IRC chat logs for #ltsp on irc.freenode.net (webchat)


Channel log from 19 March 2011   (all times are UTC)

00:27LoveStorm (storm@thepcrepair.info) left irc: Ping timeout: 264 seconds
00:31LoveStorm (~Storm@thepcrepair.info) joined #ltsp.
00:36LoveStorm (~Storm@thepcrepair.info) left irc: Ping timeout: 264 seconds
03:06wwx (~ww@raamat.polva.ee) left irc: Read error: Connection reset by peer
03:07wwx (~ww@raamat.polva.ee) joined #ltsp.
03:24artista-frustrad (~artista_f@187.7.151.137) left irc: Quit: Leaving
03:26cyberorg (~cyberorg@opensuse/member/Cyberorg) left irc: Remote host closed the connection
03:28cyberorg (~cyberorg@opensuse/member/Cyberorg) joined #ltsp.
03:33bobby_C (~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) joined #ltsp.
04:34drdata (~drdata@pool-173-57-189-158.dllstx.fios.verizon.net) left irc: Ping timeout: 255 seconds
04:35Faithful (~Faithful@202.189.73.144) joined #ltsp.
05:49Trixboxer (~Trixboxer@office.supportdepartment.net) joined #ltsp.
05:54andygraybeal (~andy.gray@obsidian.casanueva.com) joined #ltsp.
06:03wwx (~ww@raamat.polva.ee) left irc: Read error: Connection reset by peer
06:03wwx (~ww@raamat.polva.ee) joined #ltsp.
06:05shogunx (~shogunx@rrcs-67-79-182-228.se.biz.rr.com) left irc: Read error: Operation timed out
06:41Faithful (~Faithful@202.189.73.144) left irc: Read error: Connection reset by peer
06:43Faithful (~Faithful@202.189.73.144) joined #ltsp.
06:51alexqwesa (~alex@109.172.15.11) left irc: Read error: Connection reset by peer
07:03alexqwesa (~alex@alexo-veto.broker.freenet6.net) joined #ltsp.
07:07alexqwesa (~alex@alexo-veto.broker.freenet6.net) left irc: Remote host closed the connection
07:10alexqwesa (~alex@alexo-veto.broker.freenet6.net) joined #ltsp.
07:37cyberorg (~cyberorg@opensuse/member/Cyberorg) left irc: Ping timeout: 255 seconds
07:56drdata (~drdata@tw2.udallas.edu) joined #ltsp.
08:00drdata (~drdata@tw2.udallas.edu) left irc: Read error: Connection reset by peer
08:24MorningSon (~MorningSo@cpe-70-114-21-95.satx.res.rr.com) joined #ltsp.
08:51
<moldy>
hi
08:52
about 30% of the time, my clients hang at boot (before the ldm screen appears). they still respond to ping, but i cannot login with ssh. switching to the local console does not work. can anyone suggest a good method to debug this?
09:06Faithful (~Faithful@202.189.73.144) left irc: Ping timeout: 250 seconds
09:14cyberorg (~cyberorg@opensuse/member/Cyberorg) joined #ltsp.
09:16
<moldy>
https://bugs.launchpad.net/ltsp/+bug/589034 could be this issue, i guess
09:43mistik1 (mistik1@unaffiliated/mistik1) left irc: Read error: Connection reset by peer
09:43mistik1 (mistik1@unaffiliated/mistik1) joined #ltsp.
09:58shogunx (~shogunx@2001:4978:106:1:21e:33ff:fe47:32f5) joined #ltsp.
10:02Kicer86 (~Kicer86@host-5db0eeee.sileman.net.pl) joined #ltsp.
11:12Lns (~Lns@pdpc/supporter/professional/lns) left irc: Quit: Leaving
11:56alkisg (~alkisg@ubuntu/member/alkisg) joined #ltsp.
12:09shogunx (~shogunx@2001:4978:106:1:21e:33ff:fe47:32f5) left irc: Read error: No route to host
12:11shogunx (~shogunx@2001:4978:106:1:21e:33ff:fe47:32f5) joined #ltsp.
12:32wwx (~ww@raamat.polva.ee) left irc: Read error: Connection reset by peer
12:33wwx (~ww@raamat.polva.ee) joined #ltsp.
13:02vagrantc (~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net) joined #ltsp.
13:06Trixboxer (~Trixboxer@office.supportdepartment.net) left irc: Quit: "Achievement is not the end, its the beginning of new journey !!!"
13:13
<highvoltage>
vagrantc: are you around?
13:14
<vagrantc>
highvoltage: barely
13:15
<highvoltage>
vagrantc: what do you think of http://jonathancarter.org/files/dump/ldm_spacefun.png ?
13:16
<vagrantc>
oh-ho!
13:16roasted (~Jason@c-174-54-217-48.hsd1.pa.comcast.net) joined #ltsp.
13:16
<roasted>
What's up fellas?
13:17
<vagrantc>
highvoltage: i'll have to think of it later ... but thanks!
13:18
highvoltage: looks good from here! licensing all good? :)
13:18
<highvoltage>
vagrantc: ok, I'll paste the tarball so that you could test it directly too if you'd like. and my pleasure :)
13:18
<roasted>
I want to edit my lts.conf to force all clients to get thin client images instead of fat, just for tesing purposes here. I'm not seeing the entry in the man page. Anybody know the entry offhand?
13:18
<highvoltage>
vagrantc: roasted I got it all from http://wiki.debian.org/DebianArt/Themes/SpaceFun - I'll have to get the exact licensing/copyright. will do that next...
13:18
vagrantc: but since all that artwork is already in debian I'm sure it will be no problem
13:19
<vagrantc>
highvoltage: it's just modified variants of the other artwork? should be clean then.
13:19
<highvoltage>
(oops, don't know where that "roasted" comes from)
13:19
<roasted>
was that a mistake to tag me in that highvoltage ?
13:19
lol
13:19
<highvoltage>
roasted: yes, sorry :)
13:19
<roasted>
I was like whaat?
13:19
<highvoltage>
vagrantc: yep
13:20
vagrantc: http://jonathancarter.org/files/temp/spacefun.tar (I know you're kind of away now but there it is)
13:25
<roasted>
where's the path of lts.conf? The only one I dig up is the example.
13:27
<highvoltage>
/var/lib/tftpboot/ltsp/<arch>/lts.conf
13:27
<roasted>
does one not exist by default?
13:28
yeah, weird. I don't have one in that dir.
13:28
<highvoltage>
indeed, it doesn't exist by default
13:28
<roasted>
ah okay
13:29
is read perms with root:root okay?
13:29
<highvoltage>
yep
13:30
<vagrantc>
lts.conf on debian is usually /opt/ltsp/i386/etc/lts.conf
13:30
<roasted>
I figured it would exist and just be empty or something.
13:30
yeah I have one in that tdirectory
13:30
<vagrantc>
unless you're doing squashfs NBD stuff
13:30
<roasted>
everything is commented out though
13:30
should I use that one?
13:30
or create in var?
13:31
<highvoltage>
roasted: you don't know whether you're using debian or ubuntu? :)
13:31
<roasted>
I'm using Ubuntu
13:32
But the lts.conf file I have is empty except comments.
13:32
I wasn't sure if that was the one to use or what.
13:32
And you mentioning the /var path threw me off :P
13:32
<highvoltage>
roasted: wat does it say in the commented out section?
13:32
roasted: there might be something interesting in there
13:32
<roasted>
This sit hed efault lts.conf file for ltsp5. blah blah blah
13:32
is this the default*
13:32
<highvoltage>
roasted: it might even tell you the location of the right path if you look carefully!
13:32
<abeehc>
it doesn't end up in var/lib/tftpboot by default
13:33
<roasted>
hey
13:33Last message repeated 1 time(s).
13:33
<abeehc>
but yeah the file your looking at should say where it's supposed to go
13:33
<roasted>
so I DO have to create it in var
13:33
<abeehc>
asi recall
13:33
yeah
13:33
<roasted>
well isnt htat nifty
13:33
<highvoltage>
roasted: yes, that's what I keep telling you! and that's what the file you're looking at is telling you!
13:33
<abeehc>
no need to update image of course
13:33
yeah he's right though the file you are looking at clearly says that
13:33
<roasted>
ah really? just reboot clients?
13:33
<abeehc>
at least in the distrib i';ve used
13:33
yeah just reboot
13:36
<roasted>
testing fat vs thin on my little netbook here
13:37
I almost feel bad now... I had my boss order more ram for my server... it only had 2gb of RAM so I had him get 8 GB for the real test-drive of this.
13:38
but now that I plugged fat clients into the mix with our existing hardware, I guess we don't need that much ram.
13:38
<abeehc>
it doesn't hurt hehe
13:38
<roasted>
we won't tell him that :P
13:38
it's interesting to watch the RAM usage on the server when I'm playing with fat vs thin setups. I can definitely see the differences.
13:38
fat clients are screaming my name now.
13:38
<abeehc>
i wanna run my server on bare metal but no moneys to do so
13:39
that's good news, they are also screaming my name
13:39
<roasted>
are you looking to deploy LTSP?
13:39
<abeehc>
i've had it deployed for over a year now
13:39
<roasted>
nice
13:39
good results?
13:39
<abeehc>
but th normal thinclient setup
13:39
you know I';m not sure
13:39
<roasted>
ah you're only using thin?
13:39
<abeehc>
yes
13:39
<roasted>
what aren't yuou sure about?
13:39
<abeehc>
well i did start with local-apps but that sucked
13:40
<roasted>
thin clients w/ local apps?
13:40
<abeehc>
we've gone back and forth if it's a ton of savings versus booting to rdesktop windows terminal server
13:40
Yeah
13:40
<roasted>
I read about that a little bit, but it doesn't really appeal to me
13:40
<abeehc>
thinclients with localapps was a bit painful;
13:40
it's much better for flash and whatnot and you selectively choose the apps that the client runs
13:40
<roasted>
it kind of seems like a more controlled way to have a buffer in between fat and thin clients
13:40
<abeehc>
but in the gnome environment i struggled a bit
13:40
<roasted>
but wouldnt fat clients do the same job?
13:41
I suppose the nice thing about local apps w/ thin would be if you had "in between gear"
13:41
<abeehc>
Yeah
13:41
<roasted>
meaning not really the horsepower to run fat clients
13:41
but you wanted more beef than thin clients
13:41
<abeehc>
right yeah that's where it might be decent
13:41
my clients are decent enough to do fat, can't wait to get it out there
13:41
<roasted>
nice, nice
13:41
how many clients are you using
13:42
<abeehc>
some of my awesome users.. tend to reboot the clients when they have trouble
13:42
with fat clients it might actually make some difference hehe
13:42
about 25
13:42
<roasted>
nice
13:42
one server?
13:42
<abeehc>
maybe 18 logged in at peaks
13:42
yeah
13:42
sorta
13:42
<roasted>
what specs are on the server box?
13:42
<abeehc>
it's a vm
13:42
on a standard dl385
13:43
<roasted>
nice
13:43
<abeehc>
shared with a bunch of other servers so kind of a challenge
13:43
<roasted>
we were going to VM the server
13:43
but we decided to localize it for starters
13:43
<abeehc>
it's a good place to start
13:43
<roasted>
and considering we HAD a spare server (despite the 2gb installed RAM) it worked out
13:43
<abeehc>
I'm thinking about pushing for kvm instead of vmware for the next one i prep
13:43
<roasted>
pair of dual core 3.2 ghz procs
13:43
should work, me thinks
13:43
8gb ram and 2 gigabit nic's
13:44
<abeehc>
for sure
13:44
<roasted>
if this flies, my boss may give me the current thin client box running for another vendor
13:44
he wants to rip that lab out so bad
13:44
that server has a pair of six core procs, 32gb of ram, and 4 gigabit nics. me likey. :P
13:44
<abeehc>
hehe
13:44
3..2 is too high to not be a p4 though
13:44
no?
13:44
<roasted>
what, the server I got?
13:45
<abeehc>
the dial 3.2 is p4?
13:45
dual
13:45
<roasted>
it's AMD
13:45ogra (~ogra@ubuntu/member/ogra) left irc: Ping timeout: 248 seconds
13:45
<roasted>
pretty sure it's a 3.2
13:45
<abeehc>
bleh racing on sorry half watching
13:45
<roasted>
lol
13:45
<abeehc>
nice
13:45
<roasted>
I know the model number of the server, let me look up and make sure 3.2 was an option
13:45
yeah
13:45ogra (~ogra@p5098ed03.dip0.t-ipconnect.de) joined #ltsp.
13:45
<roasted>
3.5 was the highest option
13:46
<abeehc>
it can't be opteron then/
13:46
?
13:46
<roasted>
for intel xeon
13:46
dang maybe we're not AMD...
13:46Nick change: ogra -> Guest23759
13:46
<roasted>
I was working on 2 servers
13:46
the one with the 2gb ram It hought was AMD
13:46
maybe I had them mixed
13:46
ha yeah theres no AMD option for that box. It has to be intel.
13:47
<abeehc>
ah yeah
13:47
<roasted>
so when are you moving your gear to fat clients?
13:47
seems like an easy trasnition
13:47
run the command, grab some lunch, some back, update, blam
13:47
<abeehc>
i have a heavily modified chroot
13:47
it's gonna be a day or so and i've got no time lately
13:47
esp when the thinclients work well enough
13:48
heavily isn't the right word, but i do have to modify the new fat chroot quite a bit to keep consistency from the old server
13:48
and i'm running 9.04 right now so should be exciting
13:49
main draw really was nfs_home not sure if you =can do that with thin clients but moving io off the ltsp server seems like a rad idea
13:49
<roasted>
haha isnt 9.04 dropped on support?
13:50
ha, yup :P
13:50
dropped in october
13:52
what the... "Oneiric Ocelot"
13:52
<abeehc>
my mistake, 9.10 but it being dropped wouldn't really cause me to move faster
13:52
haha
13:52
still support xp laptops .......so........
13:52
yeah
13:52
<alkisg>
For fat clients, a Pentium III CPU would do fine. Network bandwidth, disk speed, and a little RAM to cache the nbd image in RAM make the difference though.
13:52
<roasted>
yeah. we're all on XP too.
13:52
<alkisg>
(on the server)
13:53
<roasted>
ah
13:53
you scared me alkisg
13:53
<abeehc>
i think i get the equiv of a p3 gen xeon from vmware as it is hehe
13:53
<roasted>
cause I was thinking you meant that was needed for clients when I was looking at intel atom/1gb ram clients...
13:53
<alkisg>
Nah for clients I'd go for dual cores
13:53
<roasted>
dual core atoms suffice?
13:54
<alkisg>
As I said yesterday, I'd prefer something over 1000 in that scale: http://www.cpubenchmark.net/common_cpus.html
13:54
<roasted>
I had trouble finding dual core setups that were small in physical size that would fit the bill.
13:54
<alkisg>
A single core atom is at 300, too low, don't know about dual cores
13:55
<abeehc>
atoms are quiet and low power but crazzzzzzy slow
13:55
<roasted>
I don't know man. You said a fat client acts as if it's a local install, so if it can run a natively installed OS, it should work. Ubuntu flies on this darn netbook... 1.6 atom with 1gb of RAM...
13:55
<abeehc>
cheap now i think they got abandoned?
13:55
<roasted>
abandoned? There's dual core atoms now.
13:55
<alkisg>
roasted: if you think it's fine, then fat will be fine too. Personally I'd go for something better, for new setups.
13:55
<abeehc>
i had a dual core atom like 3 years ago
13:56
painful slow
13:56
<roasted>
alkisg, I'm certainly going to try to. But it depends on what money is available for spending.
13:56
abeehc, I thought dual core atoms JUST came out...
13:56
<abeehc>
not at all
13:56
they have't release a new atom for almost two years i think
13:56
they've announced a new gen though that could be interesting
13:57
<roasted>
alkisg, I've certainly noticed a big difference on server load with my little test hing here, and I'm only using my dual core laptop as a server and my netbook as a client. I can see the benefits if the clients have decent local processing power.
13:57
<alkisg>
Sure atoms work fine for normal use. But open 2-3 heavy flash apps and they crawl.
13:57
<abeehc>
yeah the problem with the current atoms on the market is arm procs are as fast in a lot of cases
13:57
<roasted>
is arm welel supported with linux?
13:58
<abeehc>
not at all but when intel launched the atoms as i recall
13:58
it was to get into that type of market and i think they missed
13:58
anyway far off topic ; i'd stil be happy with atoms of any sort for clients
13:58
<roasted>
I certainly have some homework to do when I begin to shop for clients.
13:58
<abeehc>
instead of my p4s
13:58
<roasted>
Right now, there's no money. However all of our labs are 400 watt systems. If we could justify spending for trading off energy savings, we might get it to replace the gear.
13:59
<abeehc>
yeah i have the same situation
13:59
<roasted>
But this is also assuming I can get a thin client box that uses 70w or less.
13:59
<abeehc>
and we did the math and power is way to cheap where i live
13:59
<roasted>
Which, all of the atom choices I looked at use like 40w.
13:59
<abeehc>
yeah
13:59
<roasted>
A lab of 30 systems would use nearly 12,000 watts.
13:59
A lab of 30 systems with ASUS EEE Box 1007 thin clients (atom, 1gb ram) would use less than 2,000 watts.
13:59
that's including the server energy usage too.
13:59
Times that by 25 labs, and 9 months of runtime.
14:00
<abeehc>
i tihnk it came to like 25$ a month or somthing for us
14:00
it was dissapointing
14:00
absolutely worth going over though
14:00
<roasted>
Even if we keep our existing gear and implement LTSP, we save 0 power, but we would make the hardware last longer. It's BAD right now. That gear runs so slow.
14:00
Either way, with LTSP... we win.
14:01
It's a question of how much do we want to spend today/save tomorrow to determine how far we take it.
14:01
I at least want to see thin client boxes to replace our 2 ncomputing labs. that system is a total piece of garbage.
14:04
abeehc, were you running windows clients before you dropped LTSP in place last year?
14:05
<abeehc>
no, new building so new setup
14:06
<roasted>
nice, nice
14:06
yeah we gotta cross that bridge of ditch windows/hi linux to our labs
14:06
<abeehc>
every user still needs a ts license so it kinda sucks
14:06
<roasted>
but a ton of our software is web based anyway and already cross platform (gimp, libre office, etc)
14:06
TS?
14:06
<abeehc>
ms terminalservices
14:06
<roasted>
ahh
14:07
why?
14:07
if you're on LTSP with your own gear
14:07
<abeehc>
crappy legacy software
14:07
<roasted>
are you running MS products somehow?
14:07
ah
14:07
:(
14:07
<abeehc>
core business app runs on windows hehe
14:07
maybe one day web based but it's not even ours
14:07
so my clients use rdesktop a decent amount
14:07* alkisg was able to get a lot of windows edu apps running on wine
14:07
<abeehc>
which has had it's own problems
14:07Kicer86 (~Kicer86@host-5db0eeee.sileman.net.pl) left irc: Quit: KVIrc Insomnia 4.0.2, revision: 4740, sources date: 20100627, built on: 2010-08-08 18:29:00 UTC http://www.kvirc.net/
14:07
<alkisg>
So no rdesktop here
14:08
<abeehc>
yeah i wish
14:08
try lotus approach from 1997
14:08
barely runs on windows to stat with haha
14:08
<roasted>
especially now with budget cuts, it's hard to justify paying for software or continuing to pay for software when there's free alternatives.
14:08
<abeehc>
scratch that, i have tried
14:08
<alkisg>
We even maded about 10 Gb of .deb packages out of windows programs, and even put them in a repository
14:08
<roasted>
I'm giving a presentation this week about linux and how we should adopt it.
14:11
<abeehc>
good luck i'm sure it'll go well
14:11
unless you use powerpoint
14:23
<roasted>
nope
14:23
using libreoffice presentation, to further make my point :P
14:27
<abeehc>
hehe :)
14:27
<mistik1>
sweet
14:27
keep it up
14:28* mistik1 misses LTSP
14:33
<roasted>
what's there to miss? Do you not run it anymore?
14:43MorningSon (~MorningSo@cpe-70-114-21-95.satx.res.rr.com) left irc: Quit: WeeChat 0.3.0
14:45alkisg (~alkisg@ubuntu/member/alkisg) left irc: Quit: Leaving.
14:57patrickmw (~pwright@ip68-231-4-153.ph.ph.cox.net) left irc: Ping timeout: 248 seconds
15:02mordocai (~mordocai@66.119.9.243) joined #ltsp.
15:09patrickmw (~pwright@ip68-231-4-153.ph.ph.cox.net) joined #ltsp.
15:12mordocai (~mordocai@66.119.9.243) left irc: Quit: rebooting
15:20MorningSon (~MorningSo@cpe-70-114-21-95.satx.res.rr.com) joined #ltsp.
15:41Gremble (~Ben@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com) joined #ltsp.
15:49irule (~irule@187.140.130.241) joined #ltsp.
15:49LoveStorm (storm@thepcrepair.info) joined #ltsp.
16:08Guest23759 (~ogra@p5098ed03.dip0.t-ipconnect.de) left irc: Changing host
16:08Guest23759 (~ogra@ubuntu/member/ogra) joined #ltsp.
16:08mordocai (~mordocai@66.119.9.243) joined #ltsp.
16:12vagrantc (~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net) left irc: Ping timeout: 246 seconds
16:16bobby_C (~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) left irc: Quit: Goin' down hard
16:20mordocai (~mordocai@66.119.9.243) left irc: Quit: restarting x server
16:21mordocai (~mordocai@66.119.9.243) joined #ltsp.
16:25mordocai (~mordocai@66.119.9.243) left irc: Client Quit
16:26mordocai (~mordocai@66.119.9.243) joined #ltsp.
16:56Gremble (~Ben@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com) left irc: Quit: I Leave
17:42mordocai (~mordocai@66.119.9.243) left irc: Quit: Leaving
17:58F-GT (~phantom@ppp121-44-126-79.lns20.syd6.internode.on.net) left irc: Remote host closed the connection
18:28F-GT (~phantom@ppp121-45-169-226.lns20.syd6.internode.on.net) joined #ltsp.
18:30cyberorg (~cyberorg@opensuse/member/Cyberorg) left irc: Ping timeout: 255 seconds
19:07irule (~irule@187.140.130.241) left irc: Ping timeout: 250 seconds
19:07cyberorg (~cyberorg@opensuse/member/Cyberorg) joined #ltsp.
21:21
<roasted>
If you're running fat clients, is there no way to share processing from the server? AKA - If you have a massive amount of processing power on the server, yet you're running fat clients, is there no way to make the clients use server processing power as a fallback if their local processor begins to max out?
22:00alkisg (~alkisg@ubuntu/member/alkisg) joined #ltsp.
22:04
<alkisg>
roasted: not "as a fallback", but you can start remote apps with the ltsp-remoteapps command
22:04
<roasted>
alkisg, ehh, doesn't sound like the most user friendly solution for end users.
22:04
I was just curious if that was possible.
22:04
<alkisg>
I.e. it doesn't happen automatically, you need to manage which apps you want remotely by replacing the entries in gnome menus
22:04
<roasted>
Mostly because the server I may be getting is completely crazy with processing power. I'd like to utilize it if I could.
22:04
<alkisg>
The users won't notice a difference if you modify the menus
22:05
The question is, "which apps need cpu power but not much UI"
22:05
<roasted>
alkisg, I'm just concerned about the atom thing. The only box I can find that seems worthy of the price/power savings are atom boxes, and I'd like to get something with some more processing power.
22:20
<alkisg>
For the normal desktop user, all I can think of is "dvd rippers" :D
22:26
roasted: in greece we don't use thin client, we reuse existing hardware instead, but have you looked at this one? It seems very powerful: http://www.wyse.com/products/hardware/thinclients/Z90/
22:28
<abeehc>
we recycle too
22:28
those are nice i wish
22:33roasted__ (ae36d930@gateway/web/freenode/ip.174.54.217.48) joined #ltsp.
22:33
<roasted__>
Is running localapps with thin clients really just a matter of firing up xterm and ltsp-localapps firefox, and update the image?
22:34
I thought I remember hearing localapps was a headache to set up......
22:36
<alkisg>
Does your image contain firefox?
22:37
<roasted__>
well not yet. Im just doing some reading about it.
22:37
<alkisg>
Example. You can set LTSP_FATCLIENT=False in lts.conf, and use your *fat* chroot to boot a thin client
22:37
<roasted__>
yeah I saw the lts.conf part.
22:37
<alkisg>
Nope
22:37
<roasted__>
I was just trying to read how to specify which apps are local
22:38
<alkisg>
Read what I'm writing
22:38
That's not in a wiki page
22:38
So since your fat image has all the apps, you can test localapps by booting it as a thin client
22:38
<roasted__>
oh yeah
22:38
I follow
22:39
<alkisg>
So you don't need to do anything to set localapps up. Since you already have a fat chroot, all you need to do is put a list in lts.conf
22:39
<roasted__>
I meant in the event I was using thin clients entirely
22:39
no fat
22:39
thin and local
22:40
<alkisg>
It's the same, you just put less apps in the chroot
22:40
I.e. you install them manually
22:40
<roasted__>
yeah. that's the key I'm trying to understand.
22:40
<alkisg>
sudo chroot /opt/ltsp/i386 apt-get install firefox
22:40
<abeehc>
i like the idea of fat chroot +localapps
22:40
<alkisg>
stuff like that
22:40
<abeehc>
one thing i learned is you also want firefox-gnome-support if your using gnome
22:41
<alkisg>
With a fat chroot you also get the proper themes without having a headache :)
22:41
<abeehc>
or it sucks
22:41
<roasted__>
alkisg: ahh. the example I read had xterm as sudo chroot /opt/ltspi386 apt-get install xterm. I didnt realize it was referring to xterm as an app.
22:41
<abeehc>
yeah
22:41
<roasted__>
alkisg: hm, really?
22:41
alkisg: are you recommending fat chroot, setting lts.conf to FATCLIENT=False (so everything boots thin), and then set local apps?
22:41
<alkisg>
The only reason to *not* use a fat chroot is ltsp-build-image and ltsp-update-image time, nothing else
22:42
<abeehc>
interesting
22:42
<roasted__>
alkisg: not sure I got what you meant there. what difference is there from fat/thin with build-image and update-image?
22:42
<alkisg>
Thin image = 250 mb, fat image = 2 gb
22:43
==> more time to generate the image
22:43
<roasted__>
ha. thats why I ran out of room yesterday.
22:43
ah
22:43
if that's it, I'm not worried about that
22:43
updating it isnt THAT much longer though, is it?
22:43
<alkisg>
Update like, in sudo chroot /opt/ltsp/i386 apt-get dist-upgrade?
22:43
<roasted__>
alkisg: when you say more time to generate an image, you're talking to BUILD an image, not transfer it over the network, right?
22:44
<alkisg>
I'm talking about ltsp-build-client and ltsp-update-image
22:44
Not nbd exporting, no difference there
22:44
<roasted__>
I meant update as in update-image
22:44
ltsp-update-image
22:44
<alkisg>
ltsp-update-image is the compression I wrote above
22:44
Compressing 2 gb takes more time than compressing 200 mb
22:45
<roasted__>
gotcha
22:45
I toyed with thin, I toyed with fat. Now I want to try thin with local apps and see which one of the 3 scenarios fits what Ineed.
22:45
<alkisg>
I'd only use localapps for clients with 300-500 mb ram, to have some specific apps run faster, OR in specific environments (call centers etc)
22:45
<roasted__>
alkisg: so if I had 512 clients, youd just go fat?
22:46
<alkisg>
Yup
22:46
<roasted__>
alkisg: does client cpu weigh in at all when youre deciding fat vs thin?
22:46
or do you let that decision rely on ram?
22:46
<alkisg>
But that's just my personal opinion, not everyone here shares it
22:47
I don't look at cpu when deciding thin vs fat. Even an atom with 1 ghz cpu / 1gb ram runs better as fat than as thin, imho
22:47
But if I were to buy hardware, then I'd look at cpu too :)
22:47
<roasted__>
good deal
22:47
yeah I'm still on the hunt for more options
22:48
I have time to hunt around but, ya know.
22:48
when you're dealing with school district funds, that's tax payer money. You gotta be smart about purchases. It's not a free for all spending spree.
22:48
<alkisg>
For clients with about 400 ram, maybe I'd be using debian/lxde in the chroot instead of gnome
22:49
i.e. again fat, but with a lighter de
22:49
<roasted__>
to get lxde in the chroot youd just add a tag to it when building, right?
22:49
I thought I remember seeing that in the wiki
22:49
<alkisg>
The wiki mentions ubuntu fat clients, so you'd get lubuntu there, not debian
22:50
A small test I did showed that lubuntu needs a *lot* more ram than debian/lxde
22:50
<roasted__>
right. even still, I remember seeing kubuntu or something in it.
22:50
<alkisg>
So I wouldn't choose it
22:50
<roasted__>
I'll be safe and choose gnome.
22:50
How has unity held up? Have you used it?
22:51
<alkisg>
I've seen it. I'll be sticking with 10.04 LTS for the next 2 years, and I'll see it again then, to decide if it's good enough to stay with ubuntu or to switch to debian
22:52
I tried to use it for everyday use, but I couldn't, too buggy
22:53
<roasted__>
I only tried unity with mutter, which sucks. unity with compiz (11.04) sounds super nice though in comparison.
22:53
I was curious how unity held up wiht LTSP, resource wise or if it added any headaches, issues, lag, etc.
22:54
<alkisg>
For fat clients it makes no difference than locally
22:54
<roasted__>
ahh
22:54
true
22:54
<alkisg>
For thin clients, it requires compiz etc
22:54
<roasted__>
I keep forgetting that
22:54
is compiz a bear with thin clients?
22:54
<alkisg>
Not all clients have that
22:55
It works fine with intels, but here we turn it off by default because for others it's causing too many problems
22:55
<roasted__>
yeah, good deal
22:55
I saw a video on youtube with a guy running it and he had wobbly windows, etc. But he never specified if he was using fat or thin clients
22:56
<alkisg>
Thin can do that too
22:56
The effects don't matter much, it's the local rendering that makes the difference
22:57
If I had clients good enough to support compiz, I'd use fat. So I don't give it much thought :D
22:57
(the thin+compiz combination)
22:58
<roasted__>
yeah. it sounds like thin clients are really the rarity.
22:59
now that I see the benefit of fat clients I find it hard to believe I would ever NEED dto use thin clients
22:59
unless I was running hardware that was *that* slow...
22:59
maybe I'm wrong... just my amateur opinion though, based off what little I know
22:59
<alkisg>
There are companies that don't need fat, based on the apps they run
23:00
But for schools sure I agree with you
23:00
<roasted__>
yeah
23:00
I've seen thin clients in use in shopping malls
23:00
POS terminals, etc.
23:00
but they also have DOS based text interfaces with no mouse control. :P
23:00
<alkisg>
Hehe
23:01
<roasted__>
far from wobbly windows and spinning cubes we 2011 linux users are aware of
23:01
Have you personally done deployments for districts?
23:01
or anybody for tha tmatter
23:02
<alkisg>
I've deployed ltsp in about 10 schools personally, and remote helped others to deploy it in about 100 more
23:03
<roasted__>
have those schools used existing hardware for clients?
23:03
<alkisg>
Yes
23:03
<roasted__>
nobody bought new gear?
23:03
<alkisg>
In a few cases they did, but not thin clients, just regular desktops
23:03
<roasted__>
ah
23:03
<alkisg>
They were just careful to get desktops that run with linux
23:03
<roasted__>
yeah I'm not sure what to recommend my boss in our case.
23:04
I hate to get desktops, because we miss out on energy savings. I hate to get thin clients running atoms, because if Im running fat clients, that's local processing.
23:04
anything that can't run on an atom thereby fails.
23:04
making the gear limited, even if its new/still young
23:04
<alkisg>
what can't run on atoms?
23:04
OK, you wouldn't run blender on it
23:04
<roasted__>
I'm talking down the road.
23:04
<alkisg>
Right
23:04
<roasted__>
If I buy clients, I want them to outlast desktops.
23:05
<alkisg>
That's why I was suggesting something more powerful to you :)
23:05
For down the road
23:05
<roasted__>
yeah
23:05
if I could get that ASUS box with a lower end AMD dual core, SOLD.
23:05
But it's like... once I get to the ASUS, the next step up (that I can find) IS a desktop
23:06
<alkisg>
Yeah I get your dilemma
23:07
<roasted__>
The catch is, there's no money. But we can probably get money from the business department if we can justify the energy savings.
23:07otavio (~otavio@debian/developer/otavio) left irc: Read error: Operation timed out
23:07
<roasted__>
But if we take that savings, we get atom boxes.
23:07
see, we're going in circles here :P
23:07
<alkisg>
That wyse above says it only needs 13 watts
23:07
but I bet it's more expensive than a desktop
23:08
<roasted__>
the wyse?
23:08
<alkisg>
The link I pasted above for you
23:08
(07:26:08 AM) alkisg: roasted: in greece we don't use thin client, we reuse existing hardware instead, but have you looked at this one? It seems very powerful: http://www.wyse.com/products/hardware/thinclients/Z90/
23:08
<roasted__>
ah, that's probably on my desktop
23:08
Im on my laptop now
23:09
lol. windows thin clients.
23:09
<alkisg>
I just pasted the first non-atom thin client i found :D
23:09
I.e. powerful thin clients that save energy do exist, haven't googled any more than that though
23:10
<roasted__>
I never even heard of wyse...
23:11
<alkisg>
I'd like to see more powerful (but cheaper) thin clients in the shops... I think they're the right solutions for schools.
23:11
<roasted__>
yeah that box runs about 500
23:11
1.5ghz single core or 1.6ghz dual core procs available
23:12
I can only imagine the single core was on the boxes I saw listing @ 500
23:12otavio (~otavio@189.114.111.55) joined #ltsp.
23:12otavio (~otavio@189.114.111.55) left irc: Changing host
23:12otavio (~otavio@debian/developer/otavio) joined #ltsp.
23:12
<roasted__>
at that point, my boss would probably just get desktops. :(
23:12
it's okay, I just have some homework to do.
23:12
<alkisg>
Right. But if there was more demand for "powerful thin clients", I bet prices would drop
23:12
<roasted__>
we do a lot of work with CDW. They may know of some other vendors we can buy through that they don't even have on their site.
23:12
Just never know.
23:13
I hear ya.
23:13
I really think schools are the places with that demand.
23:13
and
23:13
school districts are a small niche in the big world of technology
23:13
but hey, worst case scenario, we have atoms here. :P
23:13
<alkisg>
Yup :)
23:13
<roasted__>
either way, an atom with ubuntu is wickedly faster than xp on the boxes we have in there.
23:14
and THAT is still a benefit
23:14
where we have these clients going isnt CADD or photoshop labs or anything
23:14
just firefox/libre office intensive labs
23:14
me thinks the atom will shine nicely with that. I'm just trying to look forward.
23:14
some people buy with intentions of solving problems today. I like to make sure what I buy will last the longest.
23:15
get me a pair of six core procs in that new server. Do I need it? Nope. WIll I? No idea. Why buy it? Well, why not?
23:15
<alkisg>
You can use your server CPU/RAM to offer windows/photoshop etc VMs for your fat clients
23:15
<roasted__>
Well we won't be mixing and matching any labs.
23:15
Photoshop stays in our Photoshop lab.
23:16
Due to licensing and us not wanting to purchase fat licenses to put PS anywhere.
23:16
Things like CADD and Photoshop probably won't ever change. We'll probably always have those 60 dedicated boxes for PS and CADD.
23:17
But everything else fits the bill for fat client usage. even at the atom level. like I said, it still kills what we have now. We have 1 or 2 labs with dual core/3gb ram systems on XP Pro, and Ubuntu on my netbook over fat client is still faster than the dual core/3gb RAM XP boxes we have.
23:17
It's crazy how efficient it is.
23:19* alkisg goes to make a libreoffice presentation for ubuntu / ltsp / sch-scripts for local teachers here... in 2 weeks we have a seminar and I want to show them the benefits of that solution live, bootting a lab off of 5 external disks with ubuntu/ltsp on them
23:19
<roasted__>
enjoy. It's about time I Get to bed.
23:20
as always, alkisg... appreciate your help and advice.
23:20
<alkisg>
You're welcome, good night
23:20
<roasted__>
take care
23:20roasted__ (ae36d930@gateway/web/freenode/ip.174.54.217.48) left irc: Quit: Page closed
23:28alkisg (~alkisg@ubuntu/member/alkisg) left irc: Quit: Leaving.
23:29alkisg (~alkisg@79.167.4.153) joined #ltsp.
23:29alkisg (~alkisg@79.167.4.153) left irc: Changing host
23:29alkisg (~alkisg@ubuntu/member/alkisg) joined #ltsp.
23:50chupacabra (~chupacabr@cpe-70-112-10-77.austin.res.rr.com) left irc: Ping timeout: 240 seconds
23:51chupacabra (~chupacabr@cpe-70-112-10-77.austin.res.rr.com) joined #ltsp.
23:53MorningSon (~MorningSo@cpe-70-114-21-95.satx.res.rr.com) left irc: Quit: WeeChat 0.3.0
00:00--- Sun Mar 20 2011