IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 28 March 2008 (all times are UTC)
| 00:00 | <dberkholz> johnny: git
| |
| 00:00 | <johnny> uggh.. the evil git
| |
| 00:00 | captain_magnus has joined #ltsp | |
| 00:00 | <johnny> i do have it installed tho
| |
| 00:01 | <dberkholz> johnny: drop me an email with a dsa ssh key attached, and we'll begin getting you set up
| |
| 00:01 | <johnny> ok
| |
| 00:01 | will do that in a few
| |
| 00:01 | <dberkholz> i need to fwd that on to the overlays admins, and then they'll have to get you set up with an account
| |
| 00:02 | <johnny> aha.. so.. the waiting game again
| |
| 00:02 | guess i should sent it to you earlier if you needed it
| |
| 00:02 | <dberkholz> they're usually pretty decent about it
| |
| 00:02 | well i didn't get the overlay till a couple days ago
| |
| 00:03 | <johnny> now that i stepped away from ltsp for a bit, i'm sure i'll be a bit clearer now
| |
| 00:03 | just need to get the initramfs setup, and then do the init scripts
| |
| 00:03 | after that, should be pretty close to booting it
| |
| 00:03 | booting it properly that is
| |
| 00:04 | i have it booting now, it just needs the unionfs or aufs stuff
| |
| 00:06 | indradg has quit IRC | |
| 00:07 | <johnny> dberkholz, want me to send to your gentoo address then?
| |
| 00:08 | <dberkholz> johnny: yes please
| |
| 00:08 | <vagrantc> you don't need to implement unionfs/aufs ...
| |
| 00:08 | at least, for the most simple case.
| |
| 00:23 | subir has joined #ltsp | |
| 00:29 | <johnny> you mean nfsroot?
| |
| 00:29 | or?
| |
| 00:31 | i tend to work on part of a project until i get stuck, and then i step away from that part and move onto something else for a lil while
| |
| 00:31 | and then return to it later
| |
| 00:38 | <vagrantc> johnny: you can use tmpfs bind mounts with any root filesystem
| |
| 00:38 | johnny: i.e. i've been using it with NBD+ext2+tmpfs bind mounts
| |
| 00:38 | because squashfs and unionfs have been so unstable in debian
| |
| 00:39 | haven't tried aufs yet
| |
| 00:39 | <johnny> unionfs has a problem for me
| |
| 00:39 | in any case
| |
| 00:39 | it's not available except as a kernel patch anymore
| |
| 00:39 | it's in mm, but won't be merged into 2.6.25
| |
| 00:40 | aufs can still be compiled as an external module
| |
| 00:40 | <vagrantc> well, if you just want to get ltsp working, it would be easy to not use unionfs
| |
| 00:40 | or any union sort of filesystem
| |
| 00:40 | that's the point i'm making :)
| |
| 00:40 | <johnny> sure
| |
| 00:40 | i think i need a bit of advice on that in any case
| |
| 00:40 | <vagrantc> and i should instead be sleeping.
| |
| 00:40 | <johnny> oh :(
| |
| 00:41 | <vagrantc> johnny: there's some simple code in client/initscripts/ltsp-setup ...
| |
| 00:51 | <dberkholz> johnny: you weren't around when i mentioned this last, i think, but other option is fuse unionfs
| |
| 00:51 | <johnny> i was , and i was thinking about it
| |
| 00:51 | wasn't sure if it was any good after reading the pages
| |
| 00:51 | i found funionfs and unionfs-fuse
| |
| 00:52 | sothere are two fuse unionfs implmentations
| |
| 00:52 | <dberkholz> pick the one other people use
| |
| 00:52 | unionfs-fuse sounds familiar
| |
| 00:52 | <johnny> looks like neither :)
| |
| 00:53 | i don't know anybody personally using either one of those two
| |
| 00:53 | i'm going to look into their mailing lists next
| |
| 00:53 | i was investigating this before i left town on monday
| |
| 00:55 | <dberkholz> ah, the unionfs-fuse guy is a gentoo user
| |
| 00:56 | i knew his name sounded familiar
| |
| 00:56 | <johnny> i tried to get online when i was at the hotel.. it was nonsense tho:(
| |
| 00:56 | <dberkholz> so you could probably get good support =)
| |
| 00:56 | <johnny> i'm going to go with vagrantc's suggestion for the moment
| |
| 00:56 | we can investigate that later
| |
| 00:56 | i'm going to clean up the initscript to be more gentooish
| |
| 00:57 | <dberkholz> johnny: sounds nice. vagrantc's mentioned some ideas of making it plugin-based like the client builder
| |
| 00:57 | <cyberorg> any kind of unionfs is out, as no distro kernel maintainers are happy about them
| |
| 00:57 | <dberkholz> johnny: so if it looks like almost everything is the same, then we could look into just maintaining differences rather than a whole separate script where we need to refix the same bugs
| |
| 00:58 | <johnny> i agree dberkholz
| |
| 00:58 | <dberkholz> johnny: so i'd try to hold off on too much gentooifying, just the minimum necessary
| |
| 00:59 | <johnny> shouldn't be much work
| |
| 01:00 | <dberkholz> johnny: not in the interests of work, in the interests of merging things together later
| |
| 01:00 | <johnny> i mean to make it work at all
| |
| 01:00 | <dberkholz> oh, alright
| |
| 01:04 | <vagrantc> warren split out a bunch of functions ... so hopefully you can just whip together a gentoo-like init script that calls the appropriate functions ...
| |
| 01:04 | <dberkholz> vagrantc: nifty.
| |
| 01:04 | <vagrantc> although i'd hesitate to do a plugin-based system for initscripts, since that's a bit filesystem intensive
| |
| 01:04 | <dberkholz> about the only significant difference should be how dependencies are specified
| |
| 01:04 | vagrantc: sure ,if it's just one file ltsp/plugins.sh or whatever ,that's fine too
| |
| 01:05 | man i suck at putting commas before spaces
| |
| 01:06 | <vagrantc> although, only one of the two init scripts used by debian actually has functions split out ...
| |
| 01:10 | <johnny> which ones do you use?
| |
| 01:11 | <vagrantc> both
| |
| 01:11 | client/initscripts/ltsp-core and client/initscripts/ltsp-setup
| |
| 01:11 | one does the configuration stuff, one actually starts the sessions and things
| |
| 01:11 | oh yeah, sleep.
| |
| 01:12 | i'd best be out of here :)
| |
| 01:12 | vagrantc has quit IRC | |
| 01:17 | praveer_cool has quit IRC | |
| 01:27 | cyberorg has quit IRC | |
| 01:38 | cyberorg has joined #ltsp | |
| 01:49 | elisboa has quit IRC | |
| 01:49 | elisboa has joined #ltsp | |
| 02:07 | soneyka has joined #ltsp | |
| 02:43 | plamengr has joined #ltsp | |
| 02:44 | Pascal_1 has joined #ltsp | |
| 03:05 | <Pascal_1> salut
| |
| 03:07 | gonzaloaf has quit IRC | |
| 03:08 | gonzaloaf has joined #ltsp | |
| 03:26 | <daduke> Pascal_1: bonjour
| |
| 03:27 | <Pascal_1> bonjour !! ;-)
| |
| 03:56 | bobby_C has joined #ltsp | |
| 03:58 | Egyptian[Home] has quit IRC | |
| 03:58 | Egyptian[Home] has joined #ltsp | |
| 04:16 | tux_440volt has joined #ltsp | |
| 04:23 | ffhghgf has joined #ltsp | |
| 04:23 | soneyka has quit IRC | |
| 04:42 | mikkel has joined #ltsp | |
| 04:44 | spectra has joined #ltsp | |
| 04:55 | spectra has quit IRC | |
| 04:59 | Pascal has joined #ltsp | |
| 05:02 | Pascal has quit IRC | |
| 05:16 | praveer_cool has joined #ltsp | |
| 06:01 | subir has quit IRC | |
| 06:03 | <Pascal_1> klausade, ?? have you got some news about my problem ?
| |
| 06:04 | did you test my chroot ?
| |
| 06:05 | * gvy is back (gone 15:41:34) | |
| 06:06 | J45p3r has joined #ltsp | |
| 06:06 | praveer_cool has quit IRC | |
| 06:13 | Pascal_1 has quit IRC | |
| 06:26 | TelnetManta has quit IRC | |
| 06:28 | Egyptian[Home] has quit IRC | |
| 06:36 | hansatuofd has joined #ltsp | |
| 06:38 | hansatuofd has quit IRC | |
| 06:38 | Q-FUNK has joined #ltsp | |
| 06:39 | hansatuofd has joined #ltsp | |
| 06:45 | J45p3r has quit IRC | |
| 06:49 | hansatuofd has quit IRC | |
| 06:50 | hansatuofd has joined #ltsp | |
| 06:54 | cliebow has joined #ltsp | |
| 07:13 | Guaraldo has joined #ltsp | |
| 07:13 | cyberorg has quit IRC | |
| 07:14 | otavio has joined #ltsp | |
| 07:14 | tux_440volt has quit IRC | |
| 07:19 | Guaraldo has quit IRC | |
| 07:19 | TelnetManta has joined #ltsp | |
| 07:20 | Guaraldo has joined #ltsp | |
| 07:20 | <cliebow> botnet takes advantages of holes in Windows operating systems -- pretty much every version of Windows, excluding Windows Server 2003.
| |
| 07:20 | Guaraldo has quit IRC | |
| 07:20 | Guaraldo has joined #ltsp | |
| 07:43 | Blinny has joined #ltsp | |
| 07:44 | praveer_cool has joined #ltsp | |
| 07:54 | Pascal_1 has joined #ltsp | |
| 07:57 | Pascal_1 has quit IRC | |
| 07:59 | praveer_cool has quit IRC | |
| 08:01 | Pascal_1 has joined #ltsp | |
| 08:07 | <gvy> cliebow, looks like a boring headline :)
| |
| 08:08 | slidesinger has joined #ltsp | |
| 08:08 | <cliebow> heh..botnet is not unlike skynet..soon it will become self aware..I"LL BE BACK..
| |
| 08:08 | mhterres has joined #ltsp | |
| 08:08 | bobby_C has quit IRC | |
| 08:11 | t-kid has joined #ltsp | |
| 08:11 | <t-kid> helow
| |
| 08:12 | t-kid has quit IRC | |
| 08:19 | score has joined #ltsp | |
| 08:20 | pascal_2 has joined #ltsp | |
| 08:21 | <score> hi, i'm running gutsy's ltsp packages. it seems like over time, i've ended up with nbdrootd and nbd-server processes running when i only have around 60 thin clients. any idea whether this is a bug or normal?
| |
| 08:23 | alekibango has quit IRC | |
| 08:31 | <laga> score: i dont see what's wrong with thatr
| |
| 08:33 | <score> sorry, i guess i left out the number of processes .. 400
| |
| 08:35 | <laga> score: ok, that's quite a lot.. you should talk to ogra_cmpc
| |
| 08:38 | pascal_2 has quit IRC | |
| 08:38 | <ogra_cmpc> score, grep nbd /etc/hosts.allow
| |
| 08:38 | does that return anything ?
| |
| 08:38 | <gvy> ogra_cmpc, ! :)
| |
| 08:38 | <ogra_cmpc> hey
| |
| 08:39 | <score> ogram_cmpc: nbdrootd: ALL: keepalive
| |
| 08:39 | <ogra_cmpc> good
| |
| 08:39 | it should clean up the stale nbdroot processes within 10h or so usually
| |
| 08:39 | <score> ogra_cmpc: these have been running since march 10th
| |
| 08:39 | <ogra_cmpc> not sure ehat tcpd uses as default atm
| |
| 08:40 | thats strange
| |
| 08:40 | Egyptian[Home] has joined #ltsp | |
| 08:40 | <ogra_cmpc> it definately cleans up here
| |
| 08:41 | <score> ogra_cmpc: is there a way i can distinguish which are actually in use?
| |
| 08:42 | <ogra_cmpc> probably with a netstat command or so, sorry i'm extremly busy with work atm
| |
| 08:45 | <score> ogra_cmpc: that's fine. running netstat and sorting the results shows that there are a few conections on 2000 from the same ips.
| |
| 08:45 | i see one that has 7.. seems bad
| |
| 08:48 | dtrask has joined #ltsp | |
| 08:50 | steph_ has joined #ltsp | |
| 08:54 | <cliebow> dtrask: know od=f anything entry level for a 22 year old?
| |
| 08:58 | <dtrask> cliebow: od=f?
| |
| 08:58 | wazzat?
| |
| 09:02 | johnny has quit IRC | |
| 09:04 | tux_440volt has joined #ltsp | |
| 09:05 | <cliebow> /od=f/of
| |
| 09:17 | vagrantc has joined #ltsp | |
| 09:20 | <cliebow> !v
| |
| 09:20 | <ltspbot> cliebow: Error: "v" is not a valid command.
| |
| 09:21 | <gvy> vagrantc, ! :)
| |
| 09:22 | <dtrask> cliebow: in terms of what? Entry level what? OS computer...????
| |
| 09:22 | <gvy> vagrantc, any spare time at hand?
| |
| 09:22 | <dtrask> cliebow: not sure I understand the question
| |
| 09:22 | <cliebow> entry level it
| |
| 09:23 | <dtrask> cliebow: position?
| |
| 09:23 | <cliebow> oh..help desk or something?
| |
| 09:23 | this kid has been through smtc..knows some php..
| |
| 09:23 | <dtrask> yeah...or school tech
| |
| 09:23 | <cliebow> sure..
| |
| 09:24 | <vagrantc> gvy: probably shouldn't get too involved in things at the moment...
| |
| 09:24 | <cliebow> slackware
| |
| 09:24 | <dtrask> a school district tech would be a good place to start....great "school of hard knocks" place to get some good exposure
| |
| 09:24 | <cliebow> thought you might have heard something through the grapevine
| |
| 09:24 | <dtrask> not yet, but will keep ears open....also check servingschools.com
| |
| 09:25 | <cliebow> yep..had him do that..
| |
| 09:25 | thanks
| |
| 09:25 | <dtrask> it's the Maine database for school jobs....including technicians...I think I saw some posted last week
| |
| 09:25 | np welcome
| |
| 09:26 | <gvy> vagrantc, well, if you decide to look over led's branch, it's at r702 branched off 674.1.7 (tagged as 5.1.1)
| |
| 09:27 | vagrantc, led says he's going to maintain a patch to get working package since startx (around xfs/nc) is too broken for him to work around
| |
| 09:27 | tiagovaz has joined #ltsp | |
| 09:27 | <vagrantc> gvy: startx was renamed to xdmcp ...
| |
| 09:27 | <gvy> vagrantc, er, screen-x-common
| |
| 09:27 | <vagrantc> too broken to work around?
| |
| 09:28 | <gvy> yup
| |
| 09:28 | <vagrantc> gvy: i'd be curious to hear more than "it's broken"
| |
| 09:28 | gvy: and why doesn't led communicate directly?
| |
| 09:28 | <gvy> sleeping 5s for each probed xfs (both listening or not) is rather broken :(
| |
| 09:28 | <vagrantc> communicating by proxy seems awkward...
| |
| 09:28 | <gvy> vagrantc, his english is rather read-only i guess
| |
| 09:29 | vagrantc, yup :( imagine how the proxy "likes" that
| |
| 09:29 | vagrantc, so he prefers patches/commits :)
| |
| 09:29 | <vagrantc> gvy: the xfs code is likely to be removed entirely, and isn't enabled by default
| |
| 09:29 | <gvy> "to work around": with variable settings
| |
| 09:29 | <ogra_cmpc> using xfs nowadays is a huge waste, font rendering is done client side for 90% of the apps nowadays
| |
| 09:30 | <gvy> vagrantc, well, we will check if anything breaks without it (xfs being deprecated anyways), but not yet
| |
| 09:30 | ogra_cmpc, yup
| |
| 09:30 | the problem's 10%
| |
| 09:30 | vagrantc, er, more precisely it's netcat acting weirdly to me
| |
| 09:30 | <ogra_cmpc> fix the 10% then :) file upstream bugs
| |
| 09:30 | <gvy> but then brokennes is using netcat. :)
| |
| 09:31 | ogra_cmpc, i've recently contacted pdksh upstream :)
| |
| 09:31 | <ogra_cmpc> :)
| |
| 09:31 | <gvy> a month's turnaround, and "thanks for patches, maybe will hand some day to someone"
| |
| 09:31 | <ogra_cmpc> there are various implementations of netcat
| |
| 09:31 | <gvy> patches over 5 years or so. :)
| |
| 09:31 | <ogra_cmpc> did you try if a different one works better ?
| |
| 09:31 | <gvy> ogra_cmpc, ah, forgot that -- thanks, will try another
| |
| 09:31 | <vagrantc> gvy: if it's the timeout
| |
| 09:32 | gvy: i'd consider lowering the default timeout ...
| |
| 09:32 | <gvy> vagrantc, or that :)
| |
| 09:32 | <vagrantc> gvy: if you're using XFS, though, i'd recommend just setting XFS_SERVER and be done with it
| |
| 09:32 | * ogra_cmpc wonders what nc is used for in the xdmcp script | |
| 09:32 | <gvy> vagrantc, thanks, passing things thaaaaat way....
| |
| 09:33 | seems like bsd netcat here
| |
| 09:33 | <vagrantc> ogra_cmpc: if USE_XFS=True, and XFS_SERVER isn't set, it looks for XFS running on each of your possible login servers.
| |
| 09:33 | <ogra_cmpc> i know we use it in ldm to get the data from ldminfod ... i wasnt aware xdmcp doe use it for anything
| |
| 09:33 | ah
| |
| 09:34 | <vagrantc> so i guess we should set the timeout to 1 second, and encourage people just to set XFS_SERVER if they're actually using XFS.
| |
| 09:34 | if that 1 second delay per server is too much
| |
| 09:34 | * ogra_cmpc curses intel for hardwiring hotkeys in bios ... | |
| 09:34 | <ogra_cmpc> *sigh*
| |
| 09:34 | <gvy> vagrantc, then they should care themselves, yup
| |
| 09:34 | <ogra_cmpc> life could be so easy without silly manufacturer ideas
| |
| 09:34 | <vagrantc> and like ogra_cmpc says, xfs is basically deprecated
| |
| 09:34 | K_O-Gnom has joined #ltsp | |
| 09:34 | <gvy> ogra_cmpc, adding that to acpi and ipmi cursery upon them? :)
| |
| 09:35 | vagrantc, yeah
| |
| 09:35 | <ogra_cmpc> no acpi events from hardwired keys
| |
| 09:35 | neither a way to send any
| |
| 09:35 | <Blinny> ogra_cmpc: dell put a button on the latest batch of laptops we ordered that boots to a separate partition. It's not in the boot order, completely bypasses any boot order directives in the BIOS..
| |
| 09:36 | <vagrantc> gvy: could you register your branch with launchpad to be mirrored?
| |
| 09:36 | <ogra_cmpc> Blinny, well, i could live with that ... for me the prob is that the display doesnt wake up properly without hitting the brightness up key after resume
| |
| 09:36 | <vagrantc> gvy: it seems rather slow to pull
| |
| 09:36 | <ogra_cmpc> which would be easy to solve if i had any handler to send an event to
| |
| 09:36 | <gvy> vagrantc, how should i do that? (i'm registered with lp)
| |
| 09:36 | vagrantc, yeah, very slow but i don't know *why*
| |
| 09:37 | cyberorg has joined #ltsp | |
| 09:37 | <gvy> you can test anything off that ftp, might be 100+kb/s to you
| |
| 09:37 | up to 1Mb/s
| |
| 09:37 | but not that dog slow
| |
| 09:37 | pascal_2 has joined #ltsp | |
| 09:37 | <Blinny> Bah. At least your computer turns on correctly. I've got people who press the power button that is right beside the real power button and then can't do any work. Oh, and it continues to boot into the 2nd partition every time you press the "real" power button, until you press the fake one again, at which point it boots the real partition.
| |
| 09:37 | <ogra_cmpc> gvy, there is an "add branch" under cod on your personal LP page
| |
| 09:37 | <Blinny> Since there's no way to disable it I put a little sticker with a 'X' over the button.
| |
| 09:37 | <ogra_cmpc> *code
| |
| 09:38 | <gvy> ogra_cmpc, thx
| |
| 09:38 | <vagrantc> gvy: https://code.launchpad.net/~vagrantc/+addbranch
| |
| 09:39 | gvy: substituting vagrantc for your username, of course.
| |
| 09:39 | i've gotta ignore irc for a while.
| |
| 09:40 | * ogra_cmpc crawls back into his hole again as well | |
| 09:41 | * gvy probably too to get something done | |
| 09:41 | <cyberorg> vagrantc, i didn't bzr push last commits, please check now :)
| |
| 09:43 | <klausade> Pascal_1: sorry, been busy. I can tell you that my pam-scrip setup that does work fine with ltsp 5.0.8debian3 does not work with your chroot. well not realy fine, seems atm only onauth and onsessionopen works, not onsessionclose. wonder why onsessionclose doens't work.
| |
| 09:43 | <Pascal_1> then same problem than me
| |
| 09:43 | what is your version of ltsp ?
| |
| 09:44 | and what depot you use ?
| |
| 09:45 | oups ltsp 5.0.8debian3
| |
| 09:45 | what is the way to install this version N
| |
| 09:45 | <klausade> Pascal_1: does onauth and onsessionopen work for you?
| |
| 09:45 | <Pascal_1> yes
| |
| 09:45 | only onclose
| |
| 09:45 | doesnt works
| |
| 09:45 | <klausade> Pascal_1: ok.
| |
| 09:45 | don't use my version of ltsp, it's old.
| |
| 09:45 | <Pascal_1> vagrantc, did you see that ?
| |
| 09:45 | there is a version wich works
| |
| 09:46 | <klausade> Pascal_1: i'll be meeting vagrantc next week in spain :-)
| |
| 09:46 | <Pascal_1> shake his hand for me
| |
| 09:46 | <ogra_cmpc> you go to extremadura as well ?
| |
| 09:46 | <Pascal_1> for his great patience about me
| |
| 09:47 | <klausade> ogra_cmpc: yes.
| |
| 09:47 | <ogra_cmpc> cool
| |
| 09:47 | <klausade> ogra_cmpc: you?
| |
| 09:47 | <ogra_cmpc> nah
| |
| 09:47 | trying to cut down travelling time atm
| |
| 09:48 | i'll be in prague end of april ... and likely in seville a month later
| |
| 09:49 | preparing a release until april 24th ... so no time to waste atm
| |
| 09:51 | klausade has left #ltsp | |
| 09:52 | klausade has joined #ltsp | |
| 09:55 | <daduke> hey vagrantc and ogra_cmpc! I'm playing with debian-live now, for the LTSP USB stick, you might remember. I have something that's booting now, and I can start ldm by hand at least, but how does LTSP determine the terminal server to connect to?
| |
| 10:00 | pascal_2 is now known as pascal_ubuntu | |
| 10:00 | pascal_debian has joined #ltsp | |
| 10:01 | plamengr has quit IRC | |
| 10:01 | <vagrantc> daduke: you'll have to set LDM_SERVER in /etc/lts.conf
| |
| 10:02 | daduke: normally, it defaults to detecting it based on your root filesystem ... but if your root filesystem is a CD or USB stick ... well, that just doesn't work :)
| |
| 10:02 | <daduke> vagrantc: ah ok. is this sorta implicit for NFS boot? I don't have it in our main lts.conf
| |
| 10:02 | vagrantc: I see. thanks.
| |
| 10:02 | <vagrantc> daduke: it's only needed if it can't detect it...
| |
| 10:03 | dtrask has quit IRC | |
| 10:03 | <daduke> vagrantc: lemme try then.
| |
| 10:03 | ffhghgf has quit IRC | |
| 10:04 | <ogra_cmpc> it grabs it from the dhsp data normally
| |
| 10:04 | *dhcp
| |
| 10:04 | in combination with the kind of chroot you use
| |
| 10:05 | ffhghgf has joined #ltsp | |
| 10:05 | <daduke> ogra_cmpc: all right. This obviously doesn't work for CD boot ;)
| |
| 10:05 | <ogra_cmpc> your image will neither know anout nbd-server nor about rootserver for nfs
| |
| 10:05 | *about
| |
| 10:05 | <daduke> right
| |
| 10:05 | ffhghgf has quit IRC | |
| 10:06 | <ogra_cmpc> even though, assuming you still use dhcp at some point to get an IP you could still abuse rootserver for it
| |
| 10:07 | <daduke> ogra_cmpc: hmm maybe, but DHCP and ltsp are different servers in our case...
| |
| 10:11 | <gvy> vagrantc, mirrored here: https://code.launchpad.net/~shigorin/ltsp/ltsp-altlinux
| |
| 10:11 | soneyka has joined #ltsp | |
| 10:12 | <gvy> ogra_cmpc, maybe using avahi for autodiscovery might be reasonable some day... at least that's what avahi maintainer suggested here. :)
| |
| 10:12 | <ogra_cmpc> thats trivial
| |
| 10:12 | i have some code on an old used laptop somewhere
| |
| 10:13 | * gvy didn't look into avahi yet, just heard a lot | |
| 10:13 | <ogra_cmpc> but only proof of concept for the python ldm
| |
| 10:25 | pascal_debian has quit IRC | |
| 10:25 | pascal_ubuntu has quit IRC | |
| 10:26 | <Pascal_1> klausade, you didnt told me wich repository you use to install your version of ltsp ?
| |
| 10:27 | <klausade> Pascal_1: from ltsp-backports, but that what last summer.
| |
| 10:27 | <Pascal_1> arf ok
| |
| 10:27 | but i dont understand why it works with an old version
| |
| 10:27 | <vagrantc> gvy, cyberorg: i probably won't have much time to review changes till april 1st
| |
| 10:27 | gvy: maybe as late as april 9th
| |
| 10:28 | <gvy> vagrantc, np
| |
| 10:29 | <vagrantc> cyberorg: i don't see why you fork 001-load-configuration-file ...
| |
| 10:29 | cyberorg: seems like the extra stuff you do could be done in an extra plugin.
| |
| 10:30 | <cyberorg> vagrantc, ok, remove that
| |
| 10:30 | <vagrantc> cyberorg: also, why leave commented code in 020-rootpath ?
| |
| 10:31 | <klausade> Pascal_1: with the old version(ldm written in python) only onaouth and onsessionopen works. Not onsessionclose
| |
| 10:31 | <cyberorg> vagrantc, i explained that, we dont want rootpath created, kiwi will not run if it exists
| |
| 10:31 | <ogra_cmpc> Pascal_1, likely because klausade uses an ancient ldm version
| |
| 10:31 | <Pascal_1> klausade, you told me that it worked for you ?
| |
| 10:31 | <cyberorg> oh, you mean it should be removed
| |
| 10:31 | <vagrantc> cyberorg: yes, but why not delete all the lines in *your* plougin that begin with #
| |
| 10:31 | <klausade> ogra_cmpc: it was better before...
| |
| 10:31 | <ogra_cmpc> before what ?
| |
| 10:31 | <Pascal_1> klausade, no if onsession close doesnt works
| |
| 10:31 | <klausade> ogra_cmpc: before computers
| |
| 10:32 | <ogra_cmpc> heh
| |
| 10:32 | <vagrantc> cyberorg: the original is in the common plugin if you need to diff
| |
| 10:32 | <Pascal_1> klausade, i understood that you make it works with your version
| |
| 10:32 | <klausade> Pascal_1: i thouht it used to work onsessionclose also. Now I'm not sure. But, I'm testing now.
| |
| 10:32 | <Pascal_1> ;-)
| |
| 10:33 | <cyberorg> vagrantc, revision 689
| |
| 10:33 | <Pascal_1> anyway vagrantc told me that he confirmed that ldm doesnt speak to pam on logout
| |
| 10:33 | <vagrantc> cyberorg: and then the ROOT=$LTSNFSPATH ... why not just use ROOT ?
| |
| 10:34 | <klausade> Pascal_1: the funny thing is that onsessionclose works with cron:CRON[5356]: PAM-script: Command is: /etc/security/onsessionclose
| |
| 10:34 | <Pascal_1> arf starnge
| |
| 10:34 | strange
| |
| 10:34 | <cyberorg> vagrantc, we have two destination path NFSPATH and SQUASHPATH
| |
| 10:34 | <Pascal_1> vagrantc, i forgot to ask you if the problem was the same with ubuntu
| |
| 10:35 | <vagrantc> cyberorg: but basically, looks ready to merge.
| |
| 10:35 | <cyberorg> but yes, i can name NFSPATH to ROOT in my scripts
| |
| 10:36 | praveer_fedora has joined #ltsp | |
| 10:37 | <cyberorg> vagrantc, i'd leave it as it is currently, just to keep everything working
| |
| 10:37 | <vagrantc> cyberorg: sure.
| |
| 10:38 | cyberorg: there was some other variable name change thing ... but maybe you fixed that up already?
| |
| 10:38 | <cyberorg> vagrantc, yes, removed those plugins, will be using the ones in common
| |
| 10:38 | <vagrantc> cyberorg: do you have commit access ?
| |
| 10:39 | <cyberorg> vagrantc, nope
| |
| 10:39 | <vagrantc> cyberorg: ok, i'll merge and push, then.
| |
| 10:40 | gvy: i think you have some more invasive changes ... ?
| |
| 10:40 | i.e. outside of just the plugin dir
| |
| 10:40 | <cyberorg> vagrantc, none
| |
| 10:40 | one was Xsession path, don't know what was the outcome of that
| |
| 10:40 | <warren> I think we can give them commit access after we are confident they know if changes could possibly break other distros.
| |
| 10:41 | <vagrantc> cyberorg: i was talking to gvy
| |
| 10:41 | <warren> Also they must know how not to push without renumbering the revision numbers
| |
| 10:41 | <cyberorg> vagrantc, ah, just saw his name :)
| |
| 10:41 | <warren> And if they are not sure if they will break another distro, ask on the list
| |
| 10:41 | <vagrantc> warren: yes. although now that we've upgraded to pack-0.92 ... we *should* be able to set the append_revisions_only flag
| |
| 10:42 | <warren> vagrantc, ah
| |
| 10:42 | <gvy> warren, halo
| |
| 10:42 | ogra_cmpc is now known as ogra | |
| 10:42 | <vagrantc> warren: need to test it, though.
| |
| 10:42 | mccann has joined #ltsp | |
| 10:42 | <warren> if anyone is interested, I'm doing a k12linux devel meeting in #fedora-meeting in about 1 hour 20 minutes.
| |
| 10:42 | <gvy> warren, i'm getting led subscribed to the list, currently he's not ready to push into trunk -- probably after the initial merge will be easier
| |
| 10:43 | that is, cleaning up bit by bit is easier than chunking around...
| |
| 10:43 | <vagrantc> gvy: if we can get one branch per feature, that would make merging a *lot* easier.
| |
| 10:44 | unfortunately, bzr's cherry-picking isn't as good as i'd like. you can do it, but it doesn't keep track of the changes very well.
| |
| 10:44 | <gvy> vagrantc, a minute...
| |
| 10:45 | <vagrantc> unless that's been fixed recently ... but i somehow doubt it
| |
| 10:46 | <cyberorg> seeing all this feedback i feel glad that rest of the civilized world uses git or svn
| |
| 10:47 | <vagrantc> well, bzr is several orders of magnitude better than svn.
| |
| 10:47 | <ogra> and its really a quaetion of how you organize your work if you get advantage of it
| |
| 10:48 | *question
| |
| 10:48 | * ogra wonders if warren reads /msg | |
| 10:51 | <vagrantc> gvy: there's definitely some changes i would reject
| |
| 10:51 | <gvy> vagrantc, could you mention the revisions?
| |
| 10:51 | <vagrantc> gvy: the changes to client/update-kernels BOOTPROMPT_OPTS ...
| |
| 10:51 | gvy: i'm just looking at the whole diff at the moment.
| |
| 10:51 | <gvy> ah
| |
| 10:53 | <vagrantc> gvy: i like the use of /etc/ltsp_chroot to figure out the chroot information ...
| |
| 10:53 | Blinny has quit IRC | |
| 10:53 | <vagrantc> i've been meaning to figure out a standardized way to do that for some time...
| |
| 10:54 | <ogra> wasnt /etc/ltsp_chroot our standarized one ?
| |
| 10:54 | <vagrantc> yes, but we've never used it
| |
| 10:54 | <ogra> i do
| |
| 10:54 | at least in ltsp-client preinst
| |
| 10:55 | <vagrantc> gvy: definitely don't like the changes to ltsp-common-functions ...
| |
| 10:55 | gvy: i refuse to use 0 and 1 as booleans in shell scripts, because the potential for confusion is *huge*
| |
| 10:56 | * warren notes that other upstream developers may have additional requirements on top of what I said earlier. | |
| 10:56 | <vagrantc> gvy: and boolean_is_false could merely be a negation of boolean_is_true.
| |
| 10:56 | gvy: is_yes and is_no just seem pointless.
| |
| 10:57 | <warren> mccann, ping
| |
| 10:57 | <mccann> warren: hey
| |
| 10:57 | <vagrantc> warren: regarding tagging versions, i think we should make a practice of the commit immediately after tagging a version to increment the version in release.conf ...
| |
| 10:58 | <warren> mccann, we have a k12linux devel meetig in #fedora-meeting at 1PM if you are available. Part of it I will be talking about ldm's issues, including ConsoleKit.
| |
| 10:58 | * vagrantc wishes ~ were universally recognized | |
| 10:58 | <warren> vagrantc, you mean before tagging?
| |
| 10:58 | vagrantc, oh I see, so the number of snapshots after is always higher
| |
| 10:58 | <vagrantc> warren: well no, because once you've tagged it, it should start working on the snapshot ... yeah
| |
| 10:59 | <warren> vagrantc, that might be a little confusing to see a checkin of a higher number after...
| |
| 10:59 | <vagrantc> but there's no universal way to suggest snapshots ...
| |
| 10:59 | <warren> vagrantc, none of the other upstreams do it this way that I know of
| |
| 10:59 | <gvy> vagrantc, well you're not forced to use is_* or 0/1 :)
| |
| 10:59 | <ogra> warren, i think th emobile team recently solved it for the startx script in ubuntu, i was planning to pull the same stuff into ldm here
| |
| 10:59 | <mccann> warren: ok I can probably sit in for a bit
| |
| 10:59 | <vagrantc> gvy: but there's *no* point to it.
| |
| 10:59 | <warren> vagrantc, well, in the context of packages, you can stop doing ~. revision numbers denote changes on top of the version number.
| |
| 10:59 | <ogra> (i havent looked how hackish that is though)
| |
| 11:00 | <warren> ogra, using what?
| |
| 11:00 | <ogra> Ck in ldm
| |
| 11:00 | <vagrantc> gvy: if support for it is in the code, people may use it, and i object to allowing those changes.
| |
| 11:00 | <warren> I mean, using what command?
| |
| 11:00 | mccann, I'll talk about ldm and ConsoleKit at the beginning of the meeting then
| |
| 11:00 | <gvy> vagrantc, for us, there's no point in heaps of stuff in ltsp... [ -n "$(echo ... | grep ...)" ] is horrific as well
| |
| 11:00 | <warren> mccann, and we can have minutes to post in public too.
| |
| 11:00 | <ogra> warren, i'll have to dig that up, there was something they backported from the last CK upstream commits
| |
| 11:00 | some script to enable the session
| |
| 11:01 | <warren> ogra, it is all within the startx script or requires external stuff?
| |
| 11:01 | <gvy> vagrantc, and i can provide at least one case *for* using distinct tests for "true" and "false"
| |
| 11:01 | it's when you default to different things
| |
| 11:01 | basically "if yes, or otherwise..."
| |
| 11:01 | and "if no, or otherwise..."
| |
| 11:02 | my shell scripting experience had quite a few of such cases
| |
| 11:02 | <ogra> warren, there was a script added to upstream CK that they call from startx now
| |
| 11:02 | <vagrantc> gvy: it is possible, yes.
| |
| 11:02 | <warren> mccann, any idea what ogra is talking about?
| |
| 11:03 | ogra, mccann is our CK maintainer.
| |
| 11:03 | <gvy> vagrantc, and i don't mind if there's proper discussion and these are dropped with reworking the code to clean it up along the agreed boolean standard :)
| |
| 11:03 | <ogra> i know :)
| |
| 11:03 | <vagrantc> gvy: well, until we have that discussion, i reject the changes.
| |
| 11:03 | <mccann> warren: yeah I mentioned it to you yesterday, ck-launch-session :)
| |
| 11:03 | <gvy> but frankly, i'm surprised to see such strong opposition to such simple thing
| |
| 11:03 | <vagrantc> i've got to go now.
| |
| 11:03 | <gvy> vagrantc, thanks anyways
| |
| 11:03 | <warren> mccann, is that part backportable to F-8 or requires too much?
| |
| 11:04 | <ogra> warren, mccann right, thats the one
| |
| 11:04 | i got a session though with our sshd, even without the script
| |
| 11:04 | <warren> mccann, if an ssh session uses ck-launch-session, is it easy to automatically unregister when the ssh connection dies/
| |
| 11:04 | <ogra> i just have 6the prob that remote sessions are not allowed anything
| |
| 11:05 | <mccann> warren: http://gitweb.freedesktop.org/?p=ConsoleKit.git;a=blob;hb=HEAD;f=tools/ck-launch-session.c
| |
| 11:05 | <warren> mccann, oh, you would do like ck-launch-session /usr/bin/gnome-session?
| |
| 11:05 | <mccann> yeah
| |
| 11:06 | <warren> so when /usr/bin/gnome-session dies, ck-launch-session can unregister from CK?
| |
| 11:06 | <mccann> that's the idea anyway
| |
| 11:06 | you should test it tho
| |
| 11:06 | <warren> mccann, will this ck-launch-session work on F-8?
| |
| 11:07 | <mccann> and test that it sets up the session correctly
| |
| 11:07 | * gvy writes an async message to ML so that boolean question can be aired offline | |
| 11:07 | <mccann> warren: it should. There isn't much to it. But you should test it
| |
| 11:07 | <warren> mccann, will do. thanks.
| |
| 11:07 | * warren gets some quick food | |
| 11:10 | Pascal_1 has quit IRC | |
| 11:12 | <gvy> 500 Internal error: http://bazaar.launchpad.net/~shigorin/ltsp/ltsp-altlinux/annotate/ledest%40gmail.com-20080327153943-sk2gdajsrinxfd7p?file_id=ltspcommonfunctions-20080312182155-3niy8jrp3s3218mg-1
| |
| 11:12 | should it be reported somewhere?
| |
| 11:13 | <ogra> gvy, ask #launchpad
| |
| 11:13 | * gvy prefers term "quick silent food" | |
| 11:13 | <gvy> ogra, tnx
| |
| 11:15 | Pascal_1 has joined #ltsp | |
| 11:18 | <warren> yesterday they said they're working on the 500 errors
| |
| 11:19 | spectra has joined #ltsp | |
| 11:22 | joebaker has joined #ltsp | |
| 11:22 | <gvy> warren, ah, hopefully it's the remnant, not the result :)
| |
| 11:23 | * gvy has had different results during different works | |
| 11:24 | Pascal_1 has quit IRC | |
| 11:27 | elisboa has quit IRC | |
| 11:51 | joebaker has quit IRC | |
| 11:52 | mhterres has quit IRC | |
| 11:53 | mhterres has joined #ltsp | |
| 11:53 | staffencasa has joined #ltsp | |
| 11:53 | <warren> mccann, #fedora-meeting in 8 minutes.
| |
| 11:55 | ogra, vagrantc: if you're interested I'm going to be talking about ldm first in #fedora-meeting
| |
| 11:55 | <ogra> i'll lurk a bit between my other work
| |
| 11:56 | <warren> k
| |
| 11:57 | <vagrantc> warren: i'll drop in, but i'm mostly preparing to leave town
| |
| 11:58 | <warren> vagrantc, k
| |
| 11:58 | vagrantc, where are you btw?
| |
| 11:58 | <vagrantc> warren: buffalo, ny ... going to nyc tomorrow ... then to madrid
| |
| 12:06 | Pascal_1 has joined #ltsp | |
| 12:08 | <gvy> vagrantc, wish you easy landings!
| |
| 12:10 | joebaker has joined #ltsp | |
| 12:10 | <gvy> vagrantc, I've finished the message to the list -- with snippets and examples, not only some frustration on tastes :) re-reading before hitting send...
| |
| 12:11 | <vagrantc> gvy: cool.
| |
| 12:11 | gvy: and of course, i'm only mentioning the things i have issue with ... though there's a few plugins or ideas i'll probably grab for debian :)
| |
| 12:12 | <gvy> vagrantc, yeah, i do understand you :)
| |
| 12:12 | but want to resolve issues, even if in constructive way *grin* :)
| |
| 12:13 | <vagrantc> even if? :)
| |
| 12:14 | gonzaloaf has quit IRC | |
| 12:14 | <gvy> vagrantc, yup 8)
| |
| 12:15 | (well jokes aside, it's not that odd if :)
| |
| 12:17 | <enter>
| |
| 12:18 | vagrantc, i'll post re BOOTPROMPT_OPTS as well, was asking questions myself too
| |
| 12:18 | still that one seems way more trivial (and the post should be tiny)
| |
| 12:19 | <vagrantc> gvy: BOOTPROMPT_OPTS is more an issue of changing default behavior, rather than adding functionality ... i like the idea of it being more flexible.
| |
| 12:19 | <gvy> vagrantc, exactly
| |
| 12:19 | <ogra> mccann, initially ldm was only a fullscreen gtk wion with 2 input fields that killed itself after input and ran ssh -X user@server /etc/X11/Xsession
| |
| 12:19 | mccann, with the c rewrite the setup got a lot more complex
| |
| 12:19 | <vagrantc> gvy: overall, i think all the update-kernels code needs a good overhaul.
| |
| 12:20 | <ogra> we use 2 ssh clients now and a comunication API between front and backend
| |
| 12:21 | <gvy> vagrantc, well, we're at r706 now (not pushed yet, more intrusive)
| |
| 12:21 | <vagrantc> gvy: i'd really like to see ALTLinux-specific changes in a separate branch from proposed general changes
| |
| 12:22 | <gvy> vagrantc, umm... me too, was picking up bzr's concept of branches with some pain (using git's all-in-one-dir)
| |
| 12:22 | <vagrantc> gvy: yes, it's totally different.
| |
| 12:23 | <gvy> and at the same time explaining all of that to led, with mini-tutorial, man bzr and git experience flying around :)
| |
| 12:23 | <ogra> mccann, the first ssh connection does the password handling and sets up the actual tunnel and a socket ... that way you can interact with sshd in case of a password expiry etc ... once the first connection is established, the greeter goes away and attaches a second client to the socket running the actual Xsession
| |
| 12:24 | <gvy> vagrantc, part of decision to do things up to the point currently published was getting feedback and not wasting more (ours and yours) time on re-shuffling branches
| |
| 12:24 | rasmuson_ has joined #ltsp | |
| 12:25 | <mccann> ogra: ok thanks
| |
| 12:25 | <gvy> i've initially thought that branches are git-cheap in bzr, but was wrong... so we decided to stack the reasonable amount of changes which should be either trivially merged (non-conflicting by design) or an easy review
| |
| 12:25 | a few hours ago i've drawn the scheme with "merge" branch and "devel" one :)
| |
| 12:25 | <ogra> mccann, if gdm can handle such a setup (and some extras like rc.d scripts before and after the second call) there would be no need for ldm
| |
| 12:26 | <vagrantc> gvy: the issues arise where we disagree on easy review :)
| |
| 12:26 | <gvy> vagrantc, so we're learning how things are done here, too, but of course not in a moment
| |
| 12:26 | vagrantc, yup :) and as you agreed to try and review, it's not my call to coerce you into anything :)
| |
| 12:26 | vagrantc, let's postpone it until you have time? or to the ML
| |
| 12:27 | missing transport doesn't really help development
| |
| 12:29 | <vagrantc> gvy: well, i'm catching a train tomorrow morning... and my todo list isn't as full as it looked initially
| |
| 12:29 | <gvy> vagrantc, ah, thought "rather immediately" as it happens
| |
| 12:30 | <vagrantc> gvy: y'all are the first to use > 095 ...
| |
| 12:30 | * vagrantc thought from day 1 it should just be 00-99 | |
| 12:32 | <vagrantc> gvy: so, after reviewing the whole diff ... those are pretty much my only issues.
| |
| 12:33 | <gvy> vagrantc, not that bad to me :)
| |
| 12:34 | <vagrantc> gvy: it's mostly plugins to ALTLinux, so you can do things however crazy you want. :)
| |
| 12:35 | <gvy> vagrantc, well, people could look inside and catch that craziness, would be bad ;)
| |
| 12:35 | <vagrantc> gvy: just the boolean changes and BOOTPROMPT_OPTS, both of which i'm sure we can work out.
| |
| 12:35 | <gvy> yup
| |
| 12:35 | <vagrantc> gvy: true, true.
| |
| 12:35 | setting a bad/good example...
| |
| 12:36 | <gvy> vagrantc, could you read the second half of the list message? there are links to good examples (by my measure) there as well
| |
| 12:36 | btw i considered proposing libshell some time down the road anyways. :)
| |
| 12:36 | <vagrantc> gvy: a fair number of the ALTLinux plugins look like they would also work for Debian with a fre changes.
| |
| 12:36 | s,fre,few,
| |
| 12:37 | gvy: i'd like to try and keep outside dependencies to a minimum... but if it's widely accepted, by all means :)
| |
| 12:41 | widely accepted and makes our code better
| |
| 12:41 | that is
| |
| 12:41 | and is stable and not changing all the time, like bzr...
| |
| 12:43 | <gvy> vagrantc, well, that might involve lugging a copy around -- don't think it's widely used atm
| |
| 12:43 | quite stable though
| |
| 12:45 | ledest has joined #ltsp | |
| 12:45 | <gvy> finally
| |
| 12:45 | vagrantc, ledest; ledest, vagrantc :)
| |
| 12:45 | * vagrantc waves to ledest | |
| 12:46 | praveer_fedora is now known as praveer_cool | |
| 12:47 | <gvy> vagrantc, so regarding bools: is "the third [empty] state" argument reasonable to you?
| |
| 12:48 | as i've posted, see no problem with is_yes()/is_no(), if these hurt then sed -ir heals :)
| |
| 12:48 | <vagrantc> gvy, ledest: when making commits that are distro-specific, we've been using "$DIST: $comments" for the commit message
| |
| 12:49 | gvy: i think is_yes, is_no is just an alias for the other names ...
| |
| 12:49 | gvy: and i haven't read your email yet, by the way ...
| |
| 12:49 | <gvy> vagrantc, exactly
| |
| 12:49 | <vagrantc> gvy: so, i don't see much reason to just define an alias ...
| |
| 12:49 | <gvy> well, a one-line form is "np" :)
| |
| 12:50 | vagrantc, btw ledest put "ALTLinux: ..." in r703 not even knowing of that rule :)
| |
| 12:51 | <vagrantc> gvy: yeah... the earlier commits are more like "Added ALTLinux-specific ..." which seems verbose. again, not a huge deal. :)
| |
| 12:51 | <gvy> vagrantc, hope we'll get used to how it uses to be here. :)
| |
| 12:54 | btw is there a "developer's corner" on the wiki? i've looked more closely and cannot see any
| |
| 12:54 | i'd drop a "$DIST: $comments" there too
| |
| 12:54 | * vagrantc finds wiki.ltsp.org difficult to use | |
| 12:57 | steph_ has left #ltsp | |
| 12:58 | Pascal_1 has quit IRC | |
| 13:00 | <vagrantc> gvy: merged ..690
| |
| 13:00 | gvy: i.e. the easy stuff :)
| |
| 13:01 | <gvy> vagrantc, http://wiki.ltsp.org/twiki/bin/view/Ltsp/DevelopersCorner :)
| |
| 13:02 | vagrantc, good :) what's hurting next?
| |
| 13:05 | elisboa has joined #ltsp | |
| 13:05 | <gvy> e.g. in r691 (expand server/ltsp-build-client, server/ltsp-update-kernels, expand server/ltsp-update-sshkeys)
| |
| 13:07 | Pascal_1 has joined #ltsp | |
| 13:12 | Q-FUNK has quit IRC | |
| 13:12 | Pascal_1 has quit IRC | |
| 13:15 | <gvy> vagrantc, yay, pushed upstream r683 :)
| |
| 13:15 | * gvy offers a hand | |
| 13:17 | indradg has joined #ltsp | |
| 13:17 | hansatuofd has quit IRC | |
| 13:20 | nicoAMG has joined #ltsp | |
| 13:20 | <rjune_> !seen gadi
| |
| 13:20 | <ltspbot> rjune_: gadi was last seen in #ltsp 2 days, 1 hour, 24 minutes, and 46 seconds ago: <Gadi> agreed
| |
| 13:20 | gonzaloaf has joined #ltsp | |
| 13:23 | johnny has joined #ltsp | |
| 13:25 | Sockra_Tease has joined #ltsp | |
| 13:26 | Sockra_Tease is now known as cpunches | |
| 13:35 | dan_young has joined #ltsp | |
| 13:46 | <johnny> uggh..launchpad codebrowsing is broke again :(
| |
| 13:46 | stupid proxy error
| |
| 13:49 | vagrantc has quit IRC | |
| 13:53 | Pascal_1 has joined #ltsp | |
| 14:00 | <warren> https://bugzilla.redhat.com/show_bug.cgi?id=436906 default pulseaudio daemon launching options
| |
| 14:01 | Isn't this another case similar to X -ac?
| |
| 14:01 | Anybody on the network can direct whatever noise they want at somebody's client?
| |
| 14:01 | <klausade> Pascal_1: you might have gotten mail.
| |
| 14:02 | <Pascal_1> yes i didnt make the link between you and the name on the mail !!
| |
| 14:03 | but the problem is about pal / ldm then i think it didnt works
| |
| 14:03 | slidesinger has quit IRC | |
| 14:03 | mccann has quit IRC | |
| 14:03 | <klausade> Pascal_1: my initial tests shows that it works here with your chroot and kdm.
| |
| 14:04 | <Pascal_1> by default it use ldm/gdm no ?
| |
| 14:04 | i've got to install kde on my debian ltsp server ?
| |
| 14:05 | <ogra> warren, the thing its that its no regression to any former ltsp versions .... surely we can impriove it :)
| |
| 14:06 | <klausade> Pascal_1: you use gdm+gnome? I can test that here later.
| |
| 14:06 | <warren> ogra, it seems extremely cavalier to ship with a gaping security hole.
| |
| 14:06 | <Pascal_1> debian is gnome default i think
| |
| 14:06 | <ogra> warren, sound was always wide open ... and i dont really se a security issue here ... probably a fun one though if students can play sound on others machines :)
| |
| 14:06 | warren, feel free to fix it :)
| |
| 14:07 | <warren> ogra, if students have headsets on, other people can really scare the heck out of them.
| |
| 14:07 | ogra, did you push fixes to xauth for your existing and new LTSP versions in Ubuntu?
| |
| 14:08 | You seriously need to treat this more seriously.
| |
| 14:08 | <ogra> yeah, we used to do that at school on the terminals ... you could really scare people while playing a nethack session with wei5rd noises, that wa a lot of fun :)
| |
| 14:08 | <warren> The xauth issue alone deserves a CVE report
| |
| 14:08 | it is very serious that anybody on the network can spy on or hijack your session
| |
| 14:08 | <ogra> warren, well, if it would be seriously something that enables you to take over the machine i would be worried
| |
| 14:09 | <warren> yes it is
| |
| 14:09 | <ogra> but thats simply not the case
| |
| 14:09 | <johnny> does it affect ltsp4 too?
| |
| 14:09 | <warren> oh really?
| |
| 14:09 | johnny, no
| |
| 14:09 | <ogra> johnny, it affects all networked sound systems
| |
| 14:09 | <warren> ogra, ldm with LDM_DIRECTX?
| |
| 14:09 | <johnny> i think he's talking about the -ac flag, not sound
| |
| 14:09 | <ogra> pulse is the first one adding a rudimentary security to that
| |
| 14:09 | <warren> ogra, in fact your ldm uses X -ac makes it insecure even with ssh
| |
| 14:10 | <johnny> sound is seperate issue
| |
| 14:10 | right?
| |
| 14:10 | <ogra> warren, are we talking about sound now or about fixed X issues ?
| |
| 14:10 | <warren> ogra, did you already push fixes to all your distros for the X -ac issue?
| |
| 14:10 | <ogra> -ac is fixed
| |
| 14:10 | and will be in the ubuntu release as wello
| |
| 14:10 | <warren> that one really deserved a CVE
| |
| 14:10 | <ogra> feel free
| |
| 14:11 | <johnny> i'll bbl folks.. gotta talk to credit card processing people
| |
| 14:12 | <ogra> id really appreciate a proper security fix for pulse ... but thats surely moretricky than you think to lock it per session ...
| |
| 14:12 | the plugins only allow filtering on IP base...
| |
| 14:12 | which doeswnt gain you anything
| |
| 14:13 | deavid has joined #ltsp | |
| 14:16 | tux_440volt is now known as Subhodip | |
| 14:21 | joebaker has quit IRC | |
| 14:21 | <gvy> ogra, just in case: a top security expert, Nakamura-san iirc, got his Sun workstation hacked some ten years ago or so
| |
| 14:21 | by a kid called Kevin Mitcnick
| |
| 14:22 | just because his X was listening to the wide open network
| |
| 14:22 | iirc an input window was opened
| |
| 14:22 | <ogra> i didnt object that the X bug is serious
| |
| 14:22 | <gvy> on the whole screen
| |
| 14:22 | capturing all the input... :)
| |
| 14:22 | ogra, well, it's rather an amusing story, not a cve comment :)
| |
| 14:23 | <ogra> but i do for sound ... we dont support capturing there, all a hacker could do would be to play a sound through your speakers
| |
| 14:23 | <laga> heh
| |
| 14:23 | dan_young has quit IRC | |
| 14:23 | <laga> i remember the security issue in windows vista.. where an attacker could play back voice commands ;)
| |
| 14:23 | <ogra> and given the lockdown options we have (that could lock out IPs but thats about it) i dont see a way to easily make it secure
| |
| 14:23 | <laga> not sure if that was ever exploited
| |
| 14:24 | opapo has joined #ltsp | |
| 14:24 | <ogra> you dont gain anything by restricting to the server IP for example
| |
| 14:24 | since all ltsp users can still play sounds through your speakers
| |
| 14:24 | <opapo> When I restart my client the server gets restarted
| |
| 14:25 | <laga> that's what you get for IRCing as root ;)
| |
| 14:25 | <gvy> warren, i hope that some day we'll just do an audit of the branch, and will have backing by Dmitry Levin for that either
| |
| 14:25 | <opapo> I installed Edubuntu server on a VM
| |
| 14:25 | <gvy> warren, since it's not exactly decent to understand that the decision to use good ol
| |
| 14:25 | <opapo> Then I had another vm that was the client
| |
| 14:25 | <warren> I just talked with Lennart (pulseaudio lead developer), he has a secure solution
| |
| 14:25 | <gvy> ol' xdmcp seems to be actually more secure as well than using ssh, ldm and finally -ac
| |
| 14:26 | <ogra> warren, one thats implementable ?
| |
| 14:26 | <warren> ogra, I'm studying it
| |
| 14:26 | <ogra> and works attached to the ssh tunnel ?
| |
| 14:26 | thats our only way to attach securely to the session
| |
| 14:26 | <gvy> well, cookies might be the way
| |
| 14:27 | <ogra> we have a secure connection
| |
| 14:27 | why should we use cookies
| |
| 14:27 | <gvy> ogra, and restricting by server ip *does* gain the inability of the rest of lan (if it's not a terminal network alone) to make fun :)
| |
| 14:27 | <ogra> all we need to manage is to tie the sound to it
| |
| 14:27 | <gvy> ogra, no -- cookies over secure connections, then use within pulse to allow/deny
| |
| 14:28 | <ogra> well, we have and use an xatom cookie already
| |
| 14:28 | <gvy> sound over ssh was a bad idea to me a few years ago... delays fly high
| |
| 14:28 | <ogra> we could just use that
| |
| 14:28 | <gvy> yup
| |
| 14:28 | <ogra> i dodnt say to puch the payload though the tunnel :)
| |
| 14:28 | <gvy> ah :)
| |
| 14:28 | <ogra> *push
| |
| 14:28 | <gvy> *punch
| |
| 14:28 | <ogra> ltspfs doesnt do that either
| |
| 14:29 | but it uses the ssh security for validation
| |
| 14:29 | <warren> xatom cookies are for ltspfs?
| |
| 14:29 | <ogra> pulse should do the same
| |
| 14:29 | <warren> not pulse right?
| |
| 14:29 | <ogra> right
| |
| 14:29 | but nothing forbids us to use them generally
| |
| 14:30 | opapo has left #ltsp | |
| 14:30 | kyron has quit IRC | |
| 14:30 | <ogra> inn any case i'm not really concerned about the issue .. if you find a fix, great if you dont, it will go on to work as it always did
| |
| 14:31 | -ac was actually a regression ... beyond being a huge security hole
| |
| 14:33 | elisboa has quit IRC | |
| 14:38 | mhterres has quit IRC | |
| 14:38 | mhterres has joined #ltsp | |
| 14:41 | praveer_cool has quit IRC | |
| 14:41 | praveer_cool has joined #ltsp | |
| 14:54 | johnny_ has joined #ltsp | |
| 15:02 | gonzaloaf has quit IRC | |
| 15:02 | * gvy is away: ~ | |
| 15:02 | <gvy> bb, going home for weekend
| |
| 15:02 | thanks & luck
| |
| 15:05 | ledest has left #ltsp | |
| 15:07 | DonSilver has joined #ltsp | |
| 15:10 | J45p3r has joined #ltsp | |
| 15:13 | Pascal_Debian has joined #ltsp | |
| 15:14 | Pascal_Debian has quit IRC | |
| 15:14 | Pascal_Debian has joined #ltsp | |
| 15:14 | Subhodip has quit IRC | |
| 15:15 | Q-FUNK has joined #ltsp | |
| 15:15 | TelnetManta has quit IRC | |
| 15:17 | Pascal_Debian has joined #ltsp | |
| 15:18 | cliebow has quit IRC | |
| 15:22 | johnny_ has quit IRC | |
| 15:24 | K_O-Gnom has quit IRC | |
| 15:25 | johnny_ has joined #ltsp | |
| 15:26 | gonzaloaf has joined #ltsp | |
| 15:27 | Pascal_1 has quit IRC | |
| 15:29 | Q-FUNK has quit IRC | |
| 15:29 | Q-FUN1 has joined #ltsp | |
| 15:29 | Q-FUN1 is now known as Q-FUNK | |
| 15:32 | DonSilver is now known as K_O-Gnom | |
| 15:33 | cpunches has quit IRC | |
| 15:34 | gonzaloaf has quit IRC | |
| 15:40 | K_O-Gnom has quit IRC | |
| 15:47 | nicoAMG has quit IRC | |
| 16:03 | Q-FUNK has quit IRC | |
| 16:03 | Q-FUNK has joined #ltsp | |
| 16:03 | tiagovaz_ has joined #ltsp | |
| 16:04 | Pascal_Debian has quit IRC | |
| 16:04 | tiagovaz_ has quit IRC | |
| 16:05 | tiagovaz has quit IRC | |
| 16:06 | Guaraldo has left #ltsp | |
| 16:09 | johnny_ has quit IRC | |
| 16:21 | gonzaloaf has joined #ltsp | |
| 16:29 | Q-FUNK has quit IRC | |
| 16:32 | K_O-Gnom has joined #ltsp | |
| 16:35 | Pascal_Debian has joined #ltsp | |
| 16:48 | indradg has quit IRC | |
| 16:59 | rasmuson_ has quit IRC | |
| 16:59 | Pascal_Debian has quit IRC | |
| 17:01 | mhterres has left #ltsp | |
| 17:01 | Pascal_Debian has joined #ltsp | |
| 17:04 | milesd has joined #ltsp | |
| 17:08 | Pascal_Debian has quit IRC | |
| 17:08 | Egyptian[Home1 has joined #ltsp | |
| 17:09 | Egyptian[Home] has quit IRC | |
| 17:09 | rjune has joined #ltsp | |
| 17:17 | slipttees has joined #ltsp | |
| 17:27 | gonzaloaf has quit IRC | |
| 17:29 | rjune has quit IRC | |
| 17:30 | rjune has joined #ltsp | |
| 17:35 | gonzaloaf has joined #ltsp | |
| 18:07 | deavid has quit IRC | |
| 18:08 | deavid has joined #ltsp | |
| 18:08 | deavid has quit IRC | |
| 18:09 | deavid has joined #ltsp | |
| 18:10 | barbosa has joined #ltsp | |
| 18:13 | mikkel has quit IRC | |
| 18:28 | milesd has quit IRC | |
| 18:38 | cliebow has joined #ltsp | |
| 19:02 | twinprism has quit IRC | |
| 19:03 | slipttees has left #ltsp | |
| 19:07 | robbie has joined #ltsp | |
| 19:08 | K_O-Gnom has quit IRC | |
| 19:11 | staffencasa has quit IRC | |
| 19:28 | Egyptian[Home1 has quit IRC | |
| 19:30 | robbie has quit IRC | |
| 19:30 | TelnetManta has joined #ltsp | |
| 19:30 | score has quit IRC | |
| 19:37 | Egyptian[Home] has joined #ltsp | |
| 19:52 | rjune has joined #ltsp | |
| 20:09 | Egyptian[Home] has quit IRC | |
| 20:21 | Egyptian[Home] has joined #ltsp | |
| 21:04 | MacIver has quit IRC | |
| 21:11 | score has joined #ltsp | |
| 21:12 | cliebow has quit IRC | |
| 21:24 | twinprism has joined #ltsp | |
| 21:27 | MacIver has joined #ltsp | |
| 21:53 | J45p3r has left #ltsp | |
| 22:04 | MacIver has quit IRC | |
| 22:10 | MacIver has joined #ltsp | |
| 23:08 | spectra has quit IRC | |
| 23:34 | chup has joined #ltsp | |
| 23:34 | chup is now known as chupacabra | |
| 23:59 | Faithful has quit IRC | |