IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 5 March 2014 (all times are UTC)
| 01:16 | gdi2k_ has left IRC (gdi2k_!~gdi2k@120.28.216.159, Ping timeout: 240 seconds) | |
| 01:29 | gdi2k_ has joined IRC (gdi2k_!~gdi2k@98.143.159.155) | |
| 01:35 | gdi2k_ has left IRC (gdi2k_!~gdi2k@98.143.159.155, Ping timeout: 240 seconds) | |
| 01:52 | gdi2k_ has joined IRC (gdi2k_!~gdi2k@120.28.216.159) | |
| 04:08 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 04:50 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 244 seconds) | |
| 04:51 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 04:52 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 04:52 | cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Remote host closed the connection) | |
| 04:53 | cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg) | |
| 05:02 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.) | |
| 05:07 | Ark74 has joined IRC (Ark74!~Ark74@189.214.42.128.cable.dyn.cableonline.com.mx) | |
| 05:25 | telex has left IRC (telex!~telex@freeshell.de, Remote host closed the connection) | |
| 05:26 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 05:26 | <cyberorg> alkisg, hi, wanted some opinions :) how many fat client can one server with giga lan work with optimally?
| |
| 05:28 | <alkisg> Hi cyberorg, there's a lot of caching locally if the clients have enough ram
| |
| 05:28 | So after the first launch of programs, the other launches are fast even on low speed networks
| |
| 05:29 | So let's say 10-30 for a single nic server, depending on what the users think of "fast" too...
| |
| 05:29 | But one can easily add more NICs or more NBD servers
| |
| 05:29 | <cyberorg> i was guessing about 100 client per server
| |
| 05:29 | <alkisg> (or aoe etc)
| |
| 05:29 | Opening libreoffice => 120 MB traffic
| |
| 05:31 | <cyberorg> for multiple NIC, they would have to be in different network, right?
| |
| 05:32 | <alkisg> No, multiple nics, one IP: https://help.ubuntu.com/community/UbuntuLTSP/Trunking
| |
| 05:33 | <cyberorg> ok, few years back bonding didn't well, will give that a try again
| |
| 05:34 | <alkisg> Also use pci-e NICs, not PCI
| |
| 05:35 | <cyberorg> we were using onboad NICs, not sure if they were pci-e
| |
| 05:37 | <alkisg> They're pci-e nowadays, but one PCI and others PCI-e is also fine
| |
| 05:38 | i(the pci bus limits network traffic for multiple NICs, while for one it's fine)
| |
| 05:41 | <Ark74> alkisg, so if i set several PCI NIC, only one will work?
| |
| 05:42 | i wasn't aware of that
| |
| 05:42 | <alkisg> http://en.wikipedia.org/wiki/List_of_device_bit_rates#Computer_buses
| |
| 05:42 | PCI 32-bit/33 MHz 1,067 Mbit/s
| |
| 05:42 | <cyberorg> alkisg, ok thanks, will pick your brain some more if we go ahead with this project, plan is to have 500+ fat clients
| |
| 05:43 | <alkisg> !local-boot
| |
| 05:43 | <ltsp`> local-boot: If you want LTSP fat clients on a low-speed network, you can put i386.img on e.g. C:\Boot\LTSP\i386.img and use this command line in pxelinux.cfg: APPEND ro initrd=ltsp/i386/initrd.img init=/sbin/init-ltsp root=/dev/sda1 rootflags=ro loop=/Boot/LTSP/i386.img; IPAPPEND 3
| |
| 05:44 | <alkisg> cyberorg: ^
| |
| 05:44 | Keep that in mind too, to have a bit of local storage
| |
| 05:45 | <Ark74> ohh!
| |
| 05:45 | alkisg, thanks to point it out ;)
| |
| 05:46 | <alkisg> np
| |
| 05:47 | <cyberorg> alkisg, i know about that, we also have nicer alternative to that provided by kiwi http://doc.opensuse.org/projects/kiwi/doc/#chap.pxe
| |
| 05:48 | <alkisg> cyberorg: what nicer alternative?
| |
| 05:48 | * alkisg wants to explore dmcache/bcache to do this automatically while the client is online... | |
| 05:48 | <alkisg> Without wasting any additional bandwidth at all
| |
| 05:50 | * alkisg waves, bb in 1h | |
| 05:50 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection) | |
| 05:51 | <cyberorg> i saw you mention inconvenience of copying image file to all the client, kiwi provides deploying the image, set up rw, either local partitions or nbd/aoe/nfs
| |
| 06:20 | work_alkisg is now known as alkisg | |
| 06:20 | <alkisg> nbd/aoe/nfs don't need deploying...
| |
| 06:21 | Deploying the image locally can be done with 1 command from epoptes, unrelated to the distro in use
| |
| 06:22 | The challenge here is to use dmcache/bcache to cache the image locally, and automatically, without wasting any additional network bandwidth at all
| |
| 06:23 | vagrantc: we still have months before jessie feature freeze, right? So, we do have time for more ltsp 5.5.x releases?
| |
| 06:26 | http://www.phoronix.com/scan.php?page=news_item&px=MTM4ODA
| |
| 06:34 | <cyberorg> alkisg, 1 command from epoptes for each client, deploying over pxe mean that if the local image is out of date it gets deployed again from the server
| |
| 06:35 | <alkisg> 1 command for all clients, you can select all of them
| |
| 06:35 | <cyberorg> alkisg, didn't quite understand how bcache would help
| |
| 06:35 | <alkisg> Basically, `dd if=/dev/nbd0 of=/path/to/local/image`
| |
| 06:35 | bcache, dmcache, enhanceio etc are caching mechanisms,
| |
| 06:36 | which mostly target caching to small SSD disks to speed up slower but bigger conventional disks
| |
| 06:36 | Our needs are a bit similar, we want to cache all the data we read from the network, locally, so that next time the client needs them, it will read them locally
| |
| 06:36 | So, for example, a client boot from NBD or AOE as usual,
| |
| 06:36 | it checks its local disk and finds no cache. It creates one
| |
| 06:37 | While booting, it reads 50 mb from that NBD image. Afterwards, when the user launches libreoffice, it reads another 120 mb because it's a fat client
| |
| 06:37 | The client writes those 170 MB to its local cache.
| |
| 06:38 | Whenever it needs them again, e.g. while booting next time, or launching libreoffice again, it reads them locally
| |
| 06:38 | At some point in the future, the NBD image gets updated
| |
| 06:38 | The client then discards the local cache and creates a new one
| |
| 06:38 | So, this results in being able to have 500+ fat clients on a single NIC...
| |
| 06:38 | Without having to deploy any image at all
| |
| 06:40 | <cyberorg> idea of pxe deployment is similar, first boot the image is written to local disk, all consequent boots happens locally, if the image on the server gets updated then the local image is written again
| |
| 06:41 | <alkisg> If the image is 1 GB, and you have 500 clients, you need 1 day to boot them...
| |
| 06:41 | A better option would be to copy the image while the client is already booted with the older local image
| |
| 06:43 | <cyberorg> yeah that would make updating image while users are still working and next boot they'd have new one
| |
| 06:43 | <alkisg> Right
| |
| 06:43 | But bcache makes all this useless, it's much better and requires no additional bandwidth
| |
| 06:45 | <cyberorg> how will that help with when booting 500 clients with 1gb image for the first time?
| |
| 06:45 | <alkisg> The clients only need 50 mb to boot, not 1 GB
| |
| 06:46 | So they'll read 50 MB, and that will get cached locally
| |
| 06:46 | If they ever read e.g. /usr/share/rare-program, then that will be cached too, if not, it won't,
| |
| 06:46 | while with transferring all the squashfs image, you transfer it even if you don't need it
| |
| 06:47 | <cyberorg> are you sure fat client needs just 50M to boot to full desktop?
| |
| 06:48 | <alkisg> Yes
| |
| 06:48 | To LDM
| |
| 06:49 | Anyways, bcache would help all ltsp users, but to be honest, for 500 clients I'd probably go for another file system that would support replication
| |
| 06:49 | xnbd, crfs, dunno
| |
| 06:50 | So that it's no longer server-based, but clients help in replication too
| |
| 06:59 | (08:42:00 πμ) alkisg: A better option would be to copy the image while the client is already booted with the older local image
| |
| 06:59 | ==> for example, if you use p2p technology there, like torrents, the image would get replicated a *lot* sooner
| |
| 06:59 | <cyberorg> i like the idea of using dm-cache to cache locally
| |
| 07:01 | <alkisg> It should also be easy to write a 'caching-nbd-client' that would add caching to nbd-client... and maybe even send it upstream, many people could need that
| |
| 07:02 | freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun) | |
| 07:03 | <alkisg> xnbd appears very active... https://bitbucket.org/hirofuchi/xnbd/commits/all maybe they could add that feature
| |
| 07:06 | This sounds more promising, for 500 clients: http://lab.openslx.org/projects/dnbd/wiki
| |
| 07:09 | <cyberorg> looks really interesting :)
| |
| 07:14 | alexxtasi has left IRC (alexxtasi!~alex@unaffiliated/alexxtasi) | |
| 07:15 | <cyberorg> hmm last commit on dnbd was in 2009
| |
| 07:36 | vmlintu has joined IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi) | |
| 07:39 | alexxtasi has joined IRC (alexxtasi!~alex@unaffiliated/alexxtasi) | |
| 07:54 | vmlintu has left IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 265 seconds) | |
| 08:05 | vmlintu has joined IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi) | |
| 08:06 | <vagrantc> alkisg: yeah, there's time for more 5.5.x ...
| |
| 08:06 | as far as i'm concerned
| |
| 08:07 | i assumed there would be, although work on ltsp6 needs to be ramped up a bit
| |
| 08:09 | ogra_` has joined IRC (ogra_`!~ogra_@p5098ed03.dip0.t-ipconnect.de) | |
| 08:10 | <markosu> I have problem. I try to figure out how to get guest autologin working (kiosk mode)? https://help.ubuntu.com/community/UbuntuLTSP/AutoLoginFeatures
| |
| 08:10 | ageis_ has joined IRC (ageis_!kevin@ageispolis.net) | |
| 08:11 | Patina has left IRC (Patina!~tomas@dhcp-5-103-223-5.seas-nve.net, *.net *.split) | |
| 08:11 | ageis has left IRC (ageis!kevin@ageispolis.net, *.net *.split) | |
| 08:11 | andygraybeal has left IRC (andygraybeal!~andy@h208.209.22.98.dynamic.ip.windstream.net, *.net *.split) | |
| 08:11 | ogra_ has left IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de, *.net *.split) | |
| 08:11 | shawnp0wers has left IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers, *.net *.split) | |
| 08:11 | <vagrantc> markosu: what have you tried, what hasn't worked?
| |
| 08:11 | <markosu> yep, I tried
| |
| 08:12 | What config I have to enable?
| |
| 08:12 | <vagrantc> depends on what your goal is
| |
| 08:13 | <markosu> goal is to get terminals autologin with guestaccounts
| |
| 08:13 | Do I need still add user fot LTS.conf?
| |
| 08:13 | <vagrantc> you want them to press a button to log in, or you want them to automatically log in?
| |
| 08:14 | <markosu> automatically log in would be the best option
| |
| 08:14 | <vagrantc> with either feature, it will by default attempt to log in as thee username with the password as the hostname ...
| |
| 08:14 | <markosu> So I have add user for LTS.conf?
| |
| 08:14 | <vagrantc> i.e. if the hostname is ltsp123, it will try to log in as user ltsp123, with password ltsp123
| |
| 08:14 | yes, the server needs to have the users present.
| |
| 08:15 | Ark74 has left IRC (Ark74!~Ark74@189.214.42.128.cable.dyn.cableonline.com.mx, Read error: Connection reset by peer) | |
| 08:15 | <vagrantc> you'll probably see failed login attempts in your logs
| |
| 08:16 | <markosu> ok, I was wondering that server generate automatic users for terminals
| |
| 08:17 | <alkisg> "Autologin" is a different concept from "kiosk"
| |
| 08:17 | <markosu> ok
| |
| 08:17 | <alkisg> LTSP does support both
| |
| 08:17 | <vagrantc> the kiosk feature runs everything locally ... kind of like a limited purpose fatclient
| |
| 08:18 | markosu: hope that helps. good luck!
| |
| 08:18 | * vagrantc waves | |
| 08:18 | <markosu> My goal is to setup LTSP-PNP system for older people without user account.
| |
| 08:18 | thanks
| |
| 08:18 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 08:18 | Patina has joined IRC (Patina!~tomas@dhcp-5-103-223-5.seas-nve.net) | |
| 08:19 | <markosu> When they logout everything information (browser history) should be wiped
| |
| 08:20 | <alkisg> That's autologin plus a script like this: http://sourceforge.net/p/ltsp/mailman/ltsp-discuss/thread/1310986533.1915.3.camel@alkis/
| |
| 08:21 | alkisg is now known as work_alkisg | |
| 08:22 | <markosu> thx
| |
| 08:36 | vmlintu has left IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 264 seconds) | |
| 08:45 | flo1546796 has joined IRC (flo1546796!~flo154679@2001:660:5001:156:26be:5ff:fe13:f2db) | |
| 08:45 | flo1546796 has joined IRC (flo1546796!~flo154679@unaffiliated/flo1546796) | |
| 09:01 | vmlintu has joined IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi) | |
| 09:01 | shawnp0wers has joined IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers) | |
| 09:01 | freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Ping timeout: 240 seconds) | |
| 09:02 | bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Ping timeout: 264 seconds) | |
| 09:03 | bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com) | |
| 09:16 | freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun) | |
| 09:32 | gdi2k_ has left IRC (gdi2k_!~gdi2k@120.28.216.159, Ping timeout: 252 seconds) | |
| 09:37 | gdi2k_ has joined IRC (gdi2k_!~gdi2k@120.28.216.159) | |
| 09:39 | cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Remote host closed the connection) | |
| 09:42 | work_alkisg is now known as alkisg | |
| 09:42 | cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg) | |
| 09:51 | flo1546796 has left IRC (flo1546796!~flo154679@unaffiliated/flo1546796, Quit: Quitte) | |
| 09:56 | flo1546796 has joined IRC (flo1546796!~flo154679@2001:660:5001:156:26be:5ff:fe13:f2db) | |
| 09:56 | flo1546796 has joined IRC (flo1546796!~flo154679@unaffiliated/flo1546796) | |
| 09:58 | andygraybeal has joined IRC (andygraybeal!~andy@h208.209.22.98.dynamic.ip.windstream.net) | |
| 10:00 | <cyberorg> alkisg, http://sourceforge.net/p/cyberorg-home/code/ci/master/tree/dm-cached-nbd/dm-cached-nbd.sh
| |
| 10:01 | <alkisg> cyberorg: does it work?
| |
| 10:01 | You can limit the bandwidth to the client to e.g. 1mbps to test better...
| |
| 10:02 | <cyberorg> alkisg, that is just a test script, tried it on this machine works
| |
| 10:02 | <alkisg> If, on the second boot, it goes fast on 1 mbps, then it works fine! :)
| |
| 10:02 | <cyberorg> would it be possible to add something like that to can you try something with that to init-ltsp.d scripts
| |
| 10:02 | alkisg, haven't tested with ltsp setup, just standalone pc
| |
| 10:02 | freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Quit: So long and thanks for all the fish) | |
| 10:03 | <alkisg> I think it would go to initramfs, not to init-ltsp.d
| |
| 10:04 | aufs/overlayfs would go on top of that
| |
| 10:05 | <cyberorg> that is the part i cannot do :)
| |
| 10:06 | <alkisg> That's distro-specific
| |
| 10:06 | Ah, you mean because it would be handled by kiwi code and not ltsp code?
| |
| 10:07 | <cyberorg> yes and also because i wouldn't know how to code that, small bits of script like this ok :)
| |
| 10:08 | will file enhancement bug for kiwi, lets see what schaefi thinks about it
| |
| 10:09 | * alkisg hopes we'll be able to do the aufs/overlayfs thing from init-ltsp.d in the future... so that there's no need for ltsp code in the initramfs at all | |
| 10:10 | ogra_` is now known as ogra | |
| 10:10 | <cyberorg> that would be ideal
| |
| 10:10 | ogra has left IRC (ogra!~ogra_@p5098ed03.dip0.t-ipconnect.de, Changing host) | |
| 10:10 | ogra has joined IRC (ogra!~ogra_@ubuntu/member/ogra) | |
| 10:18 | NeonLicht has left IRC (NeonLicht!~NeonLicht@darwin.ugr.es, Ping timeout: 252 seconds) | |
| 10:24 | NeonLicht has joined IRC (NeonLicht!~NeonLicht@darwin.ugr.es) | |
| 10:58 | adrianorg has left IRC (adrianorg!~adrianorg@187.113.248.237, Ping timeout: 265 seconds) | |
| 11:00 | adrianorg has joined IRC (adrianorg!~adrianorg@187.113.248.237) | |
| 11:25 | <alkisg> Using named NBD exports from kvm: kvm -hda nbd:localhost:10809:exportname=/opt/ltsp/i386 -cdrom /usr/lib/grub-rescue/grub-rescue-cdrom.iso -boot d
| |
| 11:25 | alkisg has left IRC (alkisg!~alkisg@plinet.ioa.sch.gr, Remote host closed the connection) | |
| 11:28 | work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr) | |
| 11:40 | <work_alkisg> xnbd-server --proxy [options] remote_host remote_port cache_disk_path cache_bitmap_path control_socket_path
| |
| 11:40 | ...cache_disk_path... :)
| |
| 11:50 | kwmiebach___ has left IRC (kwmiebach___!sid16855@gateway/web/irccloud.com/x-cvpfzbcvwkjzpzcl, Ping timeout: 265 seconds) | |
| 11:56 | kwmiebach____ has joined IRC (kwmiebach____!sid16855@gateway/web/irccloud.com/x-exrrahdrtjqhqltt) | |
| 11:56 | ncom0pl has left IRC (ncom0pl!~ncom0pl@89-71-65-172.dynamic.chello.pl, Remote host closed the connection) | |
| 11:58 | ncom0pl has joined IRC (ncom0pl!~ncom0pl@89-71-65-172.dynamic.chello.pl) | |
| 12:07 | khildin has joined IRC (khildin!~khildin@ip-213-49-116-239.dsl.scarlet.be) | |
| 13:17 | willianmazzardo has joined IRC (willianmazzardo!~textual@187.4.15.116) | |
| 13:56 | willianmazzardo has left IRC (willianmazzardo!~textual@187.4.15.116, Quit: Textual IRC Client: http://www.textualapp.com/) | |
| 14:24 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 14:44 | dsugar100 has joined IRC (dsugar100!~dsugar@columbia.tresys.com) | |
| 14:51 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection) | |
| 15:04 | <joe_sizlack> anyone using puppet and/or func with their ltsp setups?
| |
| 15:18 | alexxtasi has left IRC (alexxtasi!~alex@unaffiliated/alexxtasi) | |
| 15:24 | flo1546796 has left IRC (flo1546796!~flo154679@unaffiliated/flo1546796, Remote host closed the connection) | |
| 15:29 | adrianorg has left IRC (adrianorg!~adrianorg@187.113.248.237, Ping timeout: 265 seconds) | |
| 15:31 | adrianorg has joined IRC (adrianorg!~adrianorg@187.115.104.202) | |
| 15:35 | vmlintu has left IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 240 seconds) | |
| 15:37 | PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.155.245) | |
| 15:41 | Gremble has joined IRC (Gremble!~Ben@78-86-39-87.zone2.bethere.co.uk) | |
| 15:42 | Gremble is now known as Guest45964 | |
| 15:44 | adrianorg has left IRC (adrianorg!~adrianorg@187.115.104.202, Ping timeout: 252 seconds) | |
| 15:45 | adrianorg has joined IRC (adrianorg!~adrianorg@179.182.72.139) | |
| 15:51 | vmlintu has joined IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi) | |
| 15:52 | freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun) | |
| 15:53 | Guest45964 has left IRC (Guest45964!~Ben@78-86-39-87.zone2.bethere.co.uk, Quit: I Leave) | |
| 16:10 | mealstro1 has joined IRC (mealstro1!~Thunderbi@46.63.21.12) | |
| 16:10 | <mealstro1> hi
| |
| 16:10 | mealstro1 is now known as mealstrom | |
| 16:13 | <mealstrom> thats better. I've got some interesting problem with ltsp-server and thin client. Im using Ubuntu 12.04.04 i686. When i'm installing program on clients image --(chroot) -- this program didnt apears at links. If i launch it from terminal like ltsp-localapp blabla -- thats ok.
| |
| 16:13 | there is this program at ltsp.conf at LOCAL_APPS variable
| |
| 16:13 | any idea?
| |
| 16:18 | hydruid has joined IRC (hydruid!~hydruid@mail.twinriversheadstart.com) | |
| 16:19 | <hydruid> What OS supports LTSP clients the best?
| |
| 16:19 | I know that is a broad question, but I just deployed on about 5 computers and all of them seem to be having issues, used Ubuntu 12.04 as the host.
| |
| 16:26 | <khildin> joe_sizlack, interesting thought.... can you elaborate?
| |
| 16:29 | <hydruid> Also is there a recommendation for using the server OS or a desktop OS?
| |
| 16:34 | <khildin> hydruid, you want a graphical UI for your clients?... then it is self explaining.. :)
| |
| 16:34 | alkisg_web has joined IRC (alkisg_web!c23fefeb@gateway/web/freenode/ip.194.63.239.235) | |
| 16:35 | <hydruid> khildin: they are basically just going to rdesktop to a windows box
| |
| 16:35 | <khildin> mealstrom, hydruid ... ask the boss... he just joined... ;)
| |
| 16:36 | <hydruid> khildin: lol in the past I used the desktop version of Ubuntu and didn't run into many issues....this time I used the server version and had issues, going to try the desktop version out in just a bit :)
| |
| 16:36 | alkisg_web: good morning :)
| |
| 16:36 | <khildin> hydruid, wasn't there an option to immediately redirect to rdesktop?
| |
| 16:36 | <alkisg_web> Good afternoon :)
| |
| 16:36 | <hydruid> khildin: sure but I wanted it to be automatic
| |
| 16:37 | alkisg_web: I am having a ton of random issues with clients (black screen, freezes, and etc.) but I used the server version of Ubuntu as the host, does it make more sense to use the desktop version of Ubuntu instead?
| |
| 16:38 | <khildin> I take it you did have a look at TSP wiki? http://wiki.ltsp.org/wiki/Integration
| |
| 16:38 | LTSP*
| |
| 16:39 | <alkisg_web> Client issues are usually because of network issues or because of graphics driver issues... so no, the desktop or server variant doesn't matter much.
| |
| 16:39 | The kernel/xorg packages do matter...
| |
| 16:39 | <hydruid> khildin: yes I've seen that, I actually setup an rdesktop screen using the menu option
| |
| 16:39 | <alkisg_web> So stress your network and test with a live cd/usb stick too
| |
| 16:40 | <hydruid> alkisg_web: good to know....
| |
| 16:40 | alkisg_web: network is very small, at the main site here, they have 5 computers that are plugged into the server switch, which is 10Gbe
| |
| 16:40 | the network is rock solid, but I think it falls back to what you said, graphic driver issues
| |
| 16:41 | the clients they gave me to test are p4 single core
| |
| 16:41 | lol
| |
| 16:41 | <alkisg_web> p4 with 256 ram => ok thin clients, with 1gb ram => ok fat clients
| |
| 16:41 | <khildin> hydruid, a p4 should be more than sufficient
| |
| 16:41 | <hydruid> I was thinking about trying debian instead, as it might have better support with older hardwaer
| |
| 16:42 | *hardware
| |
| 16:42 | <alkisg_web> What graphics card? lspci -nn -k | grep -A 2 VGA, on the client
| |
| 16:42 | <khildin> although a P4 consumes a ton of energy.... :-/
| |
| 16:42 | <hydruid> alkisg_web: I got kicked out of the lab lol but I will try that in a bit
| |
| 16:42 | I tried setting XSERVER = vga
| |
| 16:43 | <alkisg_web> 12.04 supports p4 just fine. 14.04 has issues with sis and s3virge drivers.
| |
| 16:43 | There's no vga driver
| |
| 16:43 | <hydruid> ohhh that is no longer a valid option?
| |
| 16:43 | <alkisg_web> I don't think it ever was
| |
| 16:43 | XSERVER=vesa...
| |
| 16:43 | <khildin> s3virge... that is like.... ancient...
| |
| 16:44 | <hydruid> alkisg_web: ty I will see if that helps
| |
| 16:44 | <alkisg_web> We still have a lot of 1998-2000 graphics cards here... hundreds of them
| |
| 16:44 | <hydruid> most of the clients show the ubuntu splash screen, sit there for a few seconds, then go black
| |
| 16:44 | <khildin> alkisg_web, ever tried a larger setup using RPI as client?
| |
| 16:45 | <alkisg_web> We don't buy thin clients, we only use old pcs as thin clients. We buy fat clients instead...
| |
| 16:45 | pi is good only as a thin client...
| |
| 16:45 | <khildin> yes... but they ar cheap and use very littly powr
| |
| 16:45 | <hydruid> Is Sound_Daemon = false still valid?
| |
| 16:48 | <alkisg_web> khildin: there are quad core arm devices that are suitable for fat clients and also are cheap and waste little power...
| |
| 16:49 | hydruid: SOUND=False
| |
| 16:49 | <khildin> Sound_Daemon: Values are esd, nasd, and pulse (default).
| |
| 16:50 | <hydruid> alkisg_web: cool ty, do I need to install the vesa driver in the chroot?
| |
| 16:52 | <alkisg_web> No, but you shouldn't have to use the vesa driver anyway
| |
| 16:55 | <hydruid> alkisg_web: gotcha ty....so it's recommended to identify the video card and then set XSERVER to match if it the automatic config doesn't work?
| |
| 16:55 | <alkisg_web> No, automatic config usually works fine
| |
| 16:55 | The problem is in the driver version
| |
| 16:56 | So if you have graphics issues, you need to find a kernel/xorg combination where that card works fine
| |
| 16:56 | <hydruid> I see
| |
| 16:56 | <alkisg_web> E.g. intels 8xx worked best in ubuntu 9.04, the completely broke in 10.04, and since 12.04 they mostly work, with rare crashes
| |
| 16:57 | <hydruid> fun
| |
| 16:57 | lol
| |
| 16:57 | <alkisg_web> That's not an Ubuntu issue, it's an upstream issue, so the same happens in all distros, at the respective versions...
| |
| 16:57 | vmlintu has left IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 240 seconds) | |
| 16:58 | <alkisg_web> The good thing with Ubuntu LTS editions is that it's very easy to try different kernels/xorg versions
| |
| 16:58 | xserver-xorg-lts-saucy etc
| |
| 16:58 | And by installing the debian 486 kernel one can support even non-pae clients...
| |
| 16:59 | So even P3's are fine
| |
| 17:08 | <hydruid> alkisg_web: okay was able to run that command
| |
| 17:08 | looks like a radeon, RC410
| |
| 17:09 | <alkisg_web> pci id?
| |
| 17:09 | <hydruid> subsystem: intel corp device {8086:d600}
| |
| 17:09 | ChadLepto has left IRC (ChadLepto!~chadlepto@unaffiliated/chadlepto, Quit: Leaving) | |
| 17:09 | <alkisg_web> Also verify that you run that locally and not on the server
| |
| 17:09 | !localxterm
| |
| 17:09 | <ltsp`> localxterm: Any applications that you launch on a thin client actually runs on the server, not on the client itself. If you want to open a program on the client locally, you can type 'ltsp-localapps <program>' in a run dialog or in a terminal. For example, 'ltsp-localapps xterm' to open a terminal running on the client.
| |
| 17:09 | <hydruid> it is local
| |
| 17:10 | radeon xpress 200/1100 is what it is
| |
| 17:10 | and the xorg error was = no screens found
| |
| 17:10 | let me see if I can find the pci id
| |
| 17:11 | <alkisg_web> lspci -nn -k | grep -A 2 VGA
| |
| 17:13 | Also "no screens found", is that with an empty lts.conf?
| |
| 17:14 | <hydruid> alkisg_web: yes I found it, pci id is: 01:05.0
| |
| 17:14 | <alkisg_web> 1234:5678
| |
| 17:14 | What you pasted is not a pci id
| |
| 17:15 | ~$ lspci -nn -k | grep -A 2 VGA
| |
| 17:15 | 00:02.0 VGA compatible controller [0300]: Intel Corporation 82G33/G31 Express Integrated Graphics Controller [8086:29c2] (rev 10)
| |
| 17:15 | <hydruid> ohhhh I see
| |
| 17:15 | <alkisg_web> The pci id is 8086:29c2
| |
| 17:15 | And the kernel driver in use should say radeon
| |
| 17:15 | <hydruid> 1002:5a61
| |
| 17:16 | alkisg_web: it does that say that for the kernel driver
| |
| 17:16 | <alkisg_web> And what's the output of: getltscfg -a
| |
| 17:16 | locally
| |
| 17:16 | <hydruid> alkisg_web: lts.conf is not blank......
| |
| 17:17 | freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Quit: So long and thanks for all the fish) | |
| 17:18 | <hydruid> alkisg_web: it printed the contents of lts.conf
| |
| 17:18 | <alkisg_web> ...which are..?
| |
| 17:19 | <hydruid> alkisg_web: http://pastebin.com/PCFqR2Zh
| |
| 17:22 | <alkisg_web> You forced vesa, so it might be why it can't find any screens.
| |
| 17:22 | <hydruid> let me try without
| |
| 17:22 | <mealstrom> hydruid: tnx )) , havant been here for ages.
| |
| 17:22 | <alkisg_web> Try with an empty lts.conf
| |
| 17:22 | Completely empty
| |
| 17:22 | <hydruid> kk
| |
| 17:22 | do I have to ltsp-update-image after lts.conf changes?
| |
| 17:23 | <alkisg_web> Not if you change the one in tftp...
| |
| 17:23 | <mealstrom> alkisg_web: I've got some interesting problem with ltsp-server and thin client. Im using Ubuntu 12.04.04 i686. When i'm installing program on clients image --(chroot) -- this program didnt apears at links. If i launch it from terminal like ltsp-localapp blabla -- thats ok. there is this program at ltsp.conf at LOCAL_APPS variable. any idea?
| |
| 17:24 | <alkisg_web> You need LOCAL_APPS_MENU etc, check the lts.conf manpage
| |
| 17:26 | <hydruid> alkisg_web: cheers it worked!
| |
| 17:29 | alkisg_web: ty vm
| |
| 17:29 | alkisg_web: next question :0
| |
| 17:30 | alkisg_web: on the ubuntu splash screen, if I don't hit ESC on the keyboard, the screen goes black and nothing happens
| |
| 17:30 | if I do hit ESC, then it works perfectly
| |
| 17:30 | <alkisg_web> !quiet-splash
| |
| 17:30 | <ltsp`> quiet-splash: to disable the splash screen in Ubuntu, in order to see any boot error messages, run sudo gedit /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default and remove quiet splash plymouth:force-splash vt.handoff=7
| |
| 17:30 | <alkisg_web> So remove all those permanately :)
| |
| 17:31 | *permanently
| |
| 17:31 | <hydruid> alkisg_web: ty!
| |
| 17:31 | alkisg_web: I tried that but missed the plymouth
| |
| 17:33 | alkisg_web: again ty very much, I owe you 500 beers
| |
| 17:34 | <alkisg_web> :) /me will get fat and drunk
| |
| 17:35 | <hydruid> lol
| |
| 17:39 | <ogra> just open a bar :)
| |
| 17:39 | <hydruid> :P
| |
| 17:41 | <mealstrom> alkisg_web: heh, as always forget about enabling local_apps
| |
| 17:43 | <effenberg> free beer?
| |
| 17:44 | * alkisg_web waves | |
| 17:44 | alkisg_web has left IRC (alkisg_web!c23fefeb@gateway/web/freenode/ip.194.63.239.235, Quit: Page closed) | |
| 17:48 | <mealstrom> someone from Ukraine or Russian?
| |
| 17:51 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 17:52 | <hydruid> effenberg: you know it!
| |
| 17:56 | <PhoenixSTF> hello guys I am trying to setup a ltsp on a KVM virtmachine, its is Xubuntu amd64, I can boot with a thin client but console 7 is all black while other have normal CLI login screen, anyone knows how to fix this?
| |
| 17:59 | hydruid has left IRC (hydruid!~hydruid@mail.twinriversheadstart.com, Read error: Connection reset by peer) | |
| 18:00 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 18:05 | <mealstrom> PhoenixSTF: are you using server edition ?
| |
| 18:06 | <PhoenixSTF> mealstrom, no just Xubuntu 13.10
| |
| 18:06 | with ltsp-standalone
| |
| 18:06 | mealstrom, desktop why should I user server?
| |
| 18:08 | <alkisg> PhoenixSTF: put these in your lts.conf:
| |
| 18:08 | [Default]
| |
| 18:08 | SCREEN_02=shell
| |
| 18:08 | SCREEN_07=ldm
| |
| 18:08 | Then reboot client and check your xorg logs from screen 02
| |
| 18:08 | Also do this:
| |
| 18:08 | !quiet-splash
| |
| 18:08 | <ltsp`> quiet-splash: to disable the splash screen in Ubuntu, in order to see any boot error messages, run sudo gedit /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default and remove quiet splash plymouth:force-splash vt.handoff=7
| |
| 18:10 | <PhoenixSTF> alkisg, update image?
| |
| 18:14 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 240 seconds) | |
| 18:14 | <PhoenixSTF> crap it stops at saned
| |
| 18:14 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 18:15 | <PhoenixSTF> alkisg, it stops at saned disabled; edit /etc.....
| |
| 18:15 | nbd0 unkown partition taqble
| |
| 18:15 | alkisg1 has joined IRC (alkisg1!~alkisg@ubuntu/member/alkisg) | |
| 18:17 | <alkisg1> PhoenixSTF: did you do the lts.conf changes? if yes, press alt+ctrl+f2...
| |
| 18:17 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Disconnected by services) | |
| 18:17 | alkisg1 is now known as alkisg | |
| 18:17 | vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc) | |
| 18:18 | <PhoenixSTF> yes I am there
| |
| 18:18 | <alkisg> getltscfg -a
| |
| 18:18 | Do you see your lts.conf settings with that command?
| |
| 18:18 | <PhoenixSTF> yes
| |
| 18:19 | <alkisg> And your lts.conf *only* has SCREEN_02 and SCREEN_07?
| |
| 18:19 | <PhoenixSTF> nop
| |
| 18:19 | <alkisg> Put only those, to make sure you don't break the booting with settings from lts.conf
| |
| 18:19 | Then check the output of this command: lspci -nn -k | grep -A2 VGA
| |
| 18:19 | on the client
| |
| 18:19 | <PhoenixSTF> RM_SYSTEM_SERVICES=2dnsmasq nbd-server nfs-kernel-server ssh-whopsie"
| |
| 18:20 | so I disable all extras
| |
| 18:20 | <alkisg> OK, those shouldn't matter
| |
| 18:20 | As long as you don't have things like XSERVER=vesa, autologin etc
| |
| 18:21 | <PhoenixSTF> no none of whose only something with DIRECTX
| |
| 18:21 | <alkisg> What's the output of the lspci command above?
| |
| 18:22 | mealstrom has left IRC (mealstrom!~Thunderbi@46.63.21.12, Ping timeout: 265 seconds) | |
| 18:22 | <PhoenixSTF> no such pci access methos
| |
| 18:22 | <alkisg> Did you make a typo?
| |
| 18:23 | stgraber: if you sync ltsp from debian, it'd be nice if you also synced the last 2 commits from ltsp-trunk
| |
| 18:23 | <PhoenixSTF> yes I did pt keyboard to us :(
| |
| 18:24 | VGA compatible cirrus logic GD 3446
| |
| 18:24 | *5446
| |
| 18:24 | <alkisg> That's the server
| |
| 18:24 | Not the client
| |
| 18:24 | Or are you using a VM for a client?
| |
| 18:24 | <PhoenixSTF> yes I am
| |
| 18:24 | * vagrantc looks at ltsp-trunk | |
| 18:25 | <alkisg> Try with kvm -vga vmware. Or with a real client.
| |
| 18:26 | <PhoenixSTF> -_-
| |
| 18:27 | seriously so many ours and had to bother you guys just to change the video...
| |
| 18:27 | *hours
| |
| 18:27 | <alkisg> :)
| |
| 18:27 | <PhoenixSTF> thanks m8
| |
| 18:27 | really sorry
| |
| 18:27 | <alkisg> np
| |
| 18:28 | <vagrantc> alkisg: so, should i merge those two patches from ltsp-trunk for debian as well?
| |
| 18:28 | i guess one is ubuntu-specific...
| |
| 18:28 | <alkisg> vagrantc: one of them is ubuntu-only,
| |
| 18:28 | <PhoenixSTF> btw anyone got a report on savings with Greek schools change to LTSP?
| |
| 18:28 | <alkisg> and the other one is for udhcpc, which debian users don't use much...
| |
| 18:28 | <vagrantc> right
| |
| 18:28 | bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Quit: http://www.twelvetribes.org) | |
| 18:28 | <alkisg> vagrantc: so I don't think there's any hurry to upload them to debian. And the timeout only affects 10% of the users
| |
| 18:29 | PhoenixSTF: monetary savings?
| |
| 18:29 | <vagrantc> alkisg: would probably be good to get into jessie sometime in the next few months
| |
| 18:29 | <alkisg> We'll have more things to upload till then :)
| |
| 18:30 | bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com) | |
| 18:30 | <vagrantc> i've been thinking about rewriting the debian-installer support a bit lately, so if i do some of that, might be worth redoing
| |
| 18:30 | worth another upload
| |
| 18:31 | stgraber: i've been thinking about making the debian-installer support more plugin based or something, to reduce the need for us to diverge so much
| |
| 18:31 | stgraber: plugins, or overridable functions, or whatever.
| |
| 18:33 | * vagrantc notes that https://launchpad.net/ubuntu/+source/ltsp is kind of messed up | |
| 18:37 | <alkisg> vagrantc: since the stackoverflow attempt wasn't successful, do you think we should enable answers in https://launchpad.net/ltsp ?
| |
| 18:37 | Like https://answers.launchpad.net/epoptes ?
| |
| 18:37 | * alkisg doesn't like mailing lists much, they're not searchable/linkable etc | |
| 18:38 | <vagrantc> i find forums to be impossible to find anything useful
| |
| 18:38 | but i am not the target audience
| |
| 18:38 | i won't likely be actively involved either way
| |
| 18:39 | <alkisg> It's not forums, it's questions
| |
| 18:39 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.) | |
| 18:39 | <vagrantc> it's all webbity
| |
| 18:40 | <alkisg> The idea is to have a place where the devs can answer questions and then reuse their answers
| |
| 18:40 | <vagrantc> but at least the launchpad questions stuff sends me emails
| |
| 18:40 | <alkisg> Reusing stuff is very difficult in mailing lists...
| |
| 18:40 | <vagrantc> so it's a dynamically generated FAQ, essentially
| |
| 18:40 | ?
| |
| 18:41 | <alkisg> Pretty similar, yeah
| |
| 18:41 | * vagrantc looks at the web ui | |
| 18:41 | <alkisg> Check https://answers.launchpad.net/epoptes as an example
| |
| 18:42 | * vagrantc finds it funny that answers.launchpad.net is titled "questions" | |
| 18:42 | <vagrantc> it seems like just another bug reporting sort of interface...
| |
| 18:43 | with different statuses than you'd see for bug reports
| |
| 18:43 | <alkisg> Right, the main difference is the allowed subjects
| |
| 18:43 | One can't open an ltsp bug report on "please advice what server to buy"
| |
| 18:43 | But he can open a question for that
| |
| 18:44 | stackexchange does have a better interface for questions
| |
| 18:44 | But it appears not enough people were interested...
| |
| 18:45 | <vagrantc> right
| |
| 18:45 | i don't see much problem with enabling answers.launchpad.net for ltsp ... as long as people can opt-out of the responses
| |
| 18:46 | <alkisg> But you're not going to use it, right?
| |
| 18:46 | <vagrantc> it depends... i do read the epoptes generated emails
| |
| 18:46 | <alkisg> (I think it's possible to answer by email too)
| |
| 18:47 | <vagrantc> i'd at least know what sort of questions people were asking
| |
| 18:47 | <alkisg> OK, let's think it over for a few days, it's very easy to enable it but I think it's a bit hard to disable it afterwards
| |
| 18:47 | <vagrantc> heh
| |
| 18:48 | you can configure which groups get the emails? if it was all the ~ltsp-maintainers or whatever group then i think we should ask them.
| |
| 18:49 | <alkisg> I think it supports setting the "answers contact"
| |
| 18:49 | We can leave that empty to notify noone by default, or put ltsp-upstream there, etc
| |
| 18:49 | The ubuntu-specific answers are automatically already enabled
| |
| 18:50 | https://answers.launchpad.net/ubuntu/+source/ltsp
| |
| 18:50 | https://answers.launchpad.net/ubuntu/+source/ltspfs, ldm etc
| |
| 18:52 | work_alkisg has left IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr, Remote host closed the connection) | |
| 18:56 | <vagrantc> alkisg: well, i'm ok with it, and would actually like to get the emails... unless it overwhelms me.
| |
| 18:57 | work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr) | |
| 19:05 | <Leolo_3> switch_root: failed to unlink unionfs: Directory not empty
| |
| 19:05 | is that ^^ something I should worry about?
| |
| 19:08 | <alkisg> What distro/version is that?
| |
| 19:09 | <Leolo_3> CentOS 6
| |
| 19:10 | LTSP 5.4
| |
| 19:10 | <alkisg> It does sound like something you should report to the distro maintainers...
| |
| 19:11 | Does it still work after that?
| |
| 19:11 | <Leolo_3> still works
| |
| 19:12 | I don't get that message when I boot the distro "normally"
| |
| 19:12 | and CentOS doesn't have a maintainer for LTSP, from what I understand
| |
| 19:14 | <alkisg> Someone should pick up, then... :)
| |
| 19:16 | <PhoenixSTF> alkisg, sorry m8 to bother again, but I have been trying to login and it stucks there trying to load
| |
| 19:16 | <alkisg> Try ssh <user>@server from screen02
| |
| 19:16 | Replace user, leave server exactly as it is
| |
| 19:16 | Don't put ip, don't change the hostname
| |
| 19:18 | <PhoenixSTF> ok I added a user with root
| |
| 19:18 | but cannot login with it
| |
| 19:18 | <alkisg> You do add the user to the server, not to the chroot, right?
| |
| 19:19 | <PhoenixSTF> yes with chroot
| |
| 19:19 | let me try and reboot
| |
| 19:19 | about befora yes monetary savings
| |
| 19:20 | <alkisg> PhoenixSTF: *don't* add the user to the chroot
| |
| 19:20 | Add the user to the server
| |
| 19:20 | <PhoenixSTF> ok
| |
| 19:21 | any groups I should add it?
| |
| 19:21 | <alkisg> No
| |
| 19:21 | <PhoenixSTF> update-image?
| |
| 19:22 | <alkisg> no
| |
| 19:23 | <PhoenixSTF> no joy
| |
| 19:23 | <alkisg> You can't login with ssh?
| |
| 19:23 | <PhoenixSTF> with ssh I can with thin client I cannot
| |
| 19:23 | tried on console 1
| |
| 19:24 | <alkisg> Don't. Try on screen_02=shell
| |
| 19:24 | !screen_02
| |
| 19:24 | <ltsp`> screen_02: To get a root shell on an Ubuntu thin client: https://help.ubuntu.com/community/UbuntuLTSP/ClientTroubleshooting#Using_a_shell_SCREEN
| |
| 19:24 | <PhoenixSTF> ok i am on sroot shell
| |
| 19:26 | gbaman has joined IRC (gbaman!~gbaman@81.134.121.245) | |
| 19:32 | <PhoenixSTF> is it possible I dont have any desktop environment installed?
| |
| 19:36 | <alkisg> (09:16:30 μμ) alkisg: Try ssh <user>@server from screen02
| |
| 19:36 | Did that work?
| |
| 19:37 | vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi) | |
| 19:37 | <PhoenixSTF> yes it did
| |
| 19:39 | <alkisg> OK then, check /usr/share/xsessions in your server
| |
| 19:41 | <PhoenixSTF> xfce.desktop and xubuntu.desktop
| |
| 19:41 | *
| |
| 19:42 | <alkisg> So you do have sessions on your server
| |
| 19:42 | Check /var/log/auth.log on your server after an ldm-login try
| |
| 19:42 | <PhoenixSTF> yes its a normal desktop distro
| |
| 19:44 | there is no ldm-login
| |
| 19:49 | <vmlintu> alkisg: You mentioned lpxelinux some time ago and I did now some boot tests over HTTP with lpxelinux.0 and it seems to work much better than gpxe/ipxe.. Haven't had a chance to try with a wide range of NICs yet, but at least for now it looks good..
| |
| 19:51 | <alkisg> Cool, so serving the kernel over http is easier with that :)
| |
| 19:51 | <vagrantc> whee
| |
| 19:52 | <vmlintu> alkisg: yes, it loads only lpxelinux.0 over tftp, everything else went over http. I was using now nginx and static files for tests..
| |
| 19:52 | <PhoenixSTF> alkisg, do you recommend any tutorial to I can try and put it up to work without bothering you too much, I obviously broke something here.
| |
| 19:53 | <alkisg> vmlintu, vagrantc, we could also dynamically generate pxelinux.cfg/default from lts.conf entries with ltspd!
| |
| 19:54 | PhoenixSTF: I don't know of any tutorials, many exists but I haven't tried any of them recently, except from the ltsp-pnp one which I wrote...
| |
| 19:56 | <vmlintu> alkisg: yes.. and if it checks for the mac address from pxelinux.cfg/01-<mac>, one can also define kernel parameters for each client..
| |
| 19:57 | <alkisg> vmlintu: I mean that we no longer will need anything from pxelinux.cfg/*
| |
| 19:57 | All the kernel parameters would go inside lts.conf
| |
| 19:57 | [mac-address]
| |
| 19:57 | KERNEL_CMDLINE="nomodeset"
| |
| 19:57 | And ltspd would translate that to the appropriate pxelinux.cfg/default, which would be the http reply, not a real file...
| |
| 19:58 | <vmlintu> yes.. we are already got rid of pxelinux.cfg/* with puavo-tftpd and lpxelinux.0 makes it even simpler
| |
| 19:58 | <alkisg> Is there any need for a custom tftp server, when one is using lpxelinux?
| |
| 19:58 | The only thing downloaded is lpxelinux.0, isn't it?
| |
| 19:59 | *via tftp...
| |
| 19:59 | <vmlintu> no, tftpd-hpa should be ok if lpxelinux.0 works with all clients
| |
| 20:00 | <vagrantc> or dnsmasq
| |
| 20:00 | there are clients it doesn't work for?
| |
| 20:00 | alkisg: also, serving up pxelinux.cfg/01-<mac> could be generated and served without it landing with all the fallbacks...
| |
| 20:01 | would be fewer hits to the tftp/http server
| |
| 20:01 | <vmlintu> I haven't done much testing yet.. earlier when I tried to use the http support in gpxe/ipxe, there were some NICs that didn't work
| |
| 20:01 | <alkisg> ipxe only supports certain nics, but lpxelinux should support all the ones that offer a pxe stack
| |
| 20:02 | <vmlintu> I tested with one client that didn't work gpxe and at least it worked now ok, so I'm hopeful
| |
| 20:02 | <alkisg> So I think only a few broken ones wouldn't work
| |
| 20:02 | <vagrantc> gpxe isn't still developed, is it?
| |
| 20:02 | <alkisg> Not really, only ipxe is
| |
| 20:02 | <vagrantc> right
| |
| 20:03 | <vmlintu> I took the lpxelinux.0 file from debian experimental packages
| |
| 20:03 | <vagrantc> hrm.
| |
| 20:03 | need to get that migrating into jessie...
| |
| 20:04 | <vmlintu> Yes, always serving the 01-<mac> file does speed up things a bit at least with puavo-tftpd
| |
| 20:07 | telex has left IRC (telex!~telex@freeshell.de, Remote host closed the connection) | |
| 20:08 | <vmlintu> I'll do probably more testing next week.. I need to find some place that has a wide variety nics to test..
| |
| 20:08 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 20:16 | gdi2k_ has left IRC (gdi2k_!~gdi2k@120.28.216.159, Remote host closed the connection) | |
| 20:18 | mealstrom has joined IRC (mealstrom!~Thunderbi@46.63.63.163) | |
| 20:19 | ball has joined IRC (ball!ball@162-202-67-158.lightspeed.livnmi.sbcglobal.net) | |
| 20:19 | <ball> Good morning!
| |
| 20:32 | vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Ping timeout: 252 seconds) | |
| 20:46 | PhoenixSTF has left IRC (PhoenixSTF!~rudi@78.29.155.245, Ping timeout: 240 seconds) | |
| 20:47 | <dead_inside> morning ball
| |
| 20:54 | <gbaman> alkisg: can sch-scripts be installed onto a debian server?
| |
| 21:14 | <alkisg> gbaman: I don't think so
| |
| 21:14 | They also contain a lot of greek settings
| |
| 21:15 | So one shouldn't try to install them. One can try to i18n them though by creating a similar, but changed, package
| |
| 21:16 | <gbaman> any reason cant run entire thing through a translator?
| |
| 21:16 | <alkisg> It shouldn't take more than a few hours to translate sch-scripts to english and to make them run in debian...
| |
| 21:19 | <gbaman> btw, alkisg and vagrantc, the teacher trialing LTSP fat clients on pis for me has uploaded a few videos of his classes trying them out https://www.youtube.com/watch?v=ZCxnMZuCyzM#t=616
| |
| 21:19 | <vagrantc> cool
| |
| 21:20 | <alkisg> They already have PCs, yet they're starting to use PIs? :D
| |
| 21:21 | <gbaman> well, in this case, yes
| |
| 21:21 | mainly because they want to make use of applications only available for the pi
| |
| 21:22 | in this case, minecraft Pi edition
| |
| 21:22 | <alkisg> Isn't that java? shouldn't it run anywhere?
| |
| 21:22 | <gbaman> no, minecraft pi edition
| |
| 21:22 | <alkisg> And?
| |
| 21:23 | <gbaman> custom version of minecraft written by Mojang for the Raspberry pi
| |
| 21:23 | <alkisg> They re-programmed it without java?
| |
| 21:23 | <gbaman> based off pocket edition
| |
| 21:23 | <alkisg> If it's java, it should run anywhere...
| |
| 21:23 | <gbaman> which is written i think, in C++?
| |
| 21:23 | they dont release the source though
| |
| 21:23 | <alkisg> I think it would just be a lighter version, in java again
| |
| 21:23 | Runnable anywhere
| |
| 21:23 | <gbaman> the cool thing they did with minecraft pi edition though was add a nice easy to use api
| |
| 21:23 | <alkisg> I don't think they'd rewrite minecraft in C++ just for pi
| |
| 21:24 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 21:24 | <gbaman> available for any language that supports sockets
| |
| 21:24 | as I said, based off pocket edition (version for IOS and android)
| |
| 21:24 | which was completely rewritten
| |
| 21:24 | but pi edition has an API
| |
| 21:24 | which is rather easy to use in python
| |
| 21:25 | so this teacher is planning next year to move to it to do GCSE CS coursework on
| |
| 21:25 | with minecraft pi..
| |
| 21:26 | the API lets students do stuff like place blocks, get block, move player etc
| |
| 21:27 | So, for him, having 20 pis in the room was going to be a headache for managing their OS, users and most importantly, user files
| |
| 21:27 | seeing as SD cards are nowhere near as reliable and HDDs
| |
| 21:27 | and loosing GCSE CS coursework isnt an option
| |
| 21:27 | <ball> Can a Pie PXE?
| |
| 21:28 | <gbaman> no
| |
| 21:28 | but one can put the kernel on the sd card and boot with it
| |
| 21:28 | needs about 30mb of sd card
| |
| 21:28 | <ball> Does it mount the SD card read-only?
| |
| 21:28 | <gbaman> yes
| |
| 21:28 | <vagrantc> there are patches in a u-boot branch to support booting off the network and reading a pxelinux configuration file
| |
| 21:28 | <ball> That's not too bad then.
| |
| 21:29 | <vagrantc> although unless you disable localdevs, it'll be writeable to users
| |
| 21:29 | <gbaman> oops vagrantc
| |
| 21:30 | I am not too worried though
| |
| 21:30 | actually, it is a good thing that they can
| |
| 21:30 | <vagrantc> yeah, then they can write their own kernel!
| |
| 21:30 | <gbaman> allows them to swap the config files allowing them to switch between booting off local sd card and network
| |
| 21:31 | which was one thing I did want to talk to you about...
| |
| 21:31 | have I already talked to you about some way to switch them vagrantc?
| |
| 21:31 | <vagrantc> you just change stuff in config.txt or cmdline.txt
| |
| 21:32 | <gbaman> yes, but would rather love something like hold down a key on boot?
| |
| 21:32 | it swaps file and reboots?
| |
| 21:34 | * alkisg doesn't like promotional programs that only run on selective platforms | |
| 21:34 | <alkisg> I'd go for something else for my CS courses if minecraft-pi is only for pi :)
| |
| 21:35 | Fortunately there are many open source programming environments available...
| |
| 21:35 | <gbaman> but any as cool as Minecraft?
| |
| 21:35 | <alkisg> Sure, cooler too
| |
| 21:35 | 3d in CS was there before minecraft existed
| |
| 21:35 | <vagrantc> gbaman: i don't think there's anything like a boot menu, no.
| |
| 21:36 | gbaman: you could write a script which does that from the booted system
| |
| 21:36 | and then reboot...
| |
| 21:36 | <gbaman> perhaps alkisg, this teacher wants to though :)
| |
| 21:36 | vagrantc: what about if the server isnt available?
| |
| 21:36 | * ball thinks back to raytracing on an 8 MHz Atari ST. | |
| 21:37 | <gbaman> need some way if the student takes it home
| |
| 21:37 | <vagrantc> gbaman: sounds like you've got some ARM assembling coding to do.
| |
| 21:37 | <alkisg> The teachers here want windows 7 + ms office 201x + photoshop + corel + ... and the organizations that manage the schools just tell them that they don't have money for those, and to settle with linux
| |
| 21:37 | <vagrantc> er, assembly
| |
| 21:38 | <gbaman> kind of guessed that vagrantc :(
| |
| 21:38 | <vagrantc> maybe C
| |
| 21:38 | <ball> ARM assembly confused the hell out of me when I first saw it.
| |
| 21:38 | <gbaman> alkisg: windows and office are the norm here
| |
| 21:38 | <ball> I was used to implicit destinations for most instructions.
| |
| 21:38 | <gbaman> this teacher isnt allowed to install linux on those desktop machines
| |
| 21:38 | <ball> ...so seeing three registers for an add instruction for example.
| |
| 21:39 | ...was something I wasn't used it.
| |
| 21:39 | <gbaman> nor is he allowed to plug anything into their school network
| |
| 21:39 | <ball> s/it/to/
| |
| 21:39 | <vagrantc> gbaman: you could also do it from the initramfs, if it always boots the same initramfs. i think i already suggested that to you before, now that i think of it... :P
| |
| 21:39 | <alkisg> We don't allow anyone here to not allow stuff, it's against the laws :)
| |
| 21:39 | <gbaman> how hard would it be vagrantc?
| |
| 21:39 | liking that alkisg :)
| |
| 21:40 | <vagrantc> gbaman: i think it would be 3 hard.
| |
| 21:40 | <gbaman> ?
| |
| 21:40 | <alkisg> On a scale of 10? :P
| |
| 21:40 | * vagrantc will not disclose the scale | |
| 21:41 | <gbaman> ..
| |
| 21:41 | helpful...
| |
| 21:41 | :)
| |
| 21:41 | * ball grins | |
| 21:41 | <vagrantc> gbaman: i don't know your coding skills to know how difficult it would be.
| |
| 21:41 | <gbaman> They arent stunning, I am a python guy
| |
| 21:41 | which isnt going to help me much at that level :D
| |
| 21:41 | <vagrantc> i would think of it as trivial.
| |
| 21:41 | * vagrantc wouldn't recommending including the python interpreter in the initramfs | |
| 21:42 | <ball> I should probably learn Python at some point, or one of the other trendy new languages.
| |
| 21:42 | * vagrantc 's typing is awful lately | |
| 21:42 | <gbaman> ha ha vagrantc
| |
| 21:42 | now that would be interesting...
| |
| 21:43 | if you were implementing it vagrantc, how long would it take you?
| |
| 21:44 | <vagrantc> 5-15 minutes
| |
| 21:44 | mostly waiting for the initramfs to build
| |
| 21:44 | <gbaman> ...
| |
| 21:44 | if I asked nicely, might you?
| |
| 21:44 | is a rather key thing for this teacher, some way the kids can take the pis home
| |
| 21:45 | <vagrantc> so, the whole point of LTSP was to not have an OS install on every machine, right?
| |
| 21:46 | <gbaman> the sd card image is stored locally so they can take it home if need be
| |
| 21:46 | <vagrantc> at which point, whhy not just give them each an SD card and have them take it home, plug it in at school, etc. ?
| |
| 21:46 | <gbaman> having them boot in the classroom though off the server means the teacher can be certain that the correct packages are installed when they are teaching
| |
| 21:47 | <vagrantc> how can you be sure they're not booted to the SD OS?
| |
| 21:47 | <gbaman> because it prompts them with a login screen?
| |
| 21:48 | <alkisg> ...write an initramfs script that tries netbooting, and if that fails, boots locally
| |
| 21:48 | <gbaman> are very different login procedures
| |
| 21:48 | <vagrantc> hmmm... it'd be a little more code, but something that booted locally if the network didn't tell it to boot to the network would be slick.
| |
| 21:48 | <alkisg> No need for keypresses etc
| |
| 21:48 | * vagrantc and alkisg think alike | |
| 21:48 | <vagrantc> now and then
| |
| 21:48 | <alkisg> The server nbd image can be cached to the sd to make sure it's consistent
| |
| 21:48 | <gbaman> only slightly creepy...
| |
| 21:48 | although impressive...
| |
| 21:49 | <alkisg> and easy to maintain (as opposed to local SD installations)
| |
| 21:49 | <gbaman> well that would be the perfect situation
| |
| 21:49 | is that even possible?
| |
| 21:49 | and it is currently nfs
| |
| 21:49 | <alkisg> Sure, it's just code
| |
| 21:49 | <vagrantc> anything not impossible is possible... it's just a question of how much work...
| |
| 21:50 | <gbaman> thats what I was thinking
| |
| 21:51 | gota admit that would be rather slick if that worked
| |
| 21:51 | <vagrantc> alkisg has been talking about such setups for a while, in one form or another implemented?
| |
| 21:52 | <gbaman> I doubt it would be an issue if the teacher was able to run a command every month or 2 that copied the image over to all sd cards?
| |
| 21:52 | dsugar100 has left IRC (dsugar100!~dsugar@columbia.tresys.com, Quit: dsugar100) | |
| 21:52 | <gbaman> wait, what happens with logins?
| |
| 21:52 | and files
| |
| 21:52 | <ball> There's a distinction between what's possible and what's practical.
| |
| 21:52 | * vagrantc nods | |
| 21:52 | <gbaman> that would normally be stored on the server
| |
| 21:52 | so what happens if a kid changes a file..
| |
| 21:53 | <vagrantc> if the next 12 generations of humanity all collectively haul on this...
| |
| 21:53 | gbaman: you could exlude some dirs, or make them local only, whatever...
| |
| 21:53 | <alkisg> The way I've been thinking that is, if the ltsp server is down (or the ltsp client is off-network), a local account would autologin
| |
| 21:53 | khildin_ has joined IRC (khildin_!~khildin@ip-213-49-84-161.dsl.scarlet.be) | |
| 21:53 | <alkisg> The local /home would be accessible to all users when online
| |
| 21:53 | <gbaman> like the sound of that alkisg
| |
| 21:54 | oooo
| |
| 21:54 | <alkisg> So they could transfer things
| |
| 21:54 | <gbaman> genius...
| |
| 21:55 | * alkisg thinks "LTSP" is not the correct initials anymore for these kinds of setups... :) | |
| 21:55 | <gbaman> am guessing though something like that is nowhere near existing..
| |
| 21:55 | <alkisg> Just a few hours of programming away...
| |
| 21:56 | khildin has left IRC (khildin!~khildin@ip-213-49-116-239.dsl.scarlet.be, Ping timeout: 264 seconds) | |
| 21:56 | * vagrantc has been pondering using freedomsponsors.org for some LTSP bounties | |
| 21:57 | <vagrantc> but maybe that would ruin everything
| |
| 21:57 | <alkisg> I think it would help
| |
| 21:57 | <gbaman> the probability though alkisg of those hours of programming being available any time soon?
| |
| 21:58 | <alkisg> gbaman: it depends... when are you available? :P
| |
| 21:58 | <gbaman> :P
| |
| 21:58 | that would require me to know stuff...
| |
| 21:58 | <alkisg> I'm sure you can learn
| |
| 21:58 | A few years ago I didn't know what ltsp is
| |
| 21:59 | freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun) | |
| 22:00 | * ball is sadly distracted from a conversation that look interesting. | |
| 22:01 | <vagrantc> !arcfour
| |
| 22:01 | <ltsp`> arcfour: http://en.wikipedia.org/wiki/RC4 is an SSH cipher which is more than 2 times faster than the default aes128-ctr. To enable it, set LDM_SSHOPTIONS="-o Ciphers=arcfour128".
| |
| 22:03 | <gbaman> vagrantc: ?
| |
| 22:03 | <vagrantc> gbaman: ??
| |
| 22:06 | <gbaman> hmm
| |
| 22:07 | so, what was the conclusion, best way to sort this issue of needing to switch between local and network OS?
| |
| 22:08 | <alkisg> Without programming anything? To install the local OS and have different boot entries for local or network booting in the bootloader menu :)
| |
| 22:08 | <gbaman> we have a bootloader menu?
| |
| 22:08 | <alkisg> Bootloaders usually support menus, I've no idea about pi
| |
| 22:08 | <gbaman> no bootloader menu on pi
| |
| 22:09 | <alkisg> Why not add one?
| |
| 22:10 | <gbaman> because... I don't have one...
| |
| 22:11 | * vagrantc hrmpfs | |
| 22:14 | <gbaman> well need some way to do it..
| |
| 22:16 | <vagrantc> simplest thing would be a simple: cp -vb cmdline_ltsp.txt cmdlinux.txt
| |
| 22:16 | <gbaman> and how would that be executed?
| |
| 22:17 | mainly as need to consider the issue that the pupil has taken hom the pi
| |
| 22:17 | <vagrantc> put a menu in the initramfs, or autodetect the network setting.
| |
| 22:17 | <gbaman> so does not have access to the server
| |
| 22:18 | <vagrantc> and reboot if the menu selects something other than what's already chosen.
| |
| 22:18 | echo Switch boot method?
| |
| 22:19 | <gbaman> ok, so initranfs contains /bin/bash?
| |
| 22:19 | <vagrantc> not bash
| |
| 22:19 | but /bin/sh
| |
| 22:19 | <gbaman> ok
| |
| 22:19 | <ball> What do they want the students to do with the Pi when they have it at home?
| |
| 22:19 | <gbaman> switch 2 config files
| |
| 22:19 | <ball> (I missed much of the above conversation)
| |
| 22:20 | <gbaman> so it switches to a local os
| |
| 22:20 | <vagrantc> gbaman: you probably only need to change cmdline.txt ... i bet it would work fine with one kernel for both environments.
| |
| 22:20 | <gbaman> hmm
| |
| 22:20 | that would make life easier...
| |
| 22:20 | as we dont need to initramfs's
| |
| 22:20 | *two
| |
| 22:21 | <vagrantc> so, experiment a bit, try it out, until you get it
| |
| 22:21 | <ball> Why not have a separate SD card for home use and one for when it's in the lab (being used as a terminal)?
| |
| 22:21 | <vagrantc> gbaman: come back with more specific questions
| |
| 22:21 | <gbaman> because sd cards cost money
| |
| 22:21 | :)
| |
| 22:21 | <ball> gbaman: US$5?
| |
| 22:21 | <vagrantc> gbaman: while your and our time doesn't :P
| |
| 22:21 | <gbaman> for 30 kids?
| |
| 22:22 | <ball> Have the kids buy their own, if they want to use the Raspberry Pi at home.
| |
| 22:22 | <gbaman> why waste the other 4gb on the sd cards
| |
| 22:22 | anyways, am off to bed, will try that tomorrow vagrantc, thanks :)
| |
| 22:22 | <ball> Why get 4G cards for the lab?
| |
| 22:22 | <vagrantc> why waste so much time developing a sub-optimal solution?
| |
| 22:23 | <gbaman> :)
| |
| 22:23 | gbaman has left IRC (gbaman!~gbaman@81.134.121.245, ) | |
| 22:23 | <ball> My head hurts like a big hurty thing.
| |
| 22:24 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection) | |
| 22:51 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 240 seconds) | |
| 23:01 | ball has left IRC (ball!ball@162-202-67-158.lightspeed.livnmi.sbcglobal.net, Quit: Bye!) | |
| 23:08 | monkwitdafunk has joined IRC (monkwitdafunk!~andchat@24.114.39.193) | |
| 23:53 | PhoenixSTF has joined IRC (PhoenixSTF!~rudiservo@78.29.155.245) | |
| 23:56 | BrotherOdd has joined IRC (BrotherOdd!~Brother@173-21-43-197.client.mchsi.com) | |