IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 19 April 2010 (all times are UTC)
| 00:03 | vagrantc has joined #ltsp | |
| 00:08 | vmlintu has joined #ltsp | |
| 00:12 | alkisg has joined #ltsp | |
| 00:16 | vagrantc has quit IRC | |
| 00:19 | vmlintu has quit IRC | |
| 00:19 | alkisg has quit IRC | |
| 00:26 | Egyptian[Home] has quit IRC | |
| 00:28 | Egyptian[Home] has joined #ltsp | |
| 01:26 | vmlintu has joined #ltsp | |
| 01:26 | herytjb has joined #ltsp | |
| 01:27 | herytjb has joined #ltsp | |
| 01:27 | herytjb has joined #ltsp | |
| 01:41 | mikkel has joined #ltsp | |
| 01:41 | alkisg has joined #ltsp | |
| 01:43 | ogra has quit IRC | |
| 01:43 | ogra has joined #ltsp | |
| 01:58 | ogra has quit IRC | |
| 01:58 | gnunux has joined #ltsp | |
| 01:59 | ogra has joined #ltsp | |
| 02:04 | alkisg has quit IRC | |
| 02:04 | alkisg has joined #ltsp | |
| 02:08 | <gnunux> hi
| |
| 02:12 | otavio has quit IRC | |
| 02:14 | alkisg has quit IRC | |
| 02:16 | pths has joined #ltsp | |
| 02:30 | ats_ayas has joined #ltsp | |
| 02:31 | ats_ayas has left #ltsp | |
| 03:34 | alkisg has joined #ltsp | |
| 04:01 | gnunux has quit IRC | |
| 04:02 | gnunux has joined #ltsp | |
| 04:29 | vmlintu has quit IRC | |
| 04:39 | litlebuda has joined #ltsp | |
| 04:40 | C_Tek has joined #ltsp | |
| 04:51 | C_Tek has quit IRC | |
| 05:20 | mischko has joined #ltsp | |
| 05:27 | alkisg has quit IRC | |
| 05:28 | alkisg has joined #ltsp | |
| 05:36 | johnny has left #ltsp | |
| 05:59 | bobby_C has joined #ltsp | |
| 06:12 | mischko_ has joined #ltsp | |
| 06:13 | mischko__ has joined #ltsp | |
| 06:15 | mischko__ has joined #ltsp | |
| 06:16 | hersonls has joined #ltsp | |
| 06:16 | mischko has quit IRC | |
| 06:17 | mischko_ has quit IRC | |
| 06:23 | pmatulis has joined #ltsp | |
| 06:28 | mischko has joined #ltsp | |
| 06:55 | jammcq has quit IRC | |
| 07:03 | rjune has quit IRC | |
| 07:05 | otavio has joined #ltsp | |
| 07:26 | evilx has joined #ltsp | |
| 07:35 | litlebuda has quit IRC | |
| 08:09 | Gadi has joined #ltsp | |
| 08:20 | <_UsUrPeR_> morning all
| |
| 08:53 | <vbundi> mornin
| |
| 08:56 | jcastro has joined #ltsp | |
| 08:56 | epaphus has joined #ltsp | |
| 09:07 | japerry has quit IRC | |
| 09:22 | mikkel has quit IRC | |
| 09:32 | litlebuda has joined #ltsp | |
| 09:35 | <alkisg> stgraber: ping? (about udhcp, whenever you got 5'...)
| |
| 09:40 | jammcq has joined #ltsp | |
| 09:40 | <jammcq> hello friends
| |
| 09:46 | <highvoltage> hello uncle jammcq
| |
| 09:55 | brettbjorgaardu has quit IRC | |
| 10:07 | jbrett has quit IRC | |
| 10:07 | jbrett has joined #ltsp | |
| 10:15 | monteslu has quit IRC | |
| 10:18 | komunista has joined #ltsp | |
| 10:19 | japerry has joined #ltsp | |
| 10:22 | litlebuda has quit IRC | |
| 10:22 | kusznir_ is now known as kusznir | |
| 10:33 | vagrantc has joined #ltsp | |
| 10:34 | gnunux has quit IRC | |
| 10:35 | treyhunsucker has joined #ltsp | |
| 10:35 | <treyhunsucker> Gadi: no reports as of yet about anyone loosing audio
| |
| 10:35 | *Cheers*
| |
| 10:38 | one thing i've noticed on this install
| |
| 10:38 | is after fulling upgrading ub 9.10 and the chroot
| |
| 10:38 | i use about 1/3 of the memory
| |
| 10:39 | i made some changes on my 2003 TS to limit the nic to only use 1 core on 1 cpu
| |
| 10:39 | *2008 TS
| |
| 10:39 | i wonder if that also helped in a strange way?
| |
| 10:39 | japerry has quit IRC | |
| 10:39 | <vbundi> cool, didn't know you could do that
| |
| 10:40 | I think when you upgrade it uninstalls things sometimes, so maybe that's why you are using less memory? less things running
| |
| 10:40 | <treyhunsucker> vbundi: it's a new tool in 2008
| |
| 10:40 | vbundi: yea but I went from using 75% of my ram
| |
| 10:40 | to using 25$
| |
| 10:40 | 25%
| |
| 10:40 | <Gadi> server side or client side?
| |
| 10:40 | <treyhunsucker> Gadi: server
| |
| 10:40 | Faithful has quit IRC | |
| 10:41 | <Gadi> then, it is probably the windows settings that helped - not the ubuntu upgrade
| |
| 10:41 | <treyhunsucker> Gadi: interesting
| |
| 10:41 | i lied, after usage this morning it's peaked at 41% usage
| |
| 10:42 | but still very interesting
| |
| 10:42 | Interrupt-Affinity Policy Tool FTW
| |
| 10:43 | <vbundi> so you are using ubuntu desktops to run rdesktop to your 2008RDS?
| |
| 10:44 | <treyhunsucker> vbundi: yes
| |
| 10:44 | <vbundi> and the 41% memory usage is your LTS?
| |
| 10:44 | <treyhunsucker> vbundi: oops let me correct myself, i'm using the rdesktop screen script
| |
| 10:45 | vbundi: using the rdesktop screen script to connect to 2008 TS, and memory is at 41%
| |
| 10:45 | <vbundi> how many clients is that
| |
| 10:45 | and 41% of what
| |
| 10:45 | <treyhunsucker> 41% of memory usage
| |
| 10:45 | <vbundi> of how much ;P
| |
| 10:45 | 8GB?
| |
| 10:45 | <treyhunsucker> vbundi: 1.5g
| |
| 10:46 | i have 100 clients
| |
| 10:46 | but not all of them are in use at a time
| |
| 10:46 | i think the average load is around 40-50
| |
| 10:46 | <vbundi> ah ok
| |
| 10:47 | do you use samba with it or are you running everything on windows
| |
| 10:47 | <treyhunsucker> windows for everything
| |
| 10:49 | <vbundi> ah, I tried messing with Server 08 a while back.. I liked it, but I wasn't having much success getting samba to work with it
| |
| 10:49 | <treyhunsucker> vbundi: yea a lot of things have changed, probably need to join your LTS to the domain
| |
| 10:50 | <vbundi> treyhunsucker: lol, can you do anything with windows without being joined to the domain? ;)
| |
| 10:50 | <treyhunsucker> vbundi: you can ping it :)
| |
| 10:51 | i absolutely hate windows
| |
| 10:51 | but there is simply no replacement for a domain style environment that has all the features
| |
| 10:53 | japerry has joined #ltsp | |
| 10:54 | <treyhunsucker> stgraber: hello sir are you around?
| |
| 10:54 | <vbundi> sabayon is sweet
| |
| 10:55 | Faithful has joined #ltsp | |
| 10:55 | <vbundi> what a great idea
| |
| 10:56 | <treyhunsucker> what is sabayon
| |
| 10:58 | <vbundi> it lets you make profiles and preferences for users
| |
| 10:58 | <treyhunsucker> vbundi: from a live cd?
| |
| 10:59 | <vbundi> when you run it and create a profile... it runs a session that is just like a normal desktop... but any preferences that you save... ie fontsize or background or desktop icons..you can save as a template to apply to users or groups
| |
| 10:59 | <treyhunsucker> vbundi: cool
| |
| 11:00 | <komunista> hi all
| |
| 11:00 | <vbundi> yeah, a little simpler than running through 1000 policies and checkboxes etc
| |
| 11:00 | hi
| |
| 11:04 | <elias_a> Hello, folks!
| |
| 11:07 | <treyhunsucker> hello: elias_a
| |
| 11:07 | komunista: hello
| |
| 11:07 | elias_a: hello
| |
| 11:08 | <elias_a> What are you guys up to?
| |
| 11:18 | <vbundi> gettin italc running
| |
| 11:18 | <treyhunsucker> elias_a: working on italc also
| |
| 11:18 | vbundi: you install it on the server?
| |
| 11:18 | vbundi: or the clients?
| |
| 11:18 | <vbundi> both
| |
| 11:18 | just install italc-client in your chroot
| |
| 11:19 | I'm getting an error saying my keys aren't generated or my service isnt' running
| |
| 11:24 | epaphus has quit IRC | |
| 11:24 | komunista has quit IRC | |
| 11:29 | jhutchins_lt has quit IRC | |
| 11:31 | komunista has joined #ltsp | |
| 11:33 | johnny has joined #ltsp | |
| 11:33 | Faithful has quit IRC | |
| 11:37 | Faithful has joined #ltsp | |
| 11:40 | <vbundi> cool! :) got italc running
| |
| 11:45 | epaphus has joined #ltsp | |
| 11:55 | <treyhunsucker> vbundi: cool how much memory is it using on each client
| |
| 11:55 | <vbundi> not sure I'll look in a few
| |
| 12:03 | <johnny> alkisg, how much memory is being used before swap on your fat clients?
| |
| 12:03 | err lemme try that again
| |
| 12:04 | how much memory is being used on login with your fat clients
| |
| 12:11 | treyhunsucker has left #ltsp | |
| 12:11 | slidesinger has quit IRC | |
| 12:14 | atkuepker has joined #ltsp | |
| 12:30 | litlebuda has joined #ltsp | |
| 12:40 | Lns has joined #ltsp | |
| 12:45 | japerry has quit IRC | |
| 12:45 | japerry has joined #ltsp | |
| 12:47 | Egyptian[Home] has quit IRC | |
| 12:50 | komunista has quit IRC | |
| 12:52 | <alkisg> johnny: `free` says 96M, `htop` says 30M.
| |
| 12:53 | <johnny> with how much ram?
| |
| 12:53 | <alkisg> 768
| |
| 12:53 | (I'm disabling cron in my setups with an ltsp-build-client plugin)
| |
| 12:53 | <johnny> does that really help?
| |
| 12:53 | <alkisg> Yes, it stops a lot of annoying tasks
| |
| 12:53 | <johnny> for example?
| |
| 12:53 | <alkisg> E.g. updatedb, update-xapian-index etc
| |
| 12:54 | There are dozens of them in a fat client installation
| |
| 12:54 | <johnny> hmm.. with 512mb ram, it's already touching swap.. :(
| |
| 12:54 | and things get kinda slow
| |
| 12:54 | <alkisg> And that's before logon?! Wow...
| |
| 12:55 | I'm also using nfs-client, that also takes a bit of ram before logon.
| |
| 12:55 | (it's included in the 96 M)
| |
| 12:56 | <johnny> no.. on login
| |
| 12:56 | before opening anything else
| |
| 12:57 | <alkisg> After gnome logon?
| |
| 12:57 | <johnny> yes
| |
| 12:57 | <alkisg> OK let me try...
| |
| 12:57 | <johnny> i thought i was gonna be ok with 512mb ram, but i don't think so..
| |
| 12:58 | i should probably go back to my gigabit card
| |
| 12:58 | we were having problems with that before
| |
| 12:58 | i think there was a bad driver during jaunty
| |
| 12:58 | <alkisg> What I'm telling my users is "512 is the threshold, below that use the clients as thin, above that, as fat"
| |
| 12:59 | With exactly 512 it's hard to decide... with a good setup you should be able to work but not with a lot of apps open simultaneously...
| |
| 13:00 | <johnny> usually ti's just firefox and/or openoffice
| |
| 13:00 | i wish i could upgrade them more
| |
| 13:00 | <alkisg> Well firefox with 1 tab is very very different than firefox with 10 tabs :)
| |
| 13:01 | <johnny> i wish i had a pc tech who could tell me what's wrong with this computer i used to use as a server :(
| |
| 13:01 | <alkisg> After logon and with just an xterm open, `free` tells me 282M
| |
| 13:01 | <johnny> yeah.. that sounds about right then.. with 768
| |
| 13:01 | <alkisg> No, with 512
| |
| 13:01 | <johnny> oh.. hmm
| |
| 13:01 | <alkisg> And I'm talking about the used memory
| |
| 13:01 | Free == 227M
| |
| 13:02 | <johnny> maybe it's just my network link that's getting congested
| |
| 13:02 | i'm going to switch to back to the gigabit card
| |
| 13:02 | <alkisg> johnny: ps aux should tell you the exact details...
| |
| 13:02 | If you want, upload it to pastebin
| |
| 13:02 | <johnny> i'm not there
| |
| 13:02 | <alkisg> Whenever...
| |
| 13:04 | <Lns> Does anyone have any experience with E-Rate funding and LTSP servers?
| |
| 13:09 | Gadi: ping? ^^^ ? Do you know if your boot appliances qualify for e-rate funding, since they don't typically house application software?
| |
| 13:10 | denisesball has joined #ltsp | |
| 13:10 | <denisesball> hey all, i know some key components of ltsp-cluster are being rewritten. would it be wise to wait until its completed before deploying in a large environment?
| |
| 13:10 | <Gadi> Lns: hey
| |
| 13:10 | <Lns> hey gadi =)
| |
| 13:11 | <Gadi> we are an erate vendor
| |
| 13:11 | what dya want to know?
| |
| 13:11 | (not that I am an erate expert)
| |
| 13:12 | <Lns> Gadi: I'm just trying to get some grips on E-Rate in general and what LTSP functionality is funded..I can see end-user terminals aren't, as well as things that house app. software..but it would seem that things like your boot appliance would be eligable
| |
| 13:12 | Egyptian[Home] has joined #ltsp | |
| 13:12 | <Lns> eligible*
| |
| 13:12 | <Gadi> as far as I understand, there are multiple categories within erate
| |
| 13:12 | <Lns> seems like it's a big grey area for terminal services, though i read an article that even citrix software is eligible
| |
| 13:13 | <Gadi> categories of types of equipment
| |
| 13:13 | <Lns> right
| |
| 13:13 | http://www.olis.ri.gov/grants/erate/eligible.pdf
| |
| 13:13 | <Gadi> and depending on how many kids get free lunch, that determines what categories you can buy stuff from
| |
| 13:14 | <Lns> the free lunch factors in with what categories are eligible?? weird.
| |
| 13:14 | <johnny> denisesball, probably not, but you should probably ask on the mailing list
| |
| 13:14 | <Lns> I thought it factored with the percentage overall it would fund
| |
| 13:14 | maybe both..
| |
| 13:14 | <denisesball> johnny: ok, they have this irc channel listed on their site, so thought i would take a stab
| |
| 13:15 | <johnny> sure, some of the cluster devs and users are here
| |
| 13:15 | but, not as active as regular ltsp users yet
| |
| 13:15 | <denisesball> also i find the ltsp docs incredibly confusing
| |
| 13:15 | seems to be a lot of different names for the same things
| |
| 13:16 | <vagrantc> denisesball: for example?
| |
| 13:16 | <johnny> cluster docs?
| |
| 13:16 | or regular?
| |
| 13:16 | <vagrantc> denisesball: and which docs?
| |
| 13:16 | <Lns> Gadi: ah this one is the most up to date (2010).. http://hraunfoss.fcc.gov/edocs_public/attachmatch/FCC-09-105A2.pdf
| |
| 13:17 | * vagrantc suspects that LOCALDEV, LOCAL_APPS, REMOTE_APPS, FATCLIENTS and SOUND support might confuse people | |
| 13:18 | <johnny> luckily localdev and sound are enabled by default
| |
| 13:18 | so most people dont' even have to know they exist
| |
| 13:18 | <denisesball> vagrantc/johnny: cluster docs. comparing this doc - https://www.ltsp-cluster.org/documentation/technical-introduction to this doc https://help.ubuntu.com/community/UbuntuLTSP/LTSP-Cluster
| |
| 13:18 | <johnny> oh.. for cluster
| |
| 13:19 | those aren't the ltsp docs
| |
| 13:19 | <vagrantc> ah, barely know anything about the cluster stuff
| |
| 13:19 | <johnny> they have yet to be merged into the ltsp docs
| |
| 13:19 | <denisesball> right, i said cluster
| |
| 13:19 | <johnny> the ltsp docs are fine :)
| |
| 13:19 | <denisesball> trying to understand all the pieces
| |
| 13:19 | <johnny> mostly anyways :)
| |
| 13:19 | <vagrantc> "11:15 < denisesball> also i find the ltsp docs incredibly confusing"
| |
| 13:19 | <johnny> denisesball, cluster is still somewhat of a seperate project
| |
| 13:19 | <vagrantc> :)
| |
| 13:19 | <denisesball> right, i clarified that...
| |
| 13:19 | i never had a problem with ltsp docs
| |
| 13:19 | <johnny> your best bet is on tlhe mailing list
| |
| 13:20 | unless stgraber wants to talk..
| |
| 13:20 | <denisesball> i also dont see how to have a fully redundant setup with ltsp-cluster without the root servers being redundant
| |
| 13:20 | which is kinda the whole attraction for me
| |
| 13:24 | jhutchins_lt has joined #ltsp | |
| 13:25 | <denisesball> johnny: would that be the regular ltsp-discuss mailing list then?
| |
| 13:27 | <johnny> not sure, if they don't have a mailing list on ltsp-cluster.org
| |
| 13:27 | then yes
| |
| 13:27 | otherwise no
| |
| 13:31 | japerry has quit IRC | |
| 13:34 | <alkisg> vagrantc: the udhcp debian/ubuntu package is using ancient code (more than 5 years old), because udhcp is now being maintained in busybox. So the busybox people adviced me to notify the udhcp package maintainer to either drop the package completely, or to use busybox as the source to build it.
| |
| 13:34 | What's the correct way to handle that? I suppose I could file a bug in launchpad/ubuntu, but I don't think that would help... should I file a bug in debian even though I use ubuntu?
| |
| 13:36 | <vagrantc> alkisg: whatever feels comfortable to you...
| |
| 13:36 | <alkisg> Neither :)
| |
| 13:36 | <vagrantc> hah
| |
| 13:36 | <alkisg> Does LTSP in debian use udhcp, if it exists?
| |
| 13:36 | <vagrantc> ah!
| |
| 13:36 | alkisg: i install the hooks, but do not install the package.
| |
| 13:37 | alkisg: no need to do anything: http://packages.qa.debian.org/u/udhcp/news/20100228T163929Z.html
| |
| 13:37 | <alkisg> Ah, nice!
| |
| 13:37 | Now we need to go to the second phase, i.e. ask the busybox-initramfs maintainers to add udhcpc as a build target :)
| |
| 13:37 | stgraber: ^^^
| |
| 13:38 | <vagrantc> weird... it was removed from testing, but not from unstable...
| |
| 13:38 | but it's not present in unstable...
| |
| 13:39 | <alkisg> vagrantc: if you run /usr/lib/initramfs-tools/bin/busybox, does it tell you that udhcpc is an available command in debian?
| |
| 13:40 | <vagrantc> alkisg: from the initramfs?
| |
| 13:40 | <alkisg> No, directly
| |
| 13:40 | (if you have busybox-initramfs installed, that is...)
| |
| 13:40 | (either in the server or the chroot)
| |
| 13:40 | <vagrantc> there's no /usr/lib/initramfs-tools
| |
| 13:41 | <alkisg> dpkg -L busybox-initramfs
| |
| 13:41 | <vagrantc> there's no busybox-initramfs package in debian
| |
| 13:41 | <alkisg> Nah, ubuntu-ism... :(
| |
| 13:41 | OK, then the busybox itself?
| |
| 13:42 | (busybox-static, I believe)
| |
| 13:42 | <vagrantc> seems like debian busybox package ships a udhcpc/udhcpd package
| |
| 13:42 | <alkisg> Aaahhh! http://packages.debian.org/sid/source/busybox
| |
| 13:43 | <vagrantc> that was mysterious
| |
| 13:43 | <alkisg> OK, that's the best news then, it requires the less work from me :D
| |
| 13:43 | They actually dropped the ancient code and are building it from busybox
| |
| 13:43 | <vagrantc> you just have to say, "Ubuntu, do what debian is doing!"
| |
| 13:43 | <alkisg> Right :)
| |
| 13:43 | Thanks!
| |
| 13:45 | <vagrantc> though busybox (and hence, udhcpc) seems blocked from migration to testing
| |
| 13:45 | <alkisg> What does that mean?
| |
| 13:46 | <vagrantc> presumeably since busybox so core to debian-installer, they've frozen it's migration, and the version in testing doesn't have the udhcp stuff ... though i'm guessing they'll let it move before release.
| |
| 13:47 | denisesball has left #ltsp | |
| 13:55 | * alkisg filed LP #566845, now we just need a good way to detect the new version in the LTSP sources so that we use the -C parameter instead of the broken --clientid... | |
| 14:00 | rjune has joined #ltsp | |
| 14:15 | <Gadi> alkisg: so clientid doesn't work atm?
| |
| 14:19 | <alkisg> Gadi: it *only* works with dhcp3-server
| |
| 14:19 | <_UsUrPeR_> hey all. Can somebody point me toward the correct way to boot a non-thincan etherboot client in 10.04?
| |
| 14:19 | <alkisg> It breaks on windows dhcp servers, cisco routers etc...
| |
| 14:19 | <_UsUrPeR_> Previously, I was able to use mknbi to create a new initramfs, but that does not seem to cut it anymore.
| |
| 14:19 | <Lns> vagrantc: have you ever had to deal with session processes lingering after user logout, and/or gnome-watchdog?
| |
| 14:19 | <alkisg> So for Lucid I have to stop using udhcp, it's broken for me
| |
| 14:20 | <vagrantc> Lns: it's been a recurring problem over the years, with varying degrees of evilness.
| |
| 14:21 | Lns: the biggest thing i did to improve it was to teach users that ctrl-alt-backspace wasn't the same thing as logging out cleanly.
| |
| 14:21 | <Lns> vagrantc: gotcha.. I think I'll test out gnome-watchdog on lenny. Wonder if it'll even install ;)
| |
| 14:21 | <vagrantc> there were several generations of freegeekers who taught each sucessive generation to just use ctrl-alt-backspace...
| |
| 14:22 | <Lns> vagrantc: hehe... =) Well, I can see that you could always 1) disable CTRL+ALT+BACKSPACE, or 2) have something like g-w clean up afterwards if used.. Personally I love the neat shortcuts X provides and don't like it when distros disable it :)
| |
| 14:24 | looks like gnome-watchdog installs (after installing gtkdialog of course as a dep)
| |
| 14:24 | brb, gonna re-log in to enable it
| |
| 14:24 | Lns has quit IRC | |
| 14:25 | Lns has joined #ltsp | |
| 14:26 | <Lns> coolio
| |
| 14:28 | <atkuepker> Lns: we're seeing it even after proper logouts. Haven't devised a good solution yet, but it's on my to-do list for this quarter.
| |
| 14:28 | <Lns> atkuepker: you've seen gnome-watchdog right?
| |
| 14:29 | <Gadi> alkisg: what does that mean? you mean sending clientid="" only works?
| |
| 14:29 | <alkisg> Gadi, with a cisco router, my clients hang all the time
| |
| 14:30 | So I have to use IPAPPEND=3 (==bypass udhcpc completely) in order for them to work
| |
| 14:30 | The dhcp rfc states that clientid should be unique for each client on the local network
| |
| 14:30 | <atkuepker> Lns: no, not really.
| |
| 14:30 | <Lns> atkuepker: you running debian/ubuntu ltsp?
| |
| 14:30 | <alkisg> Gadi: so I guess that the cisco router considers all my LTSP clients as the same client or something like that...
| |
| 14:30 | <atkuepker> Am still coming up to speed after 8 years of KDE & Fedora
| |
| 14:31 | but, yes
| |
| 14:31 | <Lns> atkuepker: https://help.ubuntu.com/community/UbuntuLTSP/GnomeWatchdog
| |
| 14:31 | atkuepker: it's worked great for my sites for a couple years now
| |
| 14:32 | <Gadi> alkisg: but hangs when specifying a clientid or just hangs when using udhcpc in general?
| |
| 14:32 | <alkisg> Gadi: udhcpc by default generates a unique clientid, so it works
| |
| 14:32 | But in LTSP we force it to use clientid ''
| |
| 14:32 | ...and that's what breaks it
| |
| 14:33 | The problem is that by default that old udhcpc version causes dhcp3-server to generate duplicate leases
| |
| 14:33 | <Gadi> alkisg: I actually tried something wrt clientid the other day that wasnt working - where would one specify a custom clientid?
| |
| 14:33 | <vagrantc> Lns: i've heard mention of this gnome-watchdog thing ...
| |
| 14:34 | * vagrantc wonders why it hasn't been uploaded to debian/ubuntu | |
| 14:34 | <alkisg> Gadi: the proper solution is to use a newer udhcpc version, one that supports the new -C switch
| |
| 14:34 | -C == "don't provide a clientid at all"
| |
| 14:34 | That's acceptable by the dhcp rfc, and is not causing duplicate leases
| |
| 14:34 | <Lns> vagrantc: It was put together by the guy who did schooltool i think, it was an independent project that never got any dist. attention i think
| |
| 14:34 | <abeehc> gnome-watchdog works ok for me.. not so great for auto-login users i find
| |
| 14:35 | <vagrantc> hm.
| |
| 14:35 | <Lns> abeehc: yes, I haven't tried gnome-watchdog with auto/guest users before
| |
| 14:36 | <Gadi> alkisg: right, but I actually need a clientID
| |
| 14:36 | :)
| |
| 14:36 | <alkisg> Gadi, why?
| |
| 14:36 | dhclient doesn't work for you?
| |
| 14:36 | <Gadi> I use it to distinguish between thin clients and PCs
| |
| 14:36 | <Lns> vagrantc: I'd love to see it in deb/ubu, since my domain seems to be the only place that even hosts the .deb file anymore ;)
| |
| 14:36 | <alkisg> or PXE ?
| |
| 14:36 | Gadi, clientid != vendor id
| |
| 14:36 | <Gadi> ah, really?
| |
| 14:37 | is there a vendor id flag for udhcpc?
| |
| 14:37 | <alkisg> Uh, /me looks...
| |
| 14:37 | <Lns> and obviously it's pretty crucial.. though ogra has said before that *some* cleanup is performed in newer LTSPs, but only when the user re-logs in (which doesn't cover any runaway procs after they log out)
| |
| 14:38 | <alkisg> Gadi: line 166 in : http://git.busybox.net/busybox/tree/networking/udhcp/dhcpc.c?h=1_16_stable&id=53283adb24765a7afb4d6298661c3c1a8d6f5601
| |
| 14:38 | (that's the newer udhcp, I hope I never see again the old one :))
| |
| 14:38 | <Gadi> ah, -V
| |
| 14:40 | <alkisg> Right. So I think we should stop using --clientid completely, and use -C instead, and provide some way for the users to specify -V if they need
| |
| 14:41 | <atkuepker> Lns: Thanks. So, if /etc/check_previous_login exists, it forces users to only be logged into 1 terminal at a time, yes?
| |
| 14:42 | <Lns> atkuepker: I think it enforces it by a gtkdialog asking them if they want to kill the previous session during login
| |
| 14:42 | atkuepker: but yes, it's got the general principle of having only one session per user
| |
| 14:43 | <atkuepker> ah. That's a good start at least. Next time you visit Seattle, let me know. I'll buy you a beer. =)
| |
| 14:44 | <Gadi> crap
| |
| 14:44 | no -V in the old one
| |
| 14:45 | read: existing Ubuntu one
| |
| 14:45 | <Lns> atkuepker: ;) Hey I just host the package and made the doc for it, I didn't make the script =) but sure, I never turn down free beer!
| |
| 14:46 | mischko_ has joined #ltsp | |
| 14:46 | <alkisg> Gadi: yeah, it's ancient. We'll need to file an SRU to get it into Lucid, it seems.
| |
| 14:47 | <Lns> vagrantc: ooo, lenny+ crafts a sane sources.list including updates and stuff? :)
| |
| 14:47 | (during ltsp-build-client)
| |
| 14:48 | <Gadi> alkisg: now to see which patch I prefer - busybox/udhcpc or klibc
| |
| 14:48 | <alkisg> Gadi, klibc? What do you mean?
| |
| 14:48 | ipconfig?
| |
| 14:48 | <Gadi> yeah
| |
| 14:48 | mischko has quit IRC | |
| 14:48 | <Gadi> I think that one needed patching, too
| |
| 14:48 | I have to refresh my memory
| |
| 14:49 | <alkisg> Yes it was pretty broken, but it worked for some setups
| |
| 14:49 | treyhunsucker has joined #ltsp | |
| 14:49 | <vagrantc> Lns: i don't know that it ever didn't craft a sane sources.list.
| |
| 14:49 | <alkisg> Gadi: If it worked on your environment, you just need to remove udhcp from the chroot (not purge it, just mv it)
| |
| 14:49 | <Lns> vagrantc: i was comparing to (older?) ubu releases
| |
| 14:50 | <treyhunsucker> Gadi: bad news, audio is randomly dying on clients
| |
| 14:50 | to restate the issue
| |
| 14:51 | i am using the rdestop screen script, have SOUND= FALSE in my lts.conf in the default section. The clients randomly loose audio, not choppy or anything like that, just stops until they reboot
| |
| 14:52 | pmatulis has quit IRC | |
| 14:57 | <treyhunsucker> I found a similiar issue with jaunty, which is what i'm using, that requires the audio fix to do this: https://help.ubuntu.com/community/UbuntuLTSP/UnlockChrootRootAcct
| |
| 14:57 | that a bad idea?
| |
| 14:58 | <vagrantc> treyhunsucker: just set up a SCREEN_NN=shell
| |
| 14:58 | treyhunsucker: much simpler.
| |
| 14:58 | where NN = 01, .. , 12
| |
| 15:00 | <treyhunsucker> vagrantc: i agree, after posting it they went through and did a bunch of crazy crap not worht it
| |
| 15:00 | *worth it
| |
| 15:00 | when I have sound=false setup in my lts.conf, what app controls sound while using the rdesktop screen script?
| |
| 15:06 | <Gadi> treyhunsucker: rdesktop controls everything
| |
| 15:06 | rdesktop talks directly to alsa
| |
| 15:07 | if sound just dies, it points towards the driver failing
| |
| 15:09 | mischko_ has quit IRC | |
| 15:09 | mischko has joined #ltsp | |
| 15:13 | <treyhunsucker> Gadi: interesting
| |
| 15:13 | gadi: so my next step would to be update the driver for each sound card that I have
| |
| 15:17 | <vbundi> I must be lucky... I don't think I've ever had a system where the soundcard wasn't detected
| |
| 15:17 | what kind of systems are you having problems with trey
| |
| 15:18 | <treyhunsucker> vbundi: it's several different kinds
| |
| 15:18 | vbundi: dell optiplex sx280
| |
| 15:19 | vbundi: gateway profile 5
| |
| 15:19 | those are 2 off the top of my head
| |
| 15:19 | <vbundi> optiplex won't work?
| |
| 15:19 | it's an intel board isn't it?
| |
| 15:19 | <treyhunsucker> vbundi: they have always worked until about a month ago
| |
| 15:19 | vbundi: intel yes
| |
| 15:19 | <vbundi> ok try this.. have you got one booted up?
| |
| 15:19 | <treyhunsucker> vbundi: yup yup
| |
| 15:20 | <vbundi> right click your sound applet on your gnome panel
| |
| 15:20 | err left click
| |
| 15:20 | sound preferences
| |
| 15:20 | <treyhunsucker> vbundi: i don't run gnome, I run the rdesktop screen script :)
| |
| 15:20 | <vbundi> well load gnome and try this
| |
| 15:20 | :P
| |
| 15:20 | <treyhunsucker> vbundi: go ahead and give me instructions
| |
| 15:20 | i will do it at the end of the day whe no one is around
| |
| 15:20 | <vbundi> go to your output tab
| |
| 15:21 | there's a dropdown box labeled 'Connector:'
| |
| 15:21 | <treyhunsucker> yes
| |
| 15:21 | <vbundi> my systems that have built in speakers have an option called Analog Output (LFE)
| |
| 15:21 | <treyhunsucker> ok
| |
| 15:22 | <atkuepker> treyhunsucker, we're getting some sx280s in at the end of this week for us to test. If you pop back in next week, I'll try to report back in.
| |
| 15:22 | <vbundi> that makes the internal speaker work... if it's set to that.. you wont get audio on your audio output jack
| |
| 15:22 | <treyhunsucker> vbundi: already been down that path
| |
| 15:23 | <johnny> why does the copy/paste in most of my apps always miss the first letter ;(
| |
| 15:23 | <treyhunsucker> vbundi: on various units on the lts.conf i have things like HEADPHONE_VOLUME=75 and etc
| |
| 15:23 | atkuepker: that would be awesome
| |
| 15:24 | vbundi: for that specific reason, and others things like volume=0
| |
| 15:24 | <vbundi> ah well sorry I couldn't be of any more help... I've got a couple optiplex gx520s and thincenters and thats how they behave
| |
| 15:24 | <treyhunsucker> vbundi: no i appreciate it ty
| |
| 15:25 | evilx has quit IRC | |
| 15:25 | <vbundi> actually I think we have a couple SX280s or 270s or something at the other office
| |
| 15:26 | I'm running Lucid though so I dunno how much that would help ya
| |
| 15:26 | <treyhunsucker> i might actually end up using the gnome desktop on a few
| |
| 15:26 | of course when lucid is released at the end of this month
| |
| 15:26 | if this isn't fixed i'll upgrade to it
| |
| 15:27 | <atkuepker> run today's lucid with a GX240 on my other monitor. Other than the TFTP config files getting overwritten each morning and gnome-screensaver crashing, it's going well.
| |
| 15:28 | treyhunsucker has left #ltsp | |
| 15:28 | <atkuepker> not with the onboard Intel video though. That is still seriously crashy for us for some reason.
| |
| 15:28 | <johnny> i still need to upgrade to lucid
| |
| 15:28 | <vbundi> these thinkcenters are G915 intel video and it works great
| |
| 15:29 | <johnny> because of a hard drive failure, we had to take all our sales manually yesterday
| |
| 15:29 | <vbundi> even compiz works :)
| |
| 15:29 | <johnny> i had a misunderstanding with a coworker of what exactly was being backed up
| |
| 15:29 | <vbundi> johnny: ouch
| |
| 15:29 | <johnny> only /etc/, /home/, and some /var dirs.. not all
| |
| 15:29 | <vbundi> no raid?
| |
| 15:29 | <johnny> no.. no raid.. i don't know nothin about that
| |
| 15:30 | i am buying a second hard drive tho
| |
| 15:30 | <atkuepker> thinkcenter? Is that Lenovo stuff?
| |
| 15:30 | <johnny> and using rdiff-backup nightly
| |
| 15:30 | not sure why i would use raid vs rdiff-backup
| |
| 15:30 | <vbundi> atkuepker: yep, same small form factor stuff as the optiplex's
| |
| 15:30 | <johnny> vbundi, why would i use raid over rdiff-backup?
| |
| 15:30 | <alkisg> atkuepker: try nomodeset as a kernel parameter, it helped on some intel clients for me
| |
| 15:30 | <vbundi> johnny: to avoid downtime
| |
| 15:31 | <johnny> downtime for rdiff-backup would be less than a half hour
| |
| 15:31 | <abeehc> i found nomodeset critical on a couple too
| |
| 15:31 | <atkuepker> alkisg, I'll try that out on the test server. We're buying several hundred at a time, so we're very keen on being able to use a default configuration.
| |
| 15:31 | <johnny> and we get incremental fixes
| |
| 15:31 | raid doesn't do incremental
| |
| 15:31 | at least not the kind of raid we would do
| |
| 15:32 | <vbundi> johnny: so when your hard drive dies, you pop in a new drive and image your stuff over to it?
| |
| 15:32 | <johnny> we're going to have a 2nd one
| |
| 15:32 | 2 equivalent 500g drives
| |
| 15:32 | <vbundi> oh I see
| |
| 15:32 | <johnny> and just rdiff-backup nightly
| |
| 15:32 | <vagrantc> i'll be tagging ldm-trunk and ltsp-trunk today...
| |
| 15:32 | <johnny> since rdiff-backup gives me incremental and full backups
| |
| 15:32 | <vagrantc> unless anyone screams
| |
| 15:32 | <johnny> and i think it should even be bootlable with one small change
| |
| 15:33 | <vbundi> so it's similar to raid in the sense that your data is mirrored, but not inside the physical machine...
| |
| 15:33 | <johnny> vbundi, so really.. it would be downtime enough for a reboot
| |
| 15:33 | just pick the other drive to boot from
| |
| 15:33 | <vbundi> ohh
| |
| 15:33 | even inside the machine, I see
| |
| 15:33 | <johnny> technically rdiff-backup will use more space than a raided setup
| |
| 15:34 | but the incremental backup is worth it
| |
| 15:34 | i just wish it was easier to prune..
| |
| 15:34 | we were backing up to a remote server
| |
| 15:34 | <vbundi> johnny: the slight advantage with raid, would be that in the event of a failure, your machine doesn't go down
| |
| 15:34 | <johnny> ah... that's no big deal really
| |
| 15:34 | <vbundi> yeah for most stuff probably not
| |
| 15:34 | <johnny> if it's down for 5 minutes it's not gonna hurt us
| |
| 15:35 | we take a few orders manually..
| |
| 15:35 | that's it
| |
| 15:35 | <vbundi> we had a fan seize in a 2U case... killed both our drives
| |
| 15:35 | <johnny> if i didn't know we didn't have enough backup space on the server for all our msuic and whatnot
| |
| 15:35 | ogra_cmpc has quit IRC | |
| 15:35 | <johnny> i would have already gotten the other drive
| |
| 15:35 | we don't do tons of business
| |
| 15:35 | my budget goes up by about $50/wk
| |
| 15:36 | i also had a misunderstanding in what should be paid out of the tech budget vs the general fund
| |
| 15:36 | turns out our main computer is under the general fund :)
| |
| 15:36 | <vbundi> nice
| |
| 15:36 | <johnny> i was always holding back $400 to replace it
| |
| 15:36 | now i can buy something nice..
| |
| 15:37 | <vbundi> big monitor for pr0n
| |
| 15:37 | :D
| |
| 15:38 | or not.
| |
| 15:39 | * _UsUrPeR_ looks at the monitor at the wrong time | |
| 15:40 | <johnny> vbundi, it's a bookstore coffeehouse
| |
| 15:40 | i use ltsp for the internet terminals
| |
| 15:41 | wwx has quit IRC | |
| 15:42 | <vbundi> ahh
| |
| 15:42 | wwx has joined #ltsp | |
| 15:47 | ogra_cmpc has joined #ltsp | |
| 15:55 | <Gadi> alkisg: ping
| |
| 15:55 | <alkisg> pong
| |
| 15:55 | <Gadi> do you know what exact "stability issues" led stgraber to turn of compression on the image?
| |
| 15:56 | * Gadi is seeing some instabilities in lucid's mount of compressed image, but I never saw it before nbd-proxy | |
| 15:56 | <Gadi> and I wonder if it's only unstable through nbd-proxy
| |
| 15:57 | <alkisg> No, not exactly. He told me he was having problems with the compressed image but not the exact details.
| |
| 15:57 | I too never had problems with compressed images
| |
| 15:58 | ....and I had several problems with nbd-proxy even with uncompressed images...
| |
| 15:58 | <Gadi> I think when you turn off udhcpc or use IPAPPEND and the boot process goes quicker, it exposes more issues
| |
| 15:58 | <alkisg> Lucid?
| |
| 15:58 | <Gadi> I can reboot three times and get three different errors
| |
| 15:58 | yeah
| |
| 15:58 | lucid
| |
| 15:59 | <alkisg> Many scripts need to be up-startified
| |
| 15:59 | <Gadi> everything from bus errors, to squashfs errors, to no error and the boot goes through fine
| |
| 15:59 | <alkisg> E.g. with a fast core 2 duo as the client, I'm getting a "cannot mount nbd" error
| |
| 15:59 | <Gadi> this is all in initramfs
| |
| 15:59 | <alkisg> After 2-3 seconds, the network is initialized! :D
| |
| 15:59 | So I have to use break=mount, wait 3 seconds, and then press ctrl+d to boot it...
| |
| 16:00 | udhcp is in the wrong place (init-premount)
| |
| 16:00 | It should be called directly from ltsp_nbd
| |
| 16:00 | (at configure_networking)
| |
| 16:00 | Too many things are wrong in the boot process, both for ltsp clients and for standalone lucid PCs :(
| |
| 16:00 | <Gadi> well, my problems occur when I don't use udhcpc
| |
| 16:01 | * Gadi considers ripping out nbd-proxy | |
| 16:01 | <Gadi> why do we have nbd-proxy anyway?
| |
| 16:01 | <alkisg> Gadi, what do you use? ipconfig? That one has problems on its own
| |
| 16:01 | <Gadi> yeah, ipconfig
| |
| 16:01 | right now
| |
| 16:01 | what problems exactly?
| |
| 16:02 | <alkisg> E.g. it may never get an ip lease
| |
| 16:02 | * vagrantc still uses ipconfig | |
| 16:02 | <Gadi> haven't seen that problem
| |
| 16:02 | <alkisg> What do you see?
| |
| 16:02 | <vagrantc> Gadi: nbd-proxy was implemented to better support reconnections or something like that
| |
| 16:03 | hersonls has quit IRC | |
| 16:03 | <vagrantc> servers disappearing
| |
| 16:03 | <Gadi> vagrantc: evidently to the detriment of connections ;)
| |
| 16:03 | <vagrantc> round-robin servers or something
| |
| 16:03 | * alkisg tried unplugging / repluggin the client network cable, but nbd-proxy didn't help there... | |
| 16:03 | <vagrantc> Gadi: don't blame me, i still recommend NFS :)
| |
| 16:03 | <johnny> considered trying dracut instead of intramfs-tools alkisg :)
| |
| 16:03 | <alkisg> johnny: I'm not an ubuntu core developer, I can't make such decisions
| |
| 16:04 | <johnny> sure, i meant you trying it :)
| |
| 16:04 | personally
| |
| 16:10 | alkisg has quit IRC | |
| 16:14 | slidesinger has joined #ltsp | |
| 16:17 | <Lns> nfs ftw ;)
| |
| 16:17 | * Lns waves to all | |
| 16:17 | Lns has quit IRC | |
| 16:33 | jammcq has quit IRC | |
| 16:37 | otavio has quit IRC | |
| 16:38 | Barbosa has joined #ltsp | |
| 17:00 | Barbosa has quit IRC | |
| 17:12 | bobby_C has quit IRC | |
| 17:37 | bobby_C has joined #ltsp | |
| 17:41 | tstafford has quit IRC | |
| 17:41 | tstafford has joined #ltsp | |
| 18:08 | Gadi has left #ltsp | |
| 18:17 | japerry has joined #ltsp | |
| 18:19 | vagrantc has quit IRC | |
| 18:21 | bobby_C has quit IRC | |
| 18:31 | jelly-bean has joined #ltsp | |
| 18:31 | jelly-bean has left #ltsp | |
| 19:05 | japerry has quit IRC | |
| 19:15 | epaphus has quit IRC | |
| 19:32 | japerry has joined #ltsp | |
| 19:57 | tstafford has quit IRC | |
| 20:13 | jammcq has joined #ltsp | |
| 20:22 | japerry has quit IRC | |
| 20:42 | mischko has quit IRC | |
| 21:06 | loather-work has quit IRC | |
| 21:13 | loather-work has joined #ltsp | |
| 22:07 | rjune has quit IRC | |
| 23:24 | MaRX-Mode has quit IRC | |
| 23:25 | johnny has left #ltsp | |
| 23:26 | MaRX-Mode has joined #ltsp | |
| 23:27 | johnny has joined #ltsp | |