IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 28 December 2012 (all times are UTC)
| 00:13 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 00:43 | alexqwesa__ has left IRC (alexqwesa__!~alex@109.172.12.47, Quit: Хана X'ам !!!) | |
| 00:43 | alex__ has joined IRC (alex__!~alex@109.172.12.47) | |
| 00:47 | Parker955_Away is now known as Parker955 | |
| 00:59 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 260 seconds) | |
| 01:14 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 01:38 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 02:21 | Parker955 is now known as Parker955_Away | |
| 02:23 | Parker955_Away is now known as Parker955 | |
| 02:29 | <highvoltage> 4
| |
| 02:36 | <Phantomas> 3
| |
| 02:47 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 248 seconds) | |
| 03:03 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 03:55 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 252 seconds) | |
| 04:50 | ltspuser_20 has joined IRC (ltspuser_20!636b5b5d@gateway/web/freenode/ip.99.107.91.93) | |
| 04:51 | <ltspuser_20> Don't see icons or programs on my thin client after login
| |
| 05:05 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 05:12 | ltspuser_20 has left IRC (ltspuser_20!636b5b5d@gateway/web/freenode/ip.99.107.91.93, Ping timeout: 245 seconds) | |
| 05:12 | telex has left IRC (telex!~telex@freeshell.de, Ping timeout: 265 seconds) | |
| 05:20 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 05:33 | telex has left IRC (telex!~telex@freeshell.de, Ping timeout: 251 seconds) | |
| 05:35 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 06:12 | MonkWitDaFunk has joined IRC (MonkWitDaFunk!~yaaic@199.119.234.230) | |
| 06:25 | bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl) | |
| 07:31 | FuriousGeorge has joined IRC (FuriousGeorge!~chatzilla@ool-182c2966.dyn.optonline.net) | |
| 07:31 | <FuriousGeorge> hey all
| |
| 07:32 | im missing an option on my dhcp server, but I can't remember what I need. I have filename set and next-server set
| |
| 07:32 | note that the DHCP server is not the same as the LTSP server
| |
| 07:40 | im getting "connection timed out"
| |
| 07:44 | Parker955 is now known as Parker955_Away | |
| 07:51 | <alkisg> Those should suffice, try booting with ipxe, it has more verbose error messages etc
| |
| 07:52 | !ipxe
| |
| 07:52 | <ltsp> ipxe: iPXE is the successor to the etherboot/gPXE project, and can be used to netboot clients that don't have a NIC ROM with a PXE stack. To add it to grub, see !grub-ipxe. To add it to the Windows boot loader, see !win32-loader. To download floppy, CD or USB images, visit http://ipxe.org or install the ipxe package.
| |
| 07:56 | <FuriousGeorge> alkisg: ipxe is what im using here. thats what this client supports
| |
| 07:57 | <alkisg> FuriousGeorge: press ctrl+b, then type `config` and check that you receive the correct settings
| |
| 07:57 | <FuriousGeorge> alkisg: dont i have to specify the path
| |
| 07:57 | <alkisg> No
| |
| 07:57 | Unless you're using an NFS based distro... which one are you using?
| |
| 07:58 | <FuriousGeorge> im using tftp
| |
| 07:59 | i set this up 5 years ago, so im a little hazy on the details
| |
| 07:59 | <alkisg> Which linux distribution are you using for the server?
| |
| 07:59 | <FuriousGeorge> ubuntu
| |
| 07:59 | <alkisg> OK, so you don't need the nfs root there
| |
| 07:59 | Just next-server and boot filename
| |
| 08:00 | <FuriousGeorge> next-server is obviously the IP of my LTSP server (not the router), and filename is pxelinux.0, right?
| |
| 08:01 | alkisg: also, ctrl+b is not an option as im remotely connected though a java console that can't send that key combo
| |
| 08:02 | <alkisg> filename should be /ltsp/i386/pxelinux.0
| |
| 08:05 | <FuriousGeorge> alkisg: hmmm... for some reason, as soon as I add the "filename" option I go from getting "no such file
| |
| 08:05 | " to "connection timed out"
| |
| 08:05 | from what I read I should only get the latter if it's timing out getting the DHCP address
| |
| 08:06 | <alkisg> !tftp
| |
| 08:06 | <ltsp> tftp: Here's a page to help you troubleshoot TFTP problems in Ubuntu: https://help.ubuntu.com/community/UbuntuLTSP/Troubleshooting/TFTP
| |
| 08:06 | <alkisg> Verify that you can download pxelinux.0 from an external tftp client
| |
| 08:06 | Paragraph "Use an external tftp client"
| |
| 08:17 | <FuriousGeorge> trying that
| |
| 08:24 | alkisg: hmmm
| |
| 08:24 | Error occurred during the file transfer (Error code = 8):
| |
| 08:24 | Unsupported option(s) requested
| |
| 08:26 | i may just be using the client wrong... i did tftp.exe 10.0.0.20 GET /ltsp/i386/pxelinux.0
| |
| 08:26 | asmok has joined IRC (asmok!~ltsp@cs78145191.pp.htv.fi) | |
| 08:28 | <asmok> alkisg: are you using Puppet for your 300 schools? What kind of managing system? Just howtos for teachers/managers?
| |
| 08:29 | <alkisg> FuriousGeorge: try the linux client from the ubuntu server itself
| |
| 08:30 | asmok: nope, we're only providing 4 custom CDs/DVDs that have gnome-fallback as the default session and a few other settings/apps
| |
| 08:30 | And we have PPAs to send package updates to all schools
| |
| 08:30 | <asmok> ok
| |
| 08:32 | have you any alert/ticket systems for teachers/schools?
| |
| 08:34 | <alkisg> The central ticketing system for anything related to the school network, is helpdesksch.gr
| |
| 08:34 | helpdesk.sch.gr
| |
| 08:34 | We've put a category about linux/ltsp there
| |
| 08:35 | <asmok> ok
| |
| 08:35 | <alkisg> And of course we have bug trackers for epoptes, sch-scripts etc... e.g. bugs.launchpad.net/sch-scripts
| |
| 08:35 | <asmok> ok
| |
| 08:37 | thank you, i use google translator for your helpdesk, and happy new and better year for you Alkis
| |
| 08:38 | bb
| |
| 08:38 | asmok has left IRC (asmok!~ltsp@cs78145191.pp.htv.fi, Quit: Lähdössä) | |
| 08:49 | dobber has joined IRC (dobber!~dobber@213.169.45.222) | |
| 08:54 | <alkisg> asmok, in case you see the irclogs, the helpdesk requires login to see the tickets, so you won't be able to see anything at all...
| |
| 08:54 | bobby_C has joined IRC (bobby_C!~bobby@188.20.161.210) | |
| 09:05 | FuriousGeorge_ has joined IRC (FuriousGeorge_!~chatzilla@ool-182c2966.dyn.optonline.net) | |
| 09:06 | FuriousGeorge has left IRC (FuriousGeorge!~chatzilla@ool-182c2966.dyn.optonline.net, Ping timeout: 248 seconds) | |
| 09:06 | FuriousGeorge_ is now known as FuriousGeorge | |
| 09:23 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 09:25 | MonkWitDaFunk has left IRC (MonkWitDaFunk!~yaaic@199.119.234.230, Ping timeout: 252 seconds) | |
| 09:54 | gvy has joined IRC (gvy!~mike@altlinux/developer/mike) | |
| 11:10 | MonkWitDaFunk has joined IRC (MonkWitDaFunk!~yaaic@199.119.234.230) | |
| 11:12 | MonkWitDaFunk has left IRC (MonkWitDaFunk!~yaaic@199.119.234.230) | |
| 11:44 | markit has joined IRC (markit!~marco@88-149-177-66.v4.ngi.it) | |
| 11:52 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 11:54 | gvy has left IRC (gvy!~mike@altlinux/developer/mike, Quit: Leaving) | |
| 11:57 | PhoenixSTF has joined IRC (PhoenixSTF!~rudi@193.136.242.251) | |
| 12:09 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 260 seconds) | |
| 12:24 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 13:14 | [GuS] has joined IRC ([GuS]!~MysT@213-117-16-190.fibertel.com.ar) | |
| 13:14 | [GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402) | |
| 13:25 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 244 seconds) | |
| 13:31 | bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.) | |
| 13:40 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 13:52 | gvy has joined IRC (gvy!~mike@altlinux/developer/mike) | |
| 13:56 | FuriousGeorge has left IRC (FuriousGeorge!~chatzilla@ool-182c2966.dyn.optonline.net, Ping timeout: 248 seconds) | |
| 14:10 | |GuS| has joined IRC (|GuS|!~MysT@213-117-16-190.fibertel.com.ar) | |
| 14:10 | |GuS| has joined IRC (|GuS|!~MysT@unaffiliated/gus/x-663402) | |
| 14:11 | [GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Ping timeout: 260 seconds) | |
| 14:30 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 14:32 | |GuS| is now known as [GuS] | |
| 14:56 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 14:59 | Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com) | |
| 15:10 | staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 260 seconds) | |
| 15:23 | Parker955_Away is now known as Parker955 | |
| 15:26 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 15:39 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 265 seconds) | |
| 16:33 | dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection) | |
| 16:38 | vagrantc has joined IRC (vagrantc!~vagrant@c-98-232-129-196.hsd1.or.comcast.net) | |
| 16:38 | vagrantc has joined IRC (vagrantc!~vagrant@c-98-232-129-196.hsd1.or.comcast.net) | |
| 16:38 | vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc) | |
| 17:06 | sbalneav has left IRC (sbalneav!~sbalneav@mail.legalaid.mb.ca, Ping timeout: 248 seconds) | |
| 17:07 | sbalneav has joined IRC (sbalneav!~sbalneav@mail.legalaid.mb.ca) | |
| 17:12 | bobby_C has left IRC (bobby_C!~bobby@188.20.161.210, Quit: Goin' down hard) | |
| 17:27 | Parker955 is now known as Parker955_Away | |
| 17:29 | Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 17:30 | Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com) | |
| 17:31 | lotharn has left IRC (lotharn!~nick@24.154.55.32, Quit: Leaving) | |
| 17:34 | lotharn has joined IRC (lotharn!~nick@24.154.55.32) | |
| 17:56 | Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 18:31 | alex__ has left IRC (alex__!~alex@109.172.12.47, Quit: Хана X'ам !!!) | |
| 18:31 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 18:35 | alexqwesa_ has joined IRC (alexqwesa_!~alex@109.172.12.47) | |
| 18:35 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Client Quit) | |
| 18:41 | alexqwesa_ has left IRC (alexqwesa_!~alex@109.172.12.47, Quit: Хана X'ам !!!) | |
| 18:41 | alexqwesa_ has joined IRC (alexqwesa_!~alex@109.172.12.47) | |
| 18:44 | alexqwesa_ has left IRC (alexqwesa_!~alex@109.172.12.47, Client Quit) | |
| 18:44 | alexqwesa_ has joined IRC (alexqwesa_!~alex@109.172.12.47) | |
| 19:10 | bobby_C has joined IRC (bobby_C!~bobby@188.20.161.210) | |
| 19:30 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 19:33 | <alkisg> sbalneav, stgraber: we're having some LDM cleanup problems but it appears that no LDM hackers are around, could either of you comment on https://bugs.launchpad.net/ltsp/+bug/1093144 ?
| |
| 19:34 | Also... I was thinking, if LDM was rewritten in shell and the greeter maybe in python, then more people could maintain it... thoughts on that?
| |
| 19:36 | PhoenixSTF has left IRC (PhoenixSTF!~rudi@193.136.242.251, Remote host closed the connection) | |
| 19:37 | <stgraber> alkisg: very bad idea ;) that's what ldm used to be (python + shell) and it was so slow and memory hungry that we had to rewrite it in C
| |
| 19:37 | alkisg: the current plan is to drop ldm completely
| |
| 19:38 | <vagrantc> how's that plan progressing?
| |
| 19:38 | it seems a couple days a year might not be enough
| |
| 19:40 | <stgraber> well, I had it working at BTS but we need someone to test the improvements Scott did after BTS and make sure all of it is packaged, ... then start integrating
| |
| 19:41 | personally I'm busy with so many other projects at the time and not actually using LTSP myself that I won't commit to much more than updating ltsp and ldm in Ubuntu for the next 4 months or so
| |
| 19:41 | hopefully I'll get some period of time where I'm less busy then and where I can cleanup that stuff a bit
| |
| 19:43 | anyway, I expect we'll be doing maintenance on ldm for quite a while still and there's no real way around that. Rewriting it would just make things even worse by having yet-another-display-manager
| |
| 19:44 | alkisg: for that specific bug, my guess is that it's not fixable with the way ldm currently works. Although the greeter and the backend are separate processes, then backend will die when X does, so there's currently no way to run things after X is dead, at least not using ldm scripts.
| |
| 19:47 | <alkisg> slow + memory hungry => could that depend on the implementation? E.g. now ldm stays around in the user session and needs some MB ram for that, with the shell we'd only need the ldm script to hang around
| |
| 19:47 | So it would need much LESS memory than what it needs now...
| |
| 19:49 | stgraber: and about the backends dying, that's why I think some bigger changes are needed there
| |
| 19:49 | Currently things break on second login, and the way around it is to use NFS
| |
| 19:49 | So the default sshfs setup isn't working
| |
| 19:50 | (the .gvfs can't be mounted the second time)
| |
| 19:50 | <stgraber> alkisg: why does LTSP even care about .gvfs? isn't gnome-session supposed to cleanup its own mess on logout
| |
| 19:51 | <alkisg> stgraber: some processes only die when X dies, dunno why
| |
| 19:51 | <stgraber> that's the bug that needs fixing
| |
| 19:51 | <alkisg> That might be 20+ bugs in several packages
| |
| 19:51 | I don't think we can do that
| |
| 19:51 | <stgraber> ldm closes the session when Xsession returns, by that time everything in the user session should be dead, if that's not the case, those need fixing
| |
| 19:52 | though that problem may soon get itself fixed with the work I'm doing at the moment (completely unrelated to LTSP)
| |
| 19:52 | <alkisg> I think that's the main reason we restart X
| |
| 19:52 | <stgraber> we're currently busy getting upstart to manage the user sessions in place of gnome-session, which enforces a proper process hierarchy and ensures everything is closed when the session exits
| |
| 19:53 | <alkisg> stgraber: there's also the point of "what if X crashes, shouldn't LTSP cleanup after itself?"
| |
| 19:53 | upstart won't solve LTSP problems for all distros though
| |
| 19:53 | <stgraber> nope, the main reason we restart X is to flush all the X properties and similar crap. Killing X to hope that all the user process die is just plain wrong (because there's no guarantee they'll actually die when X disappears)
| |
| 19:54 | <alkisg> What about the "X crashes" point?
| |
| 19:54 | If we don't restart X, the greeter crashes
| |
| 19:54 | (tested some months ago)
| |
| 19:55 | <stgraber> that sounds like a bug that might have been worth fixing if we weren't planning on replacing ldm with lightdm :)
| |
| 19:55 | <alkisg> But at what time frame? In 5 years?
| |
| 19:55 | <stgraber> for X crashing, we may be able to do some basic cleanup from the xinit scripts, but that's about as much as we can do
| |
| 19:55 | <alkisg> I think many many problems will need to be solved to migrate to lightdm
| |
| 19:56 | The session stuff, not the authentication part
| |
| 19:56 | OK let me rephrase...
| |
| 19:57 | If I can get ldm with shell + python, working fine with the current backend functionality but without any cluster bits because I don't really know what ltsp-cluster needs, lighter on ram etc, would anyone mind?
| |
| 19:57 | It would indeed introduce some regressions
| |
| 19:57 | But it would be maintainable by the people that do have time currently to work on ltsp...
| |
| 19:58 | <vagrantc> just call it lldm :)P
| |
| 19:58 | <alkisg> :)
| |
| 19:58 | <||cw> pyldm?
| |
| 19:58 | <vagrantc> shpyldm
| |
| 19:59 | <sbalneav> alkisg: OK, I'll have a look
| |
| 19:59 | <||cw> pyldm.sh!
| |
| 19:59 | <stgraber> personaly, I would mind and I wouldn't ship it in Ubuntu as I actually have extensive experience with the C version and can easily fix it for all Ubuntu versions, whereas having a shell+python implementation will mean different bugs and a completely different implementation, causing me to spend a lot of time to re-learn that stuff
| |
| 20:00 | <||cw> call it anchietae
| |
| 20:00 | <stgraber> hey sbalneav :)
| |
| 20:00 | <alkisg> OK thank you guys, sbalneav I'll be around for a few hours if you happen to need feedback on the bug report...
| |
| 20:00 | <sbalneav> Just as a note: I'm currently over in #mate, helping get the next version of mate out the door for March.
| |
| 20:00 | Once that's done, I'm going to continue on with the PAM stuff.
| |
| 20:01 | <alkisg> Ah, cool
| |
| 20:01 | lightdm will also solve some things related to gnome-panel dying, dbus services missing etc
| |
| 20:02 | <stgraber> I really fell that the right way forward is to spend time releasing and packaging the missing bits of sbalneav's nss/pam code, then switch to lightdm and never again have to care about the display manager and greeter. A stopgap solution will just make us have to maintain yet-another-thing for possibly years...
| |
| 20:02 | <alkisg> stgraber: there are many bits that I couldn't think of how to implement them with a "generic DM" like lightdm
| |
| 20:02 | Selecting different servers, backends...
| |
| 20:02 | <sbalneav> agreed. Could we maybe set aside a two week period in Jan or Feb when we could all do an intense "hack a thon" to get the pam stuff going?
| |
| 20:03 | * vagrantc cheers sbalneav on! | |
| 20:03 | <alkisg> OK, I think lightdm got some support for rdesktop, but I don't know how ltsp-cluster would fit there
| |
| 20:03 | <stgraber> alkisg: I really hope to have some time for LTSP stuff soon, but I'm currently investing all my limited spare time into Edubuntu Server. During my work time I've been doing a fair bit of DBUS stuff, so I'm pretty sure I can now easily bridge the client and server DBUS and solve pretty much all our bugs there, if only I had time :)
| |
| 20:04 | <alkisg> I'm not sure we do want to bridge the client and server dbus...
| |
| 20:04 | <sbalneav> vagrantc: alkisg: would both of you have some time?
| |
| 20:04 | <alkisg> Fat clients will be a more and more common setup as things move along in the 3d era...
| |
| 20:04 | <stgraber> alkisg: so lightdm already supports plugins that offer extra stuff in the greeter. That's what Ubuntu uses for the remote login function. The idea is to base our greeter on something like that, which will let us add any extra function we need to the greeter pretty easily.
| |
| 20:04 | <alkisg> sbalneav: I could make some time as this is too important to pass the opportunity :)
| |
| 20:05 | <vagrantc> sbalneav: i'll usually have some time, although i feel way behind in understanding the whole thing
| |
| 20:05 | <sbalneav> That's ok, we can all get up to speed :)
| |
| 20:05 | <stgraber> alkisg: for now the priority should be to get the authentication/nss side of things to work properly, then I can easily test with a minimal lightdm setup, then we can figure out what's missing and I can talk with the lightdm upstream who happens to be a colleague of mine
| |
| 20:05 | <alkisg> stgraber: the priority currently is to make 12.04 working - it's currently broken :-/
| |
| 20:05 | <vagrantc> might also be in montreal mid-may, if there are any folks who'd want to meet up in person for a hacking spell
| |
| 20:06 | <alkisg> We have too many critical bugs that make the current ubuntu/ltsp solution unusuable
| |
| 20:06 | Schools are uninstalling 12.04 and getting back to 10.04 to get stable results
| |
| 20:06 | <sbalneav> what's one of the killer bugs?
| |
| 20:06 | <vagrantc> wasn't the original idea not to be dependent on a single DM? i mean, lightdm is nice, but what if something else comes along soonish?
| |
| 20:07 | <stgraber> alkisg: can you try and get a list of those, maybe talking to mgariepy? Pretty much everything Revolution Linux found to be a problem with 12.04 has been fixed, but maybe there's a lack of communication there and getting more bugs properly targeted may help :)
| |
| 20:07 | <alkisg> sbalneav: e.g. gnome-panel not loading - https://bugs.launchpad.net/ubuntu/+source/gnome-panel/+bug/1078679
| |
| 20:07 | sbalneav: another e.g., sshfs mounts not working at all on second login, so a local /home/username is used, so users can't find their files, and changes are not saved on exit
| |
| 20:08 | <stgraber> vagrantc: any DM will let you login to the default server. The extra lightdm changes are to make the UI a bit nicer for LTSP.
| |
| 20:08 | <alkisg> (that was the bug I linked at the start of the chat)
| |
| 20:08 | <stgraber> alkisg: ah, I guess those are all fat clients related and RLNX doesn't use fat clients, which explain why it's working fine for them
| |
| 20:09 | <alkisg> stgraber: they don't use localapps either?
| |
| 20:10 | Login on a thin client, open a localapps firefox, logout, relogin => you no longer have an sshfs /home/username mount
| |
| 20:10 | <stgraber> alkisg: hmm, they're using localapps
| |
| 20:10 | <alkisg> And that happens too frequently to ignore... not all the time though
| |
| 20:11 | <stgraber> maybe some odd setup difference made it not happen for them, or more likely, people don't tend to logout and login again with the same user
| |
| 20:12 | <alkisg> True, that's why I haven't seen it on my initial ltsp workshops where teachers would only login once to see ltsp...
| |
| 20:12 | But it's been reported too many times on classroom use
| |
| 20:14 | <stgraber> alkisg: would a lazy umount fix it?
| |
| 20:14 | <alkisg> stgraber: we're doing a lazy unmount currently, it works for the unmount part but then .gvfs refuses to be mounted again on the second login
| |
| 20:15 | I'd feel really safer if we moved the cleanup part after X dies
| |
| 20:15 | We have frequent X crashes as well ever since the KMS move...
| |
| 20:16 | <stgraber> the obvious problem with that is that we don't have the environment after X dies, so you'd need quite a few tricks for that
| |
| 20:16 | <alkisg> We do have the screen-session.d/XS* scripts
| |
| 20:16 | XK
| |
| 20:16 | <stgraber> I believe we have xinit startup scripts currently, maybe you can implement xinit shutdown scripts
| |
| 20:16 | ah, we already have those?
| |
| 20:16 | cool :)
| |
| 20:17 | <alkisg> Yes, but the environment variables aren't there
| |
| 20:17 | With a shell-based ldm, they could persist more easily
| |
| 20:17 | <stgraber> so one way would be to have something in the ldm scripts dumping the environment to a file and then have that sourced from xinit
| |
| 20:17 | <alkisg> Right
| |
| 20:17 | But for that we'd need to postpone the ssh session closing
| |
| 20:18 | Because some of the cleanup stuff needs ssh access
| |
| 20:18 | <stgraber> right, but once you have the environment in the xinit kill scripts, there's no good reason not to move the code closing the ssh tunnel, so that's fine
| |
| 20:19 | <alkisg> For example: # Clean up localapps menu
| |
| 20:19 | ssh -S ${LDM_SOCKET} ${LDM_SERVER} rm -rf "${TMP_XDG_MENU}"
| |
| 20:19 | That should be done before the ssh tunnel is closed
| |
| 20:20 | And it should also be done even if X crashes
| |
| 20:22 | vagrantc mentioned that one of the reasons that ldm switched from shell to C, was security... any specifics there?
| |
| 20:23 | <stgraber> probably some concerns about the memory
| |
| 20:23 | <vagrantc> alkisg: one of the reasons it was implemented in python rather than shell...
| |
| 20:23 | <stgraber> in C we have control on whether variables are completely flushed or not
| |
| 20:23 | <vagrantc> (and later switched to C)
| |
| 20:31 | [GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Quit: Konversation terminated!) | |
| 20:35 | <alkisg> True, that could be a possible exploit, even if a very rare one...
| |
| 20:36 | I still don't see why the base of ldm can't be in shell though
| |
| 20:36 | The backends and the greeter, ok
| |
| 20:36 | JesseC has left IRC (JesseC!~JesseCWor@wsip-98-175-20-126.br.br.cox.net, ) | |
| 20:42 | <stgraber> anything that at some point will have the password go through it can't be shell, which in ldm is pretty much every part of it
| |
| 20:43 | <vagrantc> even passed through stdin/stdout?
| |
| 20:45 | <alkisg> Yeah... programs ran later on might happen to get that part of the memory
| |
| 20:46 | x=malloc(1 MB); check contents of x...
| |
| 20:47 | We can probably find some way to bypass that, e.g. PASSWD="some big string that is supposed to overwrite the stored password", but it's not guaranteed to be safe
| |
| 20:47 | <vagrantc> but why store passwords in variables at all?
| |
| 20:47 | <alkisg> It doesn't have to be in a variable
| |
| 20:47 | <vagrantc> ok
| |
| 20:47 | <alkisg> Even pipe memory can survive after the processes die
| |
| 20:48 | Although I'm not sure if we actually care about that in the current ldm code
| |
| 20:48 | (zeroing pipe memory...)
| |
| 20:51 | JesseC has joined IRC (JesseC!~JesseCWor@wsip-98-175-20-126.br.br.cox.net) | |
| 20:51 | Guest97645 has joined IRC (Guest97645!~JesseCWor@wsip-98-175-20-126.br.br.cox.net) | |
| 20:53 | <alkisg> "which in ldm is pretty much every part of it" ==> ldm.c doesn't access the password at all, only the backends and the greeter do
| |
| 20:59 | Guest97645 is now known as JesseC | |
| 21:11 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 21:14 | kb8wmc has joined IRC (kb8wmc!~chatzilla@nat.mtp.cmsinter.net) | |
| 21:15 | kb8wmc has left IRC (kb8wmc!~chatzilla@nat.mtp.cmsinter.net) | |
| 21:26 | telex has left IRC (telex!~telex@freeshell.de, Remote host closed the connection) | |
| 21:28 | MonkWitDaFunk has joined IRC (MonkWitDaFunk!~yaaic@199.119.234.213) | |
| 21:28 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 21:35 | MonkWitDaFunk has left IRC (MonkWitDaFunk!~yaaic@199.119.234.213, Ping timeout: 250 seconds) | |
| 21:51 | MonkWitDaFunk has joined IRC (MonkWitDaFunk!~yaaic@199-7-158-73.eng.wind.ca) | |
| 21:57 | MonkWitDaFunk has left IRC (MonkWitDaFunk!~yaaic@199-7-158-73.eng.wind.ca, Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org) | |
| 22:18 | markit has left IRC (markit!~marco@88-149-177-66.v4.ngi.it, ) | |
| 22:19 | MonkWitDaFunk has joined IRC (MonkWitDaFunk!~yaaic@199-7-158-73.eng.wind.ca) | |
| 22:33 | <vagrantc> alright...
| |
| 22:33 | raspberrypi booted to LTSP based on raspbian :)
| |
| 22:33 | (though nearly everything is broken, since i couldn't use an initramfs)
| |
| 22:35 | for some reason rpi developers seem averse to building kernels with initramfs support
| |
| 22:36 | need to hack in some writeable filesystem support into an init-ltsp.d hook
| |
| 22:39 | might also play with interesting tricks with kexec
| |
| 23:00 | meh. no aufs.
| |
| 23:16 | bobby_C has left IRC (bobby_C!~bobby@188.20.161.210, Ping timeout: 255 seconds) | |
| 23:33 | <vagrantc> and some of our screen-session.d scripts running lspci assume lspci will never error...
| |