IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 29 November 2011 (all times are UTC)
| 01:08 | loather-work has left IRC (loather-work!~khudson@65-60-77-197.static-ip.telepacific.net, Quit: This computer has gone to sleep) | |
| 01:26 | adrianorg has left IRC (adrianorg!~adrianorg@189.58.228.178.dynamic.adsl.gvt.net.br, Ping timeout: 248 seconds) | |
| 01:41 | vagrantc has joined IRC (vagrantc!~vagrant@c-76-105-252-50.hsd1.or.comcast.net) | |
| 01:46 | * vagrantc finally got around to testing NFS with an aufs overlay | |
| 01:46 | <vagrantc> always on generation of technology behind :)
| |
| 01:47 | hughessd has left IRC (hughessd!~hughessd@173-164-117-109-Oregon.hfc.comcastbusiness.net, Quit: gone for some reason) | |
| 02:20 | Parker955_Away is now known as Parker955 | |
| 02:46 | monteslu__ is now known as monteslu | |
| 03:14 | NeonLicht has left IRC (NeonLicht!~NeonLicht@darwin.ugr.es, Ping timeout: 252 seconds) | |
| 03:15 | NeonLicht has joined IRC (NeonLicht!~NeonLicht@darwin.ugr.es) | |
| 04:00 | Parker955 is now known as Parker955_Away | |
| 04:17 | monteslu_ has joined IRC (monteslu_!~monteslu@ip68-109-174-213.ph.ph.cox.net) | |
| 04:20 | monteslu has left IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net, Ping timeout: 240 seconds) | |
| 04:32 | gentgeen__ has left IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net, Remote host closed the connection) | |
| 04:38 | alexqwesa__ has left IRC (alexqwesa__!~alex@alexo-veto.broker.freenet6.net, Ping timeout: 244 seconds) | |
| 04:41 | alexqwesa has joined IRC (alexqwesa!~alex@alexo-veto.broker.freenet6.net) | |
| 04:55 | gentgeen__ has joined IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net) | |
| 04:56 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 04:57 | <alkisg> vagrantc: did it work ok?
| |
| 05:03 | <vagrantc> alkisg: yes. didn't give it any stress testing ... but it basically seemed to work
| |
| 05:04 | <alkisg> How about the aufs => overlay switch?
| |
| 05:04 | *overlayfs
| |
| 05:04 | <vagrantc> no idea on that, overlayfs isn't in debian, as far as i can tell
| |
| 05:05 | <alkisg> I thought aufs was becoming unmaintained and overlayfs had some chances to go in the kernel tree... but not sure about the details there
| |
| 05:05 | <vagrantc> nor i
| |
| 05:06 | * vagrantc tries not to chase the latest and greatest | |
| 05:06 | <vagrantc> saves a lot of time :)
| |
| 05:06 | <alkisg> Sure... but I thought the aufs guy stopped updating it against the most recent kernels, that would require a fast switch
| |
| 05:07 | Well I suppose packagers can do that as well, if need be
| |
| 05:07 | gentgeen__ has left IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net, Remote host closed the connection) | |
| 05:15 | <cyberorg> aufs was dropped from suse few releases back as it was not maintained
| |
| 05:19 | <alkisg> cyberorg: and you're using overlayfs since then?
| |
| 05:19 | <cyberorg> alkisg, we use clicfs
| |
| 05:20 | it is fused based, ext3/ext4 compressed container, so no kernel involved
| |
| 05:20 | <alkisg> Hmmm I don't see it in debian/ubuntu, I wonder why...
| |
| 05:21 | vagrantc: btw about ltsp-chroot, we should also offer a switch to mount /dev/pts... and maybe we should mount /proc (at least) by default...
| |
| 05:21 | <cyberorg> http://gitorious.org/opensuse/clicfs
| |
| 05:24 | <vagrantc> knipwim: the changes you had committed to ltsp-chroot broke use of commands for me
| |
| 05:25 | alkisg: sure, could probably mount by default, but probably need to be careful that it's not already mounted
| |
| 05:25 | <alkisg> Indeed
| |
| 05:26 | * alkisg gets the kids ready for school, bbl, will read logs | |
| 05:26 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 05:44 | loather-work has joined IRC (loather-work!~khudson@wsip-216-231-4-2.oc.oc.cox.net) | |
| 05:50 | staffencasa has left IRC (staffencasa!~staffenca@128-193-144-154.oregonstate.edu, Ping timeout: 260 seconds) | |
| 06:11 | Yet_another_Bill has left IRC (Yet_another_Bill!billy@nat/redhat/x-qzkjedtprxdgzaob, Quit: Good night ^_^) | |
| 06:13 | Yet_another_Bill has joined IRC (Yet_another_Bill!billy@nat/redhat/x-dineerrlfxrmmvhx) | |
| 06:20 | mistik1 has left IRC (mistik1!mistik1@unaffiliated/mistik1, Ping timeout: 245 seconds) | |
| 06:21 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 06:22 | mistik1 has joined IRC (mistik1!~mistik1@unaffiliated/mistik1) | |
| 06:25 | alexqwesa has left IRC (alexqwesa!~alex@alexo-veto.broker.freenet6.net, Ping timeout: 244 seconds) | |
| 06:35 | <knipwim> vagrantc: that shouldn't have happened
| |
| 06:35 | can you be more specific, so i can fix it?
| |
| 06:36 | * vagrantc looks for the commit | |
| 06:36 | <vagrantc> -eval LTSP_HANDLE_DAEMONS=false chroot "$ROOT" $COMMAND
| |
| 06:36 | +LTSP_HANDLE_DAEMONS=false chroot "$ROOT" $COMMAND
| |
| 06:36 | in my setup, "ltsp-chroot apt-get update" doesn't work without the eval
| |
| 06:37 | <knipwim> check
| |
| 06:37 | i'll change it back to the old behaviour
| |
| 06:37 | tonight, if that's allright
| |
| 06:37 | <vagrantc> this is classical crazy shell argument processing
| |
| 06:38 | <alkisg> I agree that we should skip eval there though
| |
| 06:38 | And use "$@" instead of command
| |
| 06:38 | But, it needs to work of course ;)
| |
| 06:39 | * vagrantc found a handful of bugs by actually testing the code committed during and post-hackfest | |
| 06:39 | <alkisg> vagrantc: why doesn't it work without eval for you?
| |
| 06:39 | COMMAND="ls -ld /tmp"
| |
| 06:39 | LTSP_HANDLE_DAEMONS=false chroot /opt/ltsp/i386 $COMMAND
| |
| 06:39 | drwxrwxrwt 2 root root 4096 2011-11-19 00:56 /tmp
| |
| 06:40 | <vagrantc> i can't recall the exact command i used... but i think there were some quoting issues.
| |
| 06:41 | <alkisg> That's because we use $COMMAND instead of "$@"
| |
| 06:41 | Could you test with "$@"? It should be the corrent (TM) way to do this...
| |
| 06:41 | <vagrantc> i was surprised that it didn't workk
| |
| 06:42 | can't test tonight, really
| |
| 06:43 | <alkisg> No hurry at all; I was just suggesting that someone of us should test "$@" before knipwim puts back the eval again...
| |
| 06:43 | <vagrantc> i pushed the commits i thought were ok to push
| |
| 06:44 | a 17 hour train ride with distributed VCS and a properly populated approx cache for packages ... i got a lot of intensive work done!
| |
| 06:44 | or at least, a lot of testing
| |
| 06:44 | <alkisg> Ouch, 17 hours in a train? :(
| |
| 06:44 | <vagrantc> heh
| |
| 06:45 | anything less than 48 hours is easy :)
| |
| 06:45 | <alkisg> Haha /me has done that, no fun
| |
| 06:49 | Btw, for shift to properly work (i.e. to have the options remove and the actual command line parameters to remain), I think it should be done in "main", not in a function like analyze_command_line
| |
| 06:52 | ...but of course getopt gets in the way of using shift... meh
| |
| 06:58 | * alkisg thinks getopt needs to be paired with eval, so we either need to use both of them, or none of them | |
| 06:59 | <alkisg> knipwim: that last line ^
| |
| 07:00 | <knipwim> i see
| |
| 07:01 | * vagrantc might bump to ltsp 5.3.0 and upload next weekend | |
| 07:02 | <vagrantc> though basically, a backported server package needs to use backported client packages with the initramfs-scripts.d changes
| |
| 07:03 | <alkisg> man getopt => To solve this problem, this implementation can generate quoted output which must once again be interpreted by the shell (usually by using the eval command). This has the effect of preserving those characters, but you must call getopt in a way that is no longer compatible with other versions
| |
| 07:03 | <vagrantc> wheee.
| |
| 07:03 | this is one of the achiles heels of shell
| |
| 07:04 | <alkisg> Yeah... and `shift` only applies to a handful of cases
| |
| 07:06 | <knipwim> so the COMMAND is defined of what remaines of $@ after all shifts
| |
| 07:07 | <alkisg> knipwim: yes, but it's specially quoted by getopt
| |
| 07:07 | So e.g. if one runs ltsp-chroot ls "/tmp/some dir with spaces",
| |
| 07:07 | COMMAND will contain \'ls\' \'/tmp/some dir with spaces\' or something similar
| |
| 07:08 | So it's not directly executable, unless you have the old getopt that doesn't quote by default
| |
| 07:08 | <knipwim> that's why the eval is needed, check
| |
| 07:09 | <alkisg> We could skip both eval and getopt; in fact I think it'd be cleaner this way; but whatever no big deal both ways should work ok with spaces and anything
| |
| 07:52 | ry has left IRC (ry!~ry@static-71-183-64-28.nycmny.fios.verizon.net, Ping timeout: 258 seconds) | |
| 07:54 | vagrantc has left IRC (vagrantc!~vagrant@c-76-105-252-50.hsd1.or.comcast.net, Read error: Operation timed out) | |
| 08:05 | ry has joined IRC (ry!~ry@static-71-183-64-28.nycmny.fios.verizon.net) | |
| 08:15 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 08:33 | loather-work has left IRC (loather-work!~khudson@wsip-216-231-4-2.oc.oc.cox.net, Quit: This computer has gone to sleep) | |
| 09:00 | fosser_josh has joined IRC (fosser_josh!~prathames@02799a2d.bb.sky.com) | |
| 09:01 | fosser_josh has left IRC (fosser_josh!~prathames@02799a2d.bb.sky.com) | |
| 09:12 | shawnp0wers has left IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers, Ping timeout: 252 seconds) | |
| 09:13 | khildin has joined IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 09:14 | shawnp0wers has joined IRC (shawnp0wers!~spowers@71-13-74-18.static.aldl.mi.charter.com) | |
| 09:14 | shawnp0wers has joined IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers) | |
| 09:19 | Steve_the_Pirate has joined IRC (Steve_the_Pirate!~Gary@host109-155-255-130.range109-155.btcentralplus.com) | |
| 09:30 | <chrisjrob> just had a user complain that they had lost their panel (KDE Kicker), I felt their ought to be a way that I could run it for them, but just couldn't think how. In the end I used dcop to log them out and they were fine, but I felt there must have been a more elegant solution. Any thoughts?
| |
| 09:30 | s/their/there/
| |
| 09:31 | dobber has joined IRC (dobber!~dobber@213.169.45.222) | |
| 10:03 | <Hyperbyte> !xauthority
| |
| 10:03 | <ltsp> Hyperbyte: xauthority: To access the thin client X display from ssh or from a local console, try: export $(tr '\0' '\n' < /proc/$(pgrep -nx 'ldm|kdm_greet|lxdm-greeter-gt|dbus-daemon')/environ | egrep '^DISPLAY=|^XAUTHORITY='
| |
| 10:10 | <chrisjrob> Hyperbyte: I can do that fine, but how do I run a server app?
| |
| 10:11 | in this case "kicker"
| |
| 10:11 | oh
| |
| 10:12 | i assumed that when i ran apps from the client using xauthority that they were local apps
| |
| 10:12 | they must be surely
| |
| 10:17 | phew, I am right - was doubting my sanity their for a moment
| |
| 10:17 | garggh s/their/there/
| |
| 10:18 | so is there a way of running a server application, e.g. kicker?
| |
| 10:21 | <Hyperbyte> I just gave you it. You can 'su' to the user, and then run apps... but you won't be able to run graphical apps directly on their display, if you don't have their xauthority.
| |
| 10:22 | And X runs on the thin clients. Server apps also need to access X on the thin client, not just localapps.
| |
| 10:49 | Da-Geek has joined IRC (Da-Geek!Da-Geek@nat/redhat/x-nwhohiuabzwbxnbg) | |
| 10:50 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 10:52 | <alkisg> chrisjrob: setting xauthority doesn't make the program part of the user session, more things are needed, like dbus address etc
| |
| 10:53 | To correctly launch programs inside a user's session, you need some "listener" inside that session
| |
| 10:53 | <chrisjrob> ah
| |
| 10:53 | <alkisg> Epoptes offers that, if you want:
| |
| 10:53 | !epoptes
| |
| 10:53 | <ltsp> alkisg: epoptes: Epoptes is a computer lab administration and monitoring tool. It works on Ubuntu and Debian based labs with LTSP or non-LTSP servers, thin and fat clients, standalone workstations, NX clients etc. More info: http://www.epoptes.org
| |
| 10:54 | <chrisjrob> right, okay that makes sense
| |
| 10:54 | thanks Hyperbyte alkisg for your help
| |
| 10:55 | <Hyperbyte> Hey alkisg! :)
| |
| 10:55 | Good morning.
| |
| 10:55 | (just)
| |
| 10:55 | <chrisjrob> hmm 10.55 here :)
| |
| 10:55 | <alkisg> Good morning :)
| |
| 10:55 | (or afternoon anyway, 12:55 ;))
| |
| 10:56 | <chrisjrob> i see we have still not all embraced UTC
| |
| 11:35 | eicto has left IRC (eicto!~morgoth@eicto.broker.freenet6.net, Remote host closed the connection) | |
| 11:52 | Trixboxer has joined IRC (Trixboxer!~Trixboxer@office.supportdepartment.net) | |
| 11:57 | leio has left IRC (leio!~leio@gentoo/developer/leio, Ping timeout: 244 seconds) | |
| 11:58 | leio has joined IRC (leio!~leio@gentoo/developer/leio) | |
| 12:22 | khildin has left IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be, Ping timeout: 258 seconds) | |
| 12:31 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 12:40 | Parker955_Away is now known as Parker955 | |
| 12:59 | Parker955 is now known as Parker955_Away | |
| 13:05 | shawnp0wers has left IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers, Ping timeout: 244 seconds) | |
| 13:07 | shawnp0wers has joined IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers) | |
| 13:12 | gentgeen__ has joined IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net) | |
| 13:21 | <mgariepy> good morning everyone
| |
| 13:35 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.15.11) | |
| 13:36 | khildin has joined IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 13:38 | khildin has left IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be, Remote host closed the connection) | |
| 13:41 | khildin has joined IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 13:41 | corewin has joined IRC (corewin!~IceChat77@168.12.8.50) | |
| 13:41 | <corewin> Hi Everyone
| |
| 13:45 | corewin has left IRC (corewin!~IceChat77@168.12.8.50) | |
| 13:57 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 14:05 | <muppis> Is there existing lsb script for nbd server?
| |
| 14:06 | Or upstart, no matter.
| |
| 14:07 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 14:07 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 14:09 | <_UsUrPeR_> hello all
| |
| 14:12 | <alkisg> Hi _UsUrPeR_. muppis, yes, you can see a package contents with dpkg -L, e.g.: dpkg -L nbd-server
| |
| 14:12 | * _UsUrPeR_ tips his hat | |
| 14:15 | Steve_the_Pirate has left IRC (Steve_the_Pirate!~Gary@host109-155-255-130.range109-155.btcentralplus.com, Quit: Leaving) | |
| 14:15 | <muppis> alkisg, just 'yes' can sometimes be proper answer. :)
| |
| 14:20 | <_UsUrPeR_> Bah. Without an example, he would have been one of "those" guys ;)
| |
| 14:21 | khildin has left IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be, Read error: Connection timed out) | |
| 14:21 | khildin_ has joined IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 14:30 | khildin_ has left IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be, *.net *.split) | |
| 14:30 | shawnp0wers has left IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers, *.net *.split) | |
| 14:30 | Mava has left IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi, *.net *.split) | |
| 14:36 | khildin_ has joined IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 14:36 | shawnp0wers has joined IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers) | |
| 14:36 | Mava has joined IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi) | |
| 14:52 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 14:54 | Gadi has joined IRC (Gadi!~romm@ool-4571ca04.dyn.optonline.net) | |
| 15:01 | khildin_ has left IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be, Ping timeout: 252 seconds) | |
| 15:03 | dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174) | |
| 15:12 | adrianorg has joined IRC (adrianorg!~adrianorg@189.58.228.178.dynamic.adsl.gvt.net.br) | |
| 15:13 | khildin_ has joined IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 15:29 | can-o-spam1 has joined IRC (can-o-spam1!b85be904@gateway/web/freenode/ip.184.91.233.4) | |
| 15:29 | khildin_ has left IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be, Quit: I'm gone, bye bye) | |
| 15:30 | khildin_ has joined IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 15:32 | alkisg1 has joined IRC (alkisg1!~alkisg@ubuntu/member/alkisg) | |
| 15:32 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 258 seconds) | |
| 15:36 | pimpministerp has joined IRC (pimpministerp!~ppouliot@nat/novell/x-ixfxihdahzstyvxf) | |
| 15:36 | alkisg1 is now known as alkisg | |
| 15:43 | adrianorg has left IRC (adrianorg!~adrianorg@189.58.228.178.dynamic.adsl.gvt.net.br, Ping timeout: 244 seconds) | |
| 15:49 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Read error: Connection reset by peer) | |
| 15:50 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 15:54 | ogra_ has left IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de, Excess Flood) | |
| 16:00 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 16:00 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 16:02 | darkpixel_ has joined IRC (darkpixel_!~darkpixel@65.100.44.217) | |
| 16:02 | darkpixel_ has left IRC (darkpixel_!~darkpixel@65.100.44.217, Changing host) | |
| 16:02 | darkpixel_ has joined IRC (darkpixel_!~darkpixel@curetheitch/staff/darkpixel) | |
| 16:02 | steelrat405 has left IRC (steelrat405!steelrat40@COX-66-210-175-107-static.coxinet.net, ) | |
| 16:05 | <knipwim> alkisg: with the eval the "ltsp-chroot ls" works
| |
| 16:06 | <alkisg> knipwim: nice, yeah I think that should be committed
| |
| 16:09 | <knipwim> and done
| |
| 16:09 | staffencasa has joined IRC (staffencasa!~staffenca@128-193-144-154.oregonstate.edu) | |
| 16:20 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Read error: Operation timed out) | |
| 16:20 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Read error: Connection reset by peer) | |
| 16:21 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 16:21 | hughessd has joined IRC (hughessd!~hughessd@173-164-117-109-Oregon.hfc.comcastbusiness.net) | |
| 16:23 | andygraybeals has left IRC (andygraybeals!~andy@h28.215.22.98.dynamic.ip.windstream.net, Quit: Ex-Chat) | |
| 16:26 | ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de) | |
| 16:27 | Steve_the_Pirate has joined IRC (Steve_the_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com) | |
| 16:44 | dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection) | |
| 16:45 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 16:48 | jammcq has joined IRC (jammcq!~jam@70-91-230-209-BusName-Michigan.hfc.comcastbusiness.net) | |
| 16:54 | novacaine_ has joined IRC (novacaine_!~quassel@ppp-24-149.29-151.libero.it) | |
| 16:56 | vagrantc has joined IRC (vagrantc!~vagrant@c-76-105-252-50.hsd1.or.comcast.net) | |
| 16:56 | alkisg has joined IRC (alkisg!4f672b79@gateway/web/freenode/ip.79.103.43.121) | |
| 16:59 | toscalix has joined IRC (toscalix!~toscalix@178.139.100.195) | |
| 17:01 | Da-Geek has left IRC (Da-Geek!Da-Geek@nat/redhat/x-nwhohiuabzwbxnbg, Quit: Leaving) | |
| 17:05 | novacaine_ has left IRC (novacaine_!~quassel@ppp-24-149.29-151.libero.it, Remote host closed the connection) | |
| 17:07 | alkisg has left IRC (alkisg!4f672b79@gateway/web/freenode/ip.79.103.43.121, Ping timeout: 265 seconds) | |
| 17:12 | loather-work has joined IRC (loather-work!~khudson@65-60-77-197.static-ip.telepacific.net) | |
| 17:18 | abeehc has joined IRC (abeehc!~bob@S010652540047e22e.gv.shawcable.net) | |
| 17:22 | priyank has joined IRC (priyank!~priyank@122.177.206.152) | |
| 17:28 | vagrantc has left IRC (vagrantc!~vagrant@c-76-105-252-50.hsd1.or.comcast.net, Ping timeout: 260 seconds) | |
| 17:41 | Steve_the_Pirate has left IRC (Steve_the_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com, Quit: Leaving) | |
| 17:48 | <priyank> hi when I am trying to boot my thin client and I give username and password it gives error (response from server, restarting...)
| |
| 17:48 | Please tell what am i doing wrong
| |
| 17:52 | priyank has left IRC (priyank!~priyank@122.177.206.152, Quit: Leaving) | |
| 18:16 | Mava has left IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi, Ping timeout: 252 seconds) | |
| 18:18 | loather has joined IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net) | |
| 18:24 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 18:30 | alkisg1 has joined IRC (alkisg1!~alkisg@193.92.236.50.dsl.dyn.forthnet.gr) | |
| 18:30 | alkisg1 has left IRC (alkisg1!~alkisg@193.92.236.50.dsl.dyn.forthnet.gr, Changing host) | |
| 18:30 | alkisg1 has joined IRC (alkisg1!~alkisg@ubuntu/member/alkisg) | |
| 18:30 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Read error: Connection reset by peer) | |
| 18:31 | alkisg1 is now known as alkisg | |
| 18:31 | toscalix has left IRC (toscalix!~toscalix@178.139.100.195, Remote host closed the connection) | |
| 18:31 | toscalix has joined IRC (toscalix!~toscalix@178.139.100.195) | |
| 18:40 | toscalix has left IRC (toscalix!~toscalix@178.139.100.195, Remote host closed the connection) | |
| 18:52 | loather has left IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net, Quit: This computer has gone to sleep) | |
| 19:03 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 19:13 | <hughessd> have you done an update recently?
| |
| 19:13 | did you remember to update the ssh keys?
| |
| 19:14 | oops.... priyank: have you done an update recently? did you remember to update the ssh keys?
| |
| 19:15 | also, does anyone know if restarting the ssh daemon will kill the current sessions?
| |
| 19:19 | vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net) | |
| 19:21 | <hughessd> well, tried it...guess not
| |
| 19:33 | jammcq has left IRC (jammcq!~jam@70-91-230-209-BusName-Michigan.hfc.comcastbusiness.net, Quit: leaving) | |
| 19:46 | Mava has joined IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi) | |
| 19:54 | corewin has joined IRC (corewin!~IceChat77@168.12.8.50) | |
| 19:56 | BonaC has joined IRC (BonaC!~bonac@t143.ip7.netikka.fi) | |
| 19:56 | <BonaC> Hi!
| |
| 19:56 | <corewin> Hi all. Anyone know what the "Invalid card number." message means? Eveything seems to work fine.
| |
| 19:56 | <BonaC> No idea about that and I also have a problem. :D
| |
| 19:57 | I have now been trying to install LTSP on terminal Debian and now after 6 and ½ hours
| |
| 19:57 | <Gadi> corewin: it is most likely alsa complaining about no sound card
| |
| 19:57 | <BonaC> I managed to get the PXE to work, but it says something about nfs
| |
| 19:57 | <corewin> Gadi: hmm the sound is working on the thin client
| |
| 19:58 | BonaC: is it saying can't connect to nbd server?
| |
| 19:58 | <Gadi> corewin: when we adjust volume controls, we specify card 0. Could be there is no card 0
| |
| 19:58 | <BonaC> Hold on
| |
| 19:58 | I'll paste it
| |
| 19:58 | nfsmount: need a path
| |
| 19:58 | <Gadi> BonaC: do you have root-path defined in dhcpd.conf?
| |
| 19:58 | <BonaC> Hmm
| |
| 19:59 | Nope
| |
| 19:59 | <Gadi> BonaC: that could be it
| |
| 20:00 | <BonaC> Well... I did get the DHCP manually put together after four hours
| |
| 20:00 | <Gadi> BonaC: does debian not supply an /etc/ltsp/dhcpd.conf?
| |
| 20:00 | <BonaC> Yes it does
| |
| 20:00 | But there was a slight problem
| |
| 20:00 | /etc/default/isc-dhcp-server had a argument INTERFACES=""
| |
| 20:01 | I first changed it to "eth0
| |
| 20:01 | "
| |
| 20:01 | loather-work has left IRC (loather-work!~khudson@65-60-77-197.static-ip.telepacific.net, Quit: This computer has gone to sleep) | |
| 20:01 | <BonaC> After three hours I came back to it and put it " eth0 " and suddenly it started working
| |
| 20:01 | You can feel the frustration. D
| |
| 20:01 | But now, the root folder to the .conf
| |
| 20:03 | How the command went
| |
| 20:03 | Just root-folder "x"
| |
| 20:03 | ?
| |
| 20:04 | <corewin> Gadi: Do i check to see if i have a card 0 in /proc/asound/modules ?
| |
| 20:05 | <Gadi> BonaC: option root-path "/opt/ltsp/i386";
| |
| 20:05 | <BonaC> Thanks
| |
| 20:05 | <Gadi> corewin: yes - on the client
| |
| 20:08 | <BonaC> Derp
| |
| 20:08 | It didn't work
| |
| 20:08 | Still nagging
| |
| 20:08 | <Gadi> did you restart the dhcp server?
| |
| 20:08 | <BonaC> Yes
| |
| 20:09 | <Gadi> try: option root-path "<ip-of-server>:/opt/ltsp/i386";
| |
| 20:09 | <BonaC> Hold on, I forgot to sudo it
| |
| 20:10 | <corewin> Gadi: both the client and ltsp server have 0 snd_hda_inetl in the modules file
| |
| 20:10 | <Gadi> corewin: then, perhaps it is a race condition. ie, perhaps the module is not loaded when we try to adjust the sound
| |
| 20:13 | <BonaC> adi: http://bonac.dy.fi/files/DHCP.txt
| |
| 20:13 | Fuck
| |
| 20:13 | I forgot the password requirement
| |
| 20:13 | ...
| |
| 20:14 | Gadi: http://bonac.dy.fi/DHCP.txt
| |
| 20:14 | Now it should work
| |
| 20:14 | Can you see any problems in the DHCP config?
| |
| 20:15 | <Gadi> you are missing a colon
| |
| 20:15 | <BonaC> Where
| |
| 20:15 | <Gadi> between ip and path
| |
| 20:15 | <BonaC> Oh
| |
| 20:15 | ...
| |
| 20:15 | <Gadi> also, I would put the root-path in the subnet declaration
| |
| 20:15 | <BonaC> Ok, let's try that
| |
| 20:16 | can-o-spam1 has left IRC (can-o-spam1!b85be904@gateway/web/freenode/ip.184.91.233.4, Ping timeout: 265 seconds) | |
| 20:17 | <BonaC> Now it says that the file not found
| |
| 20:18 | Okay, so it's not that
| |
| 20:20 | corewin has left IRC (corewin!~IceChat77@168.12.8.50) | |
| 20:21 | <BonaC> What the fuck
| |
| 20:21 | Now it's not working at all
| |
| 20:21 | ...
| |
| 20:22 | <vagrantc> also, you have to tell /etc/isc-dhcp-server/dhcpd.conf to point to /etc/ltsp/dhcpd.conf
| |
| 20:23 | <Gadi> also, I doubt filename "[xelinux.0" is correct
| |
| 20:23 | *pxelinux.0
| |
| 20:23 | it is usually "/ltsp/i386/pxelinux.0"
| |
| 20:24 | <BonaC> ...
| |
| 20:24 | So the filename-command doesn't depend on the root-folder at all?
| |
| 20:25 | Shouldn't it be /ltsp/i386/boot/pxelinux.0
| |
| 20:25 | I mean the pxelinux.0 is in /boot/
| |
| 20:27 | <Gadi> no - filename has to do with the TFTP server
| |
| 20:27 | <BonaC> Ok
| |
| 20:27 | <Gadi> your TFTP server most likely has the root of: /var/lib/tftpboot/
| |
| 20:27 | or /srv/tftpboot/
| |
| 20:27 | <BonaC> the later one
| |
| 20:28 | <Gadi> and "filename" is relative to that
| |
| 20:28 | (because the tftp is securely chrooted)
| |
| 20:29 | when you run ltsp-update-kernels (which is done automatically when you first build the client), it copies the pxelinux stuff to the tftp directory
| |
| 20:29 | <BonaC> Ok
| |
| 20:29 | I'm gonna fuck this laptop with a chainsaw
| |
| 20:29 | This isn't going nowhere
| |
| 20:29 | host ws001 { hardware ethernet 00:C0:9F:75:78:BD; fixed-address 192.168.1.11; filename "pxelinux.0"; option root-path "/opt/ltsp/i386/"; }
| |
| 20:29 | <Gadi> I'm sure there's a channel for that
| |
| 20:29 | :)
| |
| 20:30 | <BonaC> Yes..
| |
| 20:30 | Now, there's something really wrong
| |
| 20:30 | I mean when I came on this channel
| |
| 20:30 | <Gadi> right
| |
| 20:30 | look:
| |
| 20:30 | filename "/ltsp/i386/pxelinux.0";
| |
| 20:30 | <BonaC> The thing nagged about nfs
| |
| 20:30 | DEEEEERP
| |
| 20:30 | <Gadi> option root-path "/opt/ltsp/i386";
| |
| 20:31 | vagrantc: perhaps you have a sample dhdcpd.conf for debian handy?
| |
| 20:31 | *dhcpd.conf
| |
| 20:32 | <BonaC> Okay
| |
| 20:32 | Now the
| |
| 20:32 | Are going somewhere
| |
| 20:32 | FUCK
| |
| 20:32 | IT BOOTS
| |
| 20:32 | Well...
| |
| 20:32 | Something
| |
| 20:32 | Now I see loading cursot
| |
| 20:33 | "LTSP.org - Linux Terminal Server Project"
| |
| 20:33 | And a username box
| |
| 20:33 | Oh man I love this thing now <3
| |
| 20:33 | Have to test this out
| |
| 20:33 | <Gadi> lol
| |
| 20:34 | <BonaC> Okayt
| |
| 20:34 | It seems I lack users
| |
| 20:34 | How do I add them
| |
| 20:34 | ....
| |
| 20:34 | :D
| |
| 20:35 | "Man, I'm a damn idiot"
| |
| 20:35 | <vagrantc> Gadi: the one that ships in /etc/ltsp/dhcpd.conf works for me ... just need to add an include statement to the bottom of /etc/*dhcp*/dhcpd.conf (depending on what version you're running)
| |
| 20:36 | <BonaC> Well... See, I accidentally deleted the whole damn thing by accident
| |
| 20:36 | When I tried to find the problem with the DHCP
| |
| 20:36 | :D
| |
| 20:36 | <vagrantc> it should also ship in /usr/share/doc/ltsp-server/examples
| |
| 20:37 | <BonaC> Nope, didn't find it
| |
| 20:37 | But
| |
| 20:37 | <vagrantc> do you have the ltsp-server package installed?
| |
| 20:37 | <BonaC> Now I need user and a password
| |
| 20:37 | Yes
| |
| 20:37 | * vagrantc is suspicious | |
| 20:37 | <Gadi> BonaC: from the client, you are logging into the server. So, simply add users to the server
| |
| 20:38 | <BonaC> To save time, you could also tell me where to put the users. :)
| |
| 20:38 | <Gadi> BonaC: as long as a user can ssh into the server, it can log in from the thin client
| |
| 20:38 | <BonaC> Oh
| |
| 20:38 | Well
| |
| 20:38 | let's see
| |
| 20:39 | Eep
| |
| 20:39 | I got scared
| |
| 20:39 | I thought that the audio wasn't working
| |
| 20:39 | I was hitting backspace like a madman
| |
| 20:39 | Then suddenly the loudest beep this thing can produce
| |
| 20:39 | D:
| |
| 20:40 | <vagrantc> Gadi: i tested the new-wave tweak everything from initramfs and after a few patches, it basically works
| |
| 20:40 | matrix3000 has joined IRC (matrix3000!~Eric@rrcs-70-61-255-227.central.biz.rr.com) | |
| 20:40 | <vagrantc> Gadi: pretty much requires writeable /
| |
| 20:40 | Gadi: though i had good luck with NFS with aufs
| |
| 20:41 | <Gadi> vagrantc: great!
| |
| 20:41 | * Gadi guesses he should look at the commits | |
| 20:41 | <vagrantc> Gadi: unfortunately it means a newer server can't install an older ltsp chroot without also installing newer ltsp backports in the chroot
| |
| 20:41 | since the older versions relied on tweaks to the chroot itself
| |
| 20:42 | <Gadi> not sure I see what you mean
| |
| 20:42 | the chroot should contain everything - even the initramfs
| |
| 20:43 | wouldn't a new server with an old chroot simply get an old initramfs
| |
| 20:43 | <vagrantc> Gadi: i used to be able to install a debian lenny chroot from squeeze without any real problems
| |
| 20:43 | Gadi: without requiring the backported packages in the chroot
| |
| 20:43 | <Gadi> oh! you're saying if we modify l-b-c plugins
| |
| 20:43 | <vagrantc> Gadi: but now, there will be a generation gap where that won't work
| |
| 20:44 | Gadi: by dropping support in the plugins, we loose the ability to do that ... and the ability to detect that
| |
| 20:44 | <Gadi> I think we are keeping plugins
| |
| 20:44 | and l-b-c
| |
| 20:44 | for when you want to build that way
| |
| 20:44 | <vagrantc> so we're tweaking things both in the l-b-c plugins *and* in the initramfs?
| |
| 20:44 | <Gadi> well, we can make those existing plugin tweaks smarter
| |
| 20:45 | <vagrantc> sure
| |
| 20:45 | <Gadi> there are only a handful
| |
| 20:45 | <vagrantc> oh, i probably never changed the plugins, so i probably ran with both
| |
| 20:45 | <Gadi> but, we're not getting rid of l-b-c entirely
| |
| 20:45 | <vagrantc> but at any rate, sometimes progress needs to move forward
| |
| 20:45 | <Gadi> sometimes
| |
| 20:45 | <vagrantc> overall, it's a better direction
| |
| 20:46 | <Gadi> :)
| |
| 20:46 | I am glad they worked!
| |
| 20:46 | did you try the udev-based sound volume?
| |
| 20:46 | <vagrantc> after a few patches :)
| |
| 20:46 | no, didn't try that
| |
| 20:46 | <Gadi> did you commit the patches?
| |
| 20:46 | <vagrantc> yeah, i pushed them, or at least the ones that shouldn't be disruptive
| |
| 20:47 | * vagrantc had a patch to maintain compatibility with aufs/overlayfs, but instead committed something better | |
| 20:47 | <vagrantc> maintain compatibility in ltsp_nbd ... but instead i committed it to init-bottom/ltsp :)
| |
| 20:48 | * vagrantc should disappear nowish | |
| 20:49 | <Gadi> ok ill take a look
| |
| 20:58 | certainly looks like we are close to being able to drop ltsp_nbd in favor of upstream
| |
| 20:58 | or at least we can make a much more slim version
| |
| 21:00 | loather-work has joined IRC (loather-work!~khudson@65-60-77-197.static-ip.telepacific.net) | |
| 21:03 | hughessd has left IRC (hughessd!~hughessd@173-164-117-109-Oregon.hfc.comcastbusiness.net, Remote host closed the connection) | |
| 21:04 | BonaC is now known as bonac | |
| 21:05 | bonac is now known as BonaC | |
| 21:06 | * Yoe likes hearing that :-) | |
| 21:11 | vagrantc has left IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net, Ping timeout: 255 seconds) | |
| 21:13 | Steve_the_Pirate has joined IRC (Steve_the_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com) | |
| 21:19 | * Gadi wonders if we could simply have another local-top/ script before upstream nbd's that does the addition magic and variable mangling we need .... | |
| 21:20 | Parker955_Away is now known as Parker955 | |
| 21:24 | <Yoe> Gadi: what do you need to do?
| |
| 21:25 | <Gadi> Yoe: well, I am not sure what the history is as to why we have our own lts_nbd rather than using upstream nbd...
| |
| 21:25 | but, looking at our code,
| |
| 21:25 | we have slightly different syntax for nbdroot,
| |
| 21:26 | we have nbd swap in there conditionally
| |
| 21:26 | we do all the mounting and overlayfs (rather than relying on local to do that)
| |
| 21:26 | <Yoe> Gadi: well, the history is that ltsp_nbd is older than my script :-)
| |
| 21:26 | <Gadi> we have nbd_proxy optionally
| |
| 21:27 | right, but we have a lot of junk thrown in there
| |
| 21:27 | that can be separated out
| |
| 21:27 | and in fact, we have been beginning to separate it out
| |
| 21:27 | <Yoe> IC
| |
| 21:27 | <Gadi> so, I am just looking to see if we can go all the way
| |
| 21:27 | :)
| |
| 21:27 | and I think we can
| |
| 21:27 | by cleverly mangling parameters beforehand that your script expects
| |
| 21:28 | <Yoe> I do remember talking to vagrantc in Banja Luka about all this, where he said that the nbdroot thing could just go
| |
| 21:28 | <Gadi> and then cleverly doing post-nbd-client stuff in init-bottom
| |
| 21:28 | <Yoe> but I think it might also make sense to change my script so it understands your bits without mangling stuff
| |
| 21:28 | <Gadi> well, for example,
| |
| 21:28 | if we have one convention for nbdroot
| |
| 21:29 | and you have another
| |
| 21:29 | we can simply read ours in first, parse it, and set nbdsrv and friends
| |
| 21:29 | so your script does the work
| |
| 21:29 | <Yoe> eh, hrm, that could work
| |
| 21:29 | <Gadi> OR, you can make your stuff more complicated by supporting 2 nbdroot conventions
| |
| 21:29 | :P
| |
| 21:29 | <Yoe> but I'd be wary of expecting those variables to be public interfaces
| |
| 21:30 | <Gadi> right
| |
| 21:30 | <Yoe> they're implementation details, not interfaces
| |
| 21:30 | <Gadi> well, OR
| |
| 21:30 | <Yoe> I don't think it hurts much to have an extra block of code that parses nbdroot in a different way... that way you won't also have it break by someone doing an NMU who doesn't know the code as well
| |
| 21:30 | <Gadi> we could simply change our nbdroot to your nbdroot
| |
| 21:30 | <Yoe> (or me doing an upload and forgetting all the details... which has also happened)
| |
| 21:31 | <Gadi> but, then, you would need to look for it in the env as opposed to just /proc/cmdline
| |
| 21:31 | <Yoe> yes, if possible, I suppose that would be the easiest and cleanest way forward
| |
| 21:31 | how's that?
| |
| 21:32 | <Gadi> well, right now, you set nbdroot by reading /proc/cmdline
| |
| 21:32 | if we wanted to intercept reading /rpoc/cmdline for nbdroot
| |
| 21:32 | <Yoe> oh
| |
| 21:32 | <Gadi> and then mangle nbdroot to conform to your convention
| |
| 21:32 | <Yoe> I thought you meant changing things so that ltsp would boot with my nbdroot convention rather than yours...
| |
| 21:32 | <Gadi> we would need a way for your script to see it other than cmdline
| |
| 21:33 | well, that would be clean, but not backwards-compat
| |
| 21:33 | <Yoe> I suppose not
| |
| 21:34 | What I'll do is check for a variable $NBDROOT (note capitalization), and if that exists and is non-empty, favour it over /proc/cmdline
| |
| 21:35 | <Gadi> that would be nice
| |
| 21:35 | <Yoe> it would then be parsed as if it were /proc/cmdline -- e.g., would contain stuff like "nbdroot=server,port root=/dev/nbd0"
| |
| 21:36 | <Gadi> hmm
| |
| 21:36 | that's a bit confuding
| |
| 21:36 | (confusing
| |
| 21:36 | I would have expected $NBDROOT to be the contents of nbdroot=
| |
| 21:37 | <Yoe> mm, yeah
| |
| 21:37 | I wanted to do that so you could override the contents of root=, too, but I just realize that's not actually necessary
| |
| 21:37 | you can just always use the three-option version of nbdroot=
| |
| 21:38 | <Gadi> there's a three option version now?
| |
| 21:38 | <Yoe> yes
| |
| 21:38 | * Gadi is still reading lucid's code | |
| 21:38 | <Gadi> :)
| |
| 21:38 | adrianorg has joined IRC (adrianorg!~adrianorg@189.58.228.178.dynamic.adsl.gvt.net.br) | |
| 21:38 | <Yoe> nbdroot=<server>,<port>,<device>
| |
| 21:38 | <Gadi> ah
| |
| 21:38 | <Yoe> or <server>,<name>,<device>
| |
| 21:38 | <Gadi> can you make device default to /dev/nbd0?
| |
| 21:38 | (if not specified)
| |
| 21:38 | <Yoe> I think it already does, but let me double-check
| |
| 21:40 | ah, no, it doesn't
| |
| 21:42 | ah, right -- it used to default to eth0, but that's been fixed
| |
| 21:42 | * Yoe was confused about which part it defaulted to something useful. | |
| 21:45 | <Yoe> pff
| |
| 21:45 | * Yoe refactors the README.Debian | |
| 21:45 | <Yoe> it's nowhere near readable anymore
| |
| 21:45 | <Gadi> hehe
| |
| 21:46 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 21:48 | <Gadi> yeah, so it seems to me, that we can have a local-top/ script that does things like, if you use nbd-proxy, start it up and set NBDROOT="127.0.0.1,<port>,nbd0" or some such
| |
| 21:48 | though, I guess we would have to configure networking in local-top/ then
| |
| 21:48 | hmm...
| |
| 21:48 | <Yoe> I should point out that nbd-proxy is a terrible idea, and that you should deprecate it
| |
| 21:49 | it increases your chance for deadlocks
| |
| 21:49 | <Gadi> have you discussed the topic with stgraber, mgariepy and that gang?
| |
| 21:49 | <Yoe> especially if you're going to swap over the device
| |
| 21:49 | nope, I don't know who to talk to about that. But I really really _really_ think it should go
| |
| 21:49 | <Gadi> ah, good!
| |
| 21:50 | you guys really need to talk
| |
| 21:50 | :)
| |
| 21:50 | <Yoe> perhaps that's just me, though... at any rate, I can't tell you guys what you should or should not do :-)
| |
| 21:50 | <Gadi> it only exists because it was needed
| |
| 21:50 | <Yoe> for what purposee?
| |
| 21:50 | <Gadi> and it was only needed because they didn't talk to upstream
| |
| 21:50 | :)
| |
| 21:50 | afaik
| |
| 21:50 | so, it handles things like
| |
| 21:50 | say the server is rebooted
| |
| 21:50 | but the clients are still on
| |
| 21:51 | server comes back
| |
| 21:51 | and the clients reconnect
| |
| 21:51 | and nbd is happy
| |
| 21:51 | <Yoe> that's what -persist is supposed to solve, though I'll readily admit that it doesn't work as it's supposed to
| |
| 21:51 | <stgraber> Gadi: well, in this case, upstream was the kernel ;) we didn't feel like implementing failover, dns resolving, ... in the kernel :)
| |
| 21:51 | <Yoe> you don't need to
| |
| 21:52 | you just need to return from the NBD_DO_IT ioctl() with a code that says "we lost the connection, help!"
| |
| 21:52 | and then nbd-client can do the handshake again
| |
| 21:52 | <stgraber> Yoe: yeah, -persist should be deprecated, if I remember its implementation correctly, it's close to the ugliest hack in the upstream kernel :)
| |
| 21:52 | <Yoe> stgraber: eh?
| |
| 21:52 | it should be fixed, but deprecated? I don't see why
| |
| 21:52 | <stgraber> Yoe: except that in this csae, nbd-client dies and gets respawned, which won't work as you lost your root filesystem
| |
| 21:52 | <Yoe> yes, that's what's going wrong
| |
| 21:53 | nbd-client needs to fork() to read from the device so the partition table gets read
| |
| 21:53 | Steve_the_Pirate has left IRC (Steve_the_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com, Quit: Leaving) | |
| 21:53 | <stgraber> Yoe: IIRC it's respawning nbd-client in user space from kernel space, which should never be done by a kernel module, though the alternative is to create the socket from kernel space which isn't that much better :)
| |
| 21:53 | <Yoe> no, it's not respawning nbd-client at al
| |
| 21:54 | it returns from NBD_DO_IT with an error code
| |
| 21:54 | and then waits until the process exits
| |
| 21:54 | if it does, it will start producing errors for processes trying to read from its device
| |
| 21:54 | pimpministerp is now known as primeministerp | |
| 21:54 | <Yoe> if it doesn't, it will block those processes
| |
| 21:54 | that gives nbd-client a fighting chance to reconnect
| |
| 21:56 | <Gadi> how long is that window of opportunity ? (before the process exits)
| |
| 21:57 | <Yoe> Gadi: the idea is that nbd-client won't exit until either it has set up the connection again, or it discovers that it's not going to be able to set it up anymore and gives up
| |
| 21:57 | well, and when it's set up the connection again, it won't exit either, of course
| |
| 21:57 | <Gadi> can you tell it to never give up?
| |
| 21:57 | <Yoe> that's what the -persist parameter is *supposed* to do. Only it doesn't, and I'm not entirely sure why
| |
| 21:58 | <Gadi> ah
| |
| 21:58 | <stgraber> yeah, that's the whole giving up part that's a bit annoying :) I don't mind all of my system being in IO wait for 2 hours, retrying a connect every few seconds
| |
| 21:58 | <Yoe> it might just be that Paul Clements didn't actually get the patch into the kernel like he said he would... wouldn't be the first time.
| |
| 21:58 | <stgraber> current implement of -persist works if an immediate reconnect works, which at least in the cases I've seen, never happens
| |
| 21:59 | <Yoe> also, I've had reports of -persist breaking disconnect
| |
| 21:59 | * stgraber used to be dealing with load-balanced + HA clusters where failover can take 2-3s to actually happen and only saw very few of the >5000 clients reconnect fine (they're now all reconnecting when it happens) | |
| 22:00 | <Gadi> Yoe: have you looked at the nbd-proxy code?
| |
| 22:01 | perhaps there is something there that can enhance the upstream module
| |
| 22:01 | <Yoe> I haven't, no
| |
| 22:01 | <matrix3000> wait, you have over 5000 thin clients?
| |
| 22:01 | <Gadi> may be worth a glance
| |
| 22:02 | there could be some useful bits
| |
| 22:02 | that may be able to get upstream
| |
| 22:02 | and solve a bunch of things at once
| |
| 22:05 | Yoe: the code is in the ltsp package
| |
| 22:05 | it is not in a separate package
| |
| 22:05 | if you go looking for it
| |
| 22:09 | Trixboxer has left IRC (Trixboxer!~Trixboxer@office.supportdepartment.net, Quit: "Achievement is not the end, its the beginning of new journey !!!") | |
| 22:15 | <Yoe> Gadi: Adding a default is going to be very fishy if I want to remain backwards compatible with older versions of my script
| |
| 22:15 | Gadi: but you can use the three-option variant, and default to nbd0 in your script, if you want to
| |
| 22:16 | <Gadi> ok
| |
| 22:16 | that works
| |
| 22:30 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 22:30 | BonaC is now known as bonac | |
| 22:35 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 260 seconds) | |
| 22:36 | loather has joined IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net) | |
| 22:38 | primeministerp has left IRC (primeministerp!~ppouliot@nat/novell/x-ixfxihdahzstyvxf, Ping timeout: 258 seconds) | |
| 22:43 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 248 seconds) | |
| 22:49 | <Yoe> http://anonscm.debian.org/gitweb/?p=users/wouter/nbd.git;a=commit;h=a68fec1f04c8c95d2af704eb9629ec59d077fd2a
| |
| 22:56 | loather has left IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net, Quit: This computer has gone to sleep) | |
| 22:59 | dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...) | |
| 23:02 | <andygraybeal_> matrix3000, :)) hey man what is the name of that directory project again?
| |
| 23:43 | lifeboy_ has left IRC (lifeboy_!~roland@196-215-48-150.dynamic.isadsl.co.za, Ping timeout: 248 seconds) | |
| 23:53 | khildin_ has left IRC (khildin_!~khildin@ip-80-236-227-210.dsl.scarlet.be, Remote host closed the connection) | |