|00:01||FrozenZia has left IRC (FrozenZiaemail@example.com, Read error: Operation timed out)|
|00:07||FrozenZia has joined IRC (FrozenZiafirstname.lastname@example.org)|
I think PAE is now a kernel variant, like the SMP one
|00:30||Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 245 seconds)|
|00:48||FrozenZia has left IRC (FrozenZiaemail@example.com, Read error: Operation timed out)|
|00:54||jammcq has joined IRC (firstname.lastname@example.org)|
|00:54||FrozenZia has joined IRC (FrozenZiaemail@example.com)|
|00:57||shogunx has left IRC (shogunx!~shogunx@2001:4978:106:1:5bf:bc9:6a1c:fdd0, Ping timeout: 245 seconds)|
|00:59||shogunx has joined IRC (shogunx!~shogunx@2001:4978:106:1:2891:229a:a6a4:512d)|
|01:04||FrozenZia has left IRC (FrozenZiafirstname.lastname@example.org, Read error: Operation timed out)|
|01:10||FrozenZia has joined IRC (FrozenZiaemail@example.com)|
Enslaver: no, it's all EL6 kernels
Enslaver: but the userspace doesn't require it
With pae is 0 though in this spec, ill look into it tho
Wow. This Atom thin client uses 6.2 watts while OFF.
That's pretty bad.
how do I get into the BIOS menu of the disklessworkstation 1520?
damn, this thing can't boot x86-64
|01:36||FrozenZia has left IRC (FrozenZiafirstname.lastname@example.org, Ping timeout: 264 seconds)|
|01:36||FrozenZia has joined IRC (FrozenZiaemail@example.com)|
|01:40||adrianorg_ has left IRC (firstname.lastname@example.org, Ping timeout: 264 seconds)|
|01:44||FrozenZia has left IRC (FrozenZiaemail@example.com, Ping timeout: 258 seconds)|
|01:49||FrozenZia has joined IRC (FrozenZiafirstname.lastname@example.org)|
|01:55||adriansnyman has joined IRC (email@example.com)|
|01:55||MadX_ZA has left IRC (MadX_ZAfirstname.lastname@example.org, Ping timeout: 264 seconds)|
|02:23||andygraybeal_ has left IRC (email@example.com, Ping timeout: 272 seconds)|
|02:58||adriansnyman has left IRC (firstname.lastname@example.org, Ping timeout: 256 seconds)|
|02:58||adriansnyman has joined IRC (email@example.com)|
|03:38||jammcq has left IRC (firstname.lastname@example.org, Quit: leaving)|
|04:30||dead_inside has joined IRC (email@example.com)|
|04:47||dead_inside has left IRC (firstname.lastname@example.org, Quit: Computer has gone to sleep.)|
|05:28||alexqwesa has left IRC (email@example.com, Quit: Хана X'ам !!!)|
|05:28||alexqwesa has joined IRC (firstname.lastname@example.org)|
|05:52||alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)|
|06:12||alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)|
warren: Not even the 1720's can boot 64 bit
|06:27||bobby_C has joined IRC (bobby_Cemail@example.com)|
Enslaver: probably just a BIOS limitation
Thats my excuse to women
Ingore me, its my bad and im dru k
|06:35||bobby_C has left IRC (bobby_Cfirstname.lastname@example.org, Quit: Goin' down hard)|
well it was, now its over
|07:33||alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)|
warren, actually, the 1720's processor doesn't support long mode
Enslaver, happy belated birthday! :-)
Happy birthday from me too... unless you too are under 30 like Hyperbyte so I'm not speaking to you either :P
alkisg, pff, you grumphy old man.
|08:01||* alkisg postpones speaking to Hyperbyte for not only when he's above 30, but for when he has 3+ kids as well :P|
Anyways... any updates/results from the hackfest?
I couldn't be around much due to family issues... was there any progress wrt ltsp6?
"family issues"... isn't that statement a bit redundant? :P
|08:19||adriansnyman is now known as MadX_ZA|
|09:06||zamba has left IRC (email@example.com, Ping timeout: 245 seconds)|
|09:19||bobby_C has joined IRC (bobby_Cfirstname.lastname@example.org)|
|09:19||gvy has left IRC (gvy!~mike@altlinux/developer/mike, Ping timeout: 252 seconds)|
|09:20||gvy has joined IRC (gvy!~mike@altlinux/developer/mike)|
|09:24||ltspuser_77 has joined IRC (ltspuser_77!56944317@gateway/web/freenode/ip.220.127.116.11)|
|09:25||Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)|
|09:29||dobber_ has joined IRC (email@example.com)|
|09:52||Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 260 seconds)|
|10:01||Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com)|
|10:05||Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)|
|10:33||gentgeen__ has left IRC (firstname.lastname@example.org, Read error: No route to host)|
|10:39||gentgeen__ has joined IRC (email@example.com)|
|10:47||F-GT has left IRC (F-GTfirstname.lastname@example.org, Remote host closed the connection)|
|10:49||gvy has left IRC (gvy!~mike@altlinux/developer/mike, Ping timeout: 256 seconds)|
|10:49||gvy has joined IRC (gvy!~mike@altlinux/developer/mike)|
|10:51||F-GT has joined IRC (F-GTemail@example.com)|
|11:05||Gremble has left IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com, Quit: I Leave)|
|11:48||zamba has joined IRC (firstname.lastname@example.org)|
|12:00||alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 256 seconds)|
Hi all -- suggestions for how to start solving: I have a very old crt, can't find spex for it, and w/no changes to lts.conf it boots up with double screens side-by-side...
xrandr=800x600 seemed to be a step in the wrong direction.
|12:22||pi_monk has joined IRC (email@example.com)|
|12:27||MadX_ZA has left IRC (MadX_ZAfirstname.lastname@example.org, Read error: Connection reset by peer)|
FrozenZia, that's not how lts.conf XRANDR_* variables work.
See the manpage.
|12:29||MadX_ZA has joined IRC (MadX_ZAemail@example.com)|
|12:52||mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy)|
|12:52||MadX_ZA has left IRC (MadX_ZAfirstname.lastname@example.org, Read error: Connection reset by peer)|
|12:54||MadX_ZA has joined IRC (MadX_ZAemail@example.com)|
|13:13||dsugar100 has joined IRC (firstname.lastname@example.org)|
|13:16||adrianorg_ has joined IRC (email@example.com)|
Hyperbyte: okay, perhaps what I should have said was that XRANDR_MODE_0=800x600 seemed a step in the wrong direction. Or were you refering to something besides my syntax?
As I only have 1 client currently, I was just trying things out, and the default lts.conf included that parameter for forcing all clients to a specific reso.
No, I was referring to the syntax
So why is this a step in the wrong direction?
That's not a very clear description of your problem.
|13:37||jammcq has joined IRC (firstname.lastname@example.org)|
bom dia meu amigos
Hyperbyte: the screen went into a difficult-to-describe-exactly MESS
I don't have it on right now, but I'm thinking it was sort of like wavy diagonal lines all over, and very hazy. At least with the double screens the image was rather clear (if next to impossible to read b/c of the small size)
I guess the question sort of becomes: is there a reasonably straightforward/easy way to determine a monitor's spex if they're not retreivable from the all-knowing internet?
FrozenZia, every CRT I know should support 800x600. If not, try 640x480 or 1024x768
If all of those don't work, toss it out.
|14:02||bobby_C has left IRC (bobby_Cemail@example.com, Ping timeout: 245 seconds)|
|14:03||pi_monk has left IRC (firstname.lastname@example.org, Quit: pi_monk)|
|14:28||dead_inside has joined IRC (email@example.com)|
So, after doing some reading on the weekend, I've come to a conclusion:
Our primary problem is that udisks doesn't support remote filesystems. It only looks at local entries in the /dev file.
|14:34||staffencasa has joined IRC (firstname.lastname@example.org)|
So here's what I'm proposing:
When a device is plugged, we have udev mounting work as per-usual on the thin client.
We have a listener running as the user on the thin client, watching the udev bus.
When it sees the mount, we set a xprop.
running on the server, we have another program, watching for updates to the xprop
when we see the change, we mount the ltspfs. We do it in such a way that we get the unmount option.
Not quite sure how to do that yet.
why an unmount option ?
just implement a fake :)
Hyperbyte: Ty :)
What I have to do is read a good tutorial on udev :(
Sorry, not udev
dbus I mean.
well, and policykit, consolekit/logind
dbus is just the messenger ... fi the message reaches you is handled by the other two
i guess you will have to cheat a lot here to make it think you are a local user
(and that polkit and consolekt are replaced by systemd crap doesnt really help either if you want to implement it properly)
(but nearly everything is re-implemented in systemd nowadays ... )
Im hungover :(
too much st. patricks day ?
yah, i turned 33 also yesterday
so double fun
oh, congrats young padawan !
tyty I feel more mature saying my age now, plus its symetrical
btw everyone, pushed out the new stable rpm's yesterday to the repo, clover has launched =)
yay awesome congrats wow your a genius and handsome
|15:00||laurense_ has joined IRC (email@example.com)|
i am looking into ltsp as a way to create 4-6 simple workstations, and i am wondering what the impact is when i'm using older (slower) thin clients?
laurense_: Well, so long as you use them as true thin clients, apart from viewing youtube videos being slow, they should be fine.
All my thin clients areound here are only 1ghz
Enslaver, hahah, congrats on the launch. :)
then i am wondering why thin-clients are getting faster
is it for the built in browser?
laurense_, you mean why they produce clients that have, say, 1.5ghz processors instead of 500mhz?
I think it's mostly just natural evolving of hardware. Companies stop producing slower processors because there's no demand.
and i am looking into buying some second hand thin clients for this project, and i am wondering if i can get away with buying the cheapest one possible
Sure for thin clients, -maybe-, but if it's just as cheap to put a 1.5ghz in, why bother.
because the only thing it needs to do is recieve framebuffer information from the network and send it to the screen?
laurense_, hah, you should talk to alkisg. I believe he still has installations with 386 machines with 64mb ram.
laurense_, actually, no, depending on what you want, thin clients do a lot more than that. They also work with USB disks and send those over the network... network audio consumes quite a bit too if you need that, there's local printer support, possible encryption
Plus LTSP allows you to, as you pointed out, run browsers locally on the clients. This can be a huge relief for your server.
I have on average 10 users on my terminalserver, which has a six-core AMD Phenom II processor, clocked at 3.3ghz
My server is an AMD Phenom II X4 clocked at 3.2 ghz, i would think it is enought for 6 users
When everybody is using Firefox, with flash plugin active (for advertisements or whatever) the Firefoxes can really give the CPU lots to do.
Actually, yes, that should do.
i am also looking into multiseat systems
Hm, I thought Ubuntu no longer supported multiseat.
but with the HP T5710 Clocked at 800mhz for 9 euros i think oltsp is much more interesting
Well if it costs 9 euros
Why not buy one and try?
Edubuntu has a live CD which allows you set up an LTSP server in a live environment, without installing.
Although you might as well just install it on a machine somewhere, would give you valuable experience in the process. ;-)
Hyperbyte: lightdm is capabale of doing multiseat, but it isn't documented nicely, and you need to couple the usb keyboard and mouse to a screen in a config file
which make it static
which is nasty to administer
LTSP administration is quite easy. :)
As long as your server works, the clients work. :)
and adding more workplaces is as much as buyin an extra thin client?
And having enough server capacity.
How much RAM does your server have?
4GB, but easily upgraded to 8 or 12GB
The more the better.
memory isn't that expensive anymore
I would definitely go with 12GB
i am going to look into ltsp when i get home from work
Let's say your server itself uses 500 MB, that leaves 3500 MB to share between all users... so about 600 MB per user
If you plan to run Firefox as a server application, then that is definitely not enough. Browsers consume RAM like fresh baked cookies...
i would thing 600mb is enough for simple office work, but firefox is nasty
Browsers are problematic as server applications. YouTube doesn't work fullscreen, because your network probably won't handle it and your server needs enough memory to fit all the browsers in there.
If you don't care about those two (I don't), then running them as servers applications is definitely possible.
it's a shame it will be the primary use of the thin clients
don't mind at all about youtube though
Well, same here. Our local CRM is web-based... 80% of what people do happens in their browser.
would be nice if we could learn firefox to be much more aggressive in memory management
cleaning up memory from invisible tabs, and loading again when opening a tab
Works great though as a server app, just not with flash. But flash == games && movies, generally speaking, so.... not much serious use for those in an office environment. Plus they work, just not fullscreen.
going to give ltsp a try when i get home
If you need any help, be sure to ask here.
i remember i have 2 thin-clients at home
built in a screen
but upgrade with compact-flash-cards and a local linux install
or else i surely have an old computer that can do pxe boot for initial testing
thanks for all the information so far
|15:35||bobby_C has joined IRC (bobby_Cfirstname.lastname@example.org)|
|15:43||alexqwesa has left IRC (email@example.com, Quit: Хана X'ам !!!)|
|15:43||alexqwesa has joined IRC (firstname.lastname@example.org)|
|15:56||bobby_C has left IRC (bobby_Cemail@example.com, Ping timeout: 248 seconds)|
|16:04||pi_monk has joined IRC (firstname.lastname@example.org)|
|16:11||alexqwesa has left IRC (email@example.com, Quit: Хана X'ам !!!)|
|16:11||alexqwesa has joined IRC (firstname.lastname@example.org)|
|16:12||pi_monk has left IRC (email@example.com)|
|16:32||Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 248 seconds)|
|17:08||45PAABAYS <45PAABAYS!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net> has joined #ltsp|
|17:09||vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net)|
|17:09||vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)|
|17:29||alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)|
|17:32||komunista has joined IRC (firstname.lastname@example.org)|
|17:43||sbalneav has left IRC (email@example.com, Quit: Changing server)|
|17:48||sbalneav has joined IRC (firstname.lastname@example.org)|
|18:10||highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Ping timeout: 256 seconds)|
|18:10||alexqwesa has left IRC (email@example.com, Quit: Хана X'ам !!!)|
|18:10||alexqwesa has joined IRC (firstname.lastname@example.org)|
|18:43||Parker955_Away is now known as Parker955|
|18:54||bobby_C has joined IRC (bobby_Cemail@example.com)|
sbalneav: wrt "14:36 When it sees the mount, we set a xprop. 14:37 running on the server, we have another program, watching for updates to the xprop"
==> why not just call it through the ssh socket?
"We do it in such a way that we get the unmount option. 14:38 Not quite sure how to do that yet."
==> we need an unmount "helper" in the mount command line
That needs to be suid root
There was a discussion to put that upstream in fuse, but it wasn't ever merged: http://sourceforge.net/mailarchive/forum.php?thread_name=1253960851.6399.43.camel%40localhost.localdomain&forum_name=fuse-devel
|19:12||Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)|
|19:17||prestain has left IRC (prestain!4a4d0577@gateway/web/freenode/ip.18.104.22.168, Ping timeout: 245 seconds)|
|19:22||alexqwesa has left IRC (firstname.lastname@example.org, Read error: Operation timed out)|
alkisg: Because there'll need to be 2 way communications.
Remember, udisks is going to automount on the thin client. We need a way when we disconnect from the ltspfs mount, to also unmount the device from the local (thin client) mount.
sbalneav: udisks doesn't automount on thin clients unless there's an "active user" logged in
Do we want to fake that, to allow automounting?
sbalneav: also, `ssh -S socket do-the-mount` -> can't we just make that return when it's unmounted?
alkisg: I'm looking into what's required to make the libpam-sshauth logged in user "active" now.
Well, ideally, what I'd like to do is connect the thin client dbus and remote dbus.
How would you have 2 system dbus'es on the server?
(actually, many more than 2...)
Not sure. Reading about it now :)
I guess the idea is to have the session dbus on the server communicate with the system dbus on the client,
but that'll be too messy, it'll lose all server-side services then
Btw, it'd be nice if the mount points were the same on the client and on the server
So that localapps and remoteapps could cooperate without path changes
Are we mounting into /home/username now? Thought it was /media
I meant /media/username/mountpoint
Because the default is plain /media/mountpoint in non-ltsp clients
So if you let udisks on the client automount a stick, that's where it'll end up
And it's not suitable for multiuser systems, we'd probably get collisions on the server-side this way
(so udisks needs to be changed a bit there, if we decide to use it)
Btw, why was ltspfs developed instead of e.g. reusing sshfs?
1) No local user on the thin client.
2) problems with copying the keys around.
(1) would be solved now, right?
And I'm sure we can find a way to solve (2)...
But, do we want to? Would it be any better, e.g. easier to maintain etc?
How are you going to get an sshfs mount back to the client without asking for a password? And what password would we use?
We can autogenerate a key for the specific user
I don't have the details there but let's suppose we find a good way
What are the pros? encryption, less maintainance? And the cons? Slower?
I've used program sshpass in frontend script for rsync to pass ssh password for it.
It would be fun to only allow thin client ssh connections from localhost, and use ssh port forwarding on the master client => server ssh socket ... but it would result in duplicate encryptiong => 2 times slower
Meh those ssh guys should really reconsider their status about the null cipher...
So, if we were to run sshd on the thin client, in order to get a seamless mount back, we'd have to:
1) either dynamically generate the ssh host keys every time the thin client boots, or just have all thin clients have the same host key.
2) on login, install the host key down into the users' known_hosts
known_hosts can be saved to /dev/null, so it wont ask if it changes.
Install a users pub key down on the thin client. This could be a problem if the user already has a key that asks for a password.
Äh, I meant the key can be saved to null..
the third is the biggest problem, I think.
Can't we set AuthorizedKeysFile to the client sshd_config to point to one of our autogenerated files?
E.g. in /home/username/.config/ltsp/authorized_keys
And do the sshfs with an "ssh -F configfile"?... Possible.
ok, I'll have a look at what I'd have to hack in my pam scripts in order to auto-generate all this stuff.
sbalneav: about udisks etc... I like it when ltsp reuses existing code, but I'm worried if the true thin clients will become a bit bloated then
Specifically, what does udisks need in order to operate? Just an "active user"? Or it requires a system dbus as well? Maybe accountsservice too?
Not sure. Still trying to figure that out. There's other "standalone" programs that'll handle automounting.
we could adapt one to our purposes.
|19:54||highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage)|
The idea is to not have to write the udev rules ourselves, right?
If possible, yeah.
|19:56||Parker955 is now known as Parker955_Away|
|20:03||Parker955_Away is now known as Parker955|
|20:07||adriansnyman has joined IRC (email@example.com)|
|20:08||MadX_ZA has left IRC (MadX_ZAfirstname.lastname@example.org, Read error: Connection reset by peer)|
we did ltspfs because sshfs couldn't tell us how much space was available on the remote filesystem
jammcq: but now it can, right?
$ df -h
Filesystem Size Used Avail Use% Mounted on
localhost:/etc 30G 27G 2,1G 93% /tmp/etc
Hehe... reverse sshfs: http://sourceforge.net/mailarchive/message.php?msg_id=27034864
With that one, we wouldn't need an sshd on the client :)
|20:46||andygraybeal_ has joined IRC (email@example.com)|
|21:09||zdgffdgz has joined IRC (firstname.lastname@example.org)|
|21:11||alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)|
it was not possible to get local sound in fluxbox
Local sound does works with KDE!
zdgffdgz: What do you mean?
I could get local sound in flluxbox environment, in KDE it works.
*I could NOT in fluxbox*
zdgffdgz: Loud and clear as your nick :P
|21:57||Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 245 seconds)|
oops wrong keyboard
Even more clear... :P
|22:01||designbybeck_ has left IRC (email@example.com, Remote host closed the connection)|
|22:14||dead_inside has left IRC (firstname.lastname@example.org, Ping timeout: 246 seconds)|
No sound: Ican connect with pulseaudio manager (paman) on the server, not on the client. . .
I get a 'Connection refused' on the client
|22:32||lee has left IRC (email@example.com, Ping timeout: 264 seconds)|
|22:33||lee has joined IRC (firstname.lastname@example.org)|
|22:34||komunista has left IRC (email@example.com, Quit: Leaving.)|
|22:35||dead_inside has joined IRC (firstname.lastname@example.org)|
|22:39||staffencasa has left IRC (email@example.com, Quit: Leaving)|
|22:39||dead_inside has left IRC (firstname.lastname@example.org, Ping timeout: 252 seconds)|
|22:43||telex has left IRC (email@example.com, Remote host closed the connection)|
|22:44||telex has joined IRC (firstname.lastname@example.org)|
|23:16||bobby_C has left IRC (bobby_Cemail@example.com, Ping timeout: 276 seconds)|
|23:24||vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)|