #ltsp | |
20:43 | arc_ has quit IRC | |
20:46 | psycodad has quit IRC | |
20:56 | psycodad has joined #ltsp | |
21:05 | J45p3r has quit IRC | |
21:28 | ogra has quit IRC | |
21:36 | sbalneav has joined #ltsp | |
21:36 | <sbalneav> Evening all
| |
21:36 | <jammcq> Scotty !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
| |
21:37 | <sbalneav> Hey jammcq
| |
21:37 | * jammcq is busy ripping all his cds | |
21:37 | <sbalneav> :)
| |
21:37 | <jammcq> and there's alot of them
| |
21:37 | <vagrantc> evenin folks.
| |
21:37 | <jammcq> hey vagrantc
| |
21:37 | <sbalneav> Evening vagrantc!
| |
21:38 | * vagrantc is stumbling through the new ion3 defaults | |
21:40 | <vagrantc> sbalneav: how was the drinking conference?
| |
21:44 | <sbalneav> Excellent!
| |
21:44 | We missed you, however!
| |
21:45 | <vagrantc> i was wishing to be there, but it just wasn't working out.
| |
21:45 | scheming, even.
| |
21:46 | <sbalneav> We'll get you next time, Gadget! Next time!!!
| |
21:47 | <vagrantc> i'm planning on staying in one place for the next few months, at least :)
| |
21:51 | <sbalneav> Great! Found a good gig? Is your sweetie with you?
| |
21:51 | <vagrantc> i'm living in a 300 year old carriage house :)
| |
21:51 | <sbalneav> Cool!!!
| |
21:51 | Got any pictures?
| |
21:52 | <vagrantc> i'll take some :)
| |
21:52 | <sbalneav> Hey! Here's an X question.
| |
21:52 | <vagrantc> just got my partner's camera working with linux
| |
21:52 | <sbalneav> Lets say I have an X server running on a machine, say a thin client.
| |
21:52 | I know it's on :0
| |
21:53 | But, if I'm on <alt><f2>, a text window, and I say "somexprogram -display :0", it doesn't work, because I don't have an Xauthority file locally.
| |
21:53 | Now...
| |
21:53 | <jammcq> that's right
| |
21:54 | <sbalneav> What I THINK is the proper way is:
| |
21:55 | generate an md5 sum from some random bit of info, create an auth file with it, using xauth, and then start the X server with -auth /path/to/file, so I can, somewhere else, set XAUTHORITY=/path/to/file, and make it work.
| |
21:55 | Is this the "proper" way to do this?
| |
21:55 | <jammcq> umm, there's more in that file than just an md5sum
| |
21:56 | <vagrantc> that's what we do in the debian implementation, except the /path/to/file is /tmp/$RANDOM_FILENAME
| |
21:56 | really, it should be $HOME/.Xauthority.
| |
21:56 | but ogra and i have fought over this one numerous times.
| |
21:56 | <sbalneav> That's what I said.
| |
21:57 | I should be $HOME/.Xauth.
| |
21:57 | ok, I'll twist his arm :)
| |
21:57 | * jammcq doesn't understand | |
21:57 | <vagrantc> $HOME/.Xauthority, according to the man page
| |
21:57 | <sbalneav> jammcq: Here's the thing
| |
21:57 | vagrantc: yeah, I just shortend
| |
21:58 | If I'm going to be setting X properties locally, the program needs to have both a valid display, and a valid Xauthority cookie.
| |
21:58 | <jammcq> yeah, that much I know
| |
21:59 | <sbalneav> $HOME for root isn't writable in LTSP5, so, .Xauthority can't be created by the X server.
| |
21:59 | <jammcq> yeah, that's true too
| |
22:00 | <sbalneav> Therefore, any program running locally can't connect, if it's started locally.
| |
22:00 | unless it sets XAUTHORITY environment variable to a valid file.
| |
22:01 | <jammcq> so you and vagrantc are in agreement? that's the part i'm fuzzy on
| |
22:01 | <sbalneav> Yes.
| |
22:01 | <jammcq> ah, k
| |
22:01 | <sbalneav> We should bind mount /root/.Xauthority.
| |
22:01 | <vagrantc> we could make /root/.Xauthority writeable
| |
22:01 | <sbalneav> So the X server can write to it.
| |
22:01 | Yes, absolutely.
| |
22:01 | * vagrantc notes how in agreement we are | |
22:02 | <jammcq> but putting .Xauthority in /root, doesn't that mean we'll be reading it via NFS ?
| |
22:02 | <vagrantc> sbalneav: is this part of a local apps spec or something?
| |
22:02 | jammcq: not in NFS- created at X server run time.
| |
22:03 | <jammcq> 23:00 < vagrantc> we could make /root/.Xauthority writeable
| |
22:03 | <vagrantc> jammcq: if we bind-mount an individual file, /root/.Xauthority will have to be an empty file so we can bind-mount it and write to it.
| |
22:03 | <jammcq> doesn't that mean it'll read the copy on NFS ?
| |
22:03 | why not just leave it in /tmp, and set the XAUTHORITY variable?
| |
22:03 | <vagrantc> read-only nfs, with a tmpfs file bind-mounted to it
| |
22:04 | because most programs expect it in $HOME/.Xauthority
| |
22:04 | <sbalneav> No, not specifically, but I'm working on coming up with a common method of local device notification using Xproperties, so therefore, I need to run X programs locally.
| |
22:04 | <jammcq> most programs SHOULD expect it to be where $XAUTHORITY says it is
| |
22:04 | <vagrantc> and if we leave it in /tmp, it needs to be a randomized file name, which is hard to scurely pass on to all the applications that need to use it.
| |
22:04 | jammcq: yes, but setting $XAUTHORITY properly is difficult to do.
| |
22:06 | <jammcq> oh
| |
22:06 | ok, that's fine for normal LTSP (non-local-apps)
| |
22:06 | what about when we turn on local apps?
| |
22:07 | then, it's the user who is running the apps, so /root/.Xauthority isn't the right place
| |
22:07 | <dberkholz> vagrantc: hey, do you know whether there's an easy way to set up a debian chroot?
| |
22:07 | <vagrantc> let the *dm handle the server-side
| |
22:07 | <dberkholz> i have no clue what a lot of these commands do, which makes them difficult to port to gentoo variants =P
| |
22:07 | <vagrantc> dberkholz: on a non-debian machine?
| |
22:07 | <dberkholz> vagrantc: yep
| |
22:08 | <vagrantc> dberkholz: is there a port of debootstrap or cdebootstrap for gentoo ?
| |
22:08 | <dberkholz> some base-system tarball i could just drop in place
| |
22:08 | <jammcq> vagrantc: 'server-side' ? is that the X-server-side, or the LTSP-server-side ?
| |
22:08 | <dberkholz> uh, looks like debootstrap is available
| |
22:08 | <vagrantc> jammcq: LTSP-server-side. sorry.
| |
22:09 | * jammcq doesn't see how that would work | |
22:09 | <vagrantc> it's already working.
| |
22:09 | <dberkholz> i'll try to look up how to use it for a chroot install
| |
22:09 | <jammcq> the DM is LDM, does that take care of Xauthority ?
| |
22:09 | <vagrantc> jammcq: ssh -X takes care of the LTSP-server-side ~/.Xauthority
| |
22:09 | <jammcq> right, but when you run local apps, you aren't using -X
| |
22:09 | <vagrantc> jammcq: and LDM takes care of the thin-client's Xauthority file
| |
22:10 | <jammcq> vagrantc: but where does LDM put the thin-clients Xauthority file?
| |
22:10 | <vagrantc> jammcq: in a randomized file name in /tmp
| |
22:10 | <jammcq> owned by root? or by the user?
| |
22:10 | <vagrantc> jammcq: it should go in a standardized place, and /root/.Xauthority would be proper, seeing as LDM is run as root.
| |
22:11 | <jammcq> but the apps aren't run by root
| |
22:11 | <vagrantc> ok, so we could put it in some other standard location... /var/cache/ldm or something.
| |
22:12 | but not /tmp.
| |
22:13 | <jammcq> what's the difference between /tmp and /var/cache/ldm ?
| |
22:14 | <vagrantc> because in /tmp/, all users have write access, which gives security implications- like some other user could make a symlink to your file to a file with different permissions and steal your Xauthority data.
| |
22:14 | so you have to randomize filenames in /tmp ...
| |
22:14 | which makes it very hard to figure out what Xauthority to use for local apps.
| |
22:15 | unless LDM could figure out what the Xauthority file is, and pass it to the client, so that it knows what it's Xauthority file is... that would probably be fine, too.
| |
23:17 | <kwak> hi, how do i change graphical login screen theme
| |
23:20 | <sbalneav> Are you using gdm?
| |
23:21 | <kwak> sbalneav, thanks, but i got it.
| |
23:31 | sbalneav, do you know how to change the "foot" icon of gnome main menu? i tried doing it with ggconf-editor, but didn't work
| |
23:33 | <sbalneav> No, sorry, theming's not something I know how to do.
| |
23:33 | <kwak> ok.
| |
23:33 | thanks anyway
| |
23:33 | are you using k12ltsp?
| |
23:33 | brb. lunch time
| |
23:55 | moose5435 has quit IRC | |