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


Channel log from 17 August 2015   (all times are UTC)

00:36Ark74 has joined IRC (Ark74!~Ark74@189.220.255.72.cable.dyn.cableonline.com.mx)
00:48AlexPortable has left IRC (AlexPortable!uid7568@gateway/web/irccloud.com/x-guucusfzglwhzjjk, Quit: Connection closed for inactivity)
01:07pppingme has left IRC (pppingme!~pppingme@unaffiliated/pppingme, Ping timeout: 260 seconds)
02:19pppingme has joined IRC (pppingme!~pppingme@unaffiliated/pppingme)
03:33Parker955_Away has left IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f, Ping timeout: 244 seconds)
03:34Parker955_Away has joined IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f)
03:38Parker955_Away has left IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f, Ping timeout: 244 seconds)
03:39Parker955_Away has joined IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f)
03:43Parker955_Away has left IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f, Ping timeout: 244 seconds)
03:47Parker955_Away has joined IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f)
03:52Parker955_Away has left IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f, Ping timeout: 246 seconds)
03:57Parker955_Away has joined IRC (Parker955_Away!~parker@2001:470:8:a61::8bdc:c4f)
04:08Ark74 has left IRC (Ark74!~Ark74@189.220.255.72.cable.dyn.cableonline.com.mx, Quit: Saliendo)
05:29ricotz has joined IRC (ricotz!~rico@ubuntu/member/ricotz)
05:57
<gehidore>
work_alkisg: I have the option of a 4 core phenom cpu with 8G of ram if I buy a new mobo for this machine that got physically damaged
05:57
to use as a server
06:12telex has left IRC (telex!teletype@freeshell.de, Remote host closed the connection)
06:14telex has joined IRC (telex!teletype@freeshell.de)
06:28vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
07:29work_alkisg is now known as alkisg
07:30
<alkisg>
Good morning
07:30* vagrantc waves
07:30
<alkisg>
gehidore: sorry I don't remember if we talked about thin vs fat clients, and I was going to suggest using fats if your clients can cope with it...
07:31
<gehidore>
alkisg: we had and you weren't impressed with the p4 with 1G that most of these machines are
07:31
<vagrantc>
alkisg: i'm giving dracut some serious consideration ... initramfs-tools is a bit stalled
07:31
<alkisg>
gehidore: I use such clients as fats
07:31
<gehidore>
they support up to 4G and the price is low enough I can probably budget funds for it
07:31
<alkisg>
!ltsp-pnp
07:31
<ltsp`>
ltsp-pnp: ltsp-pnp is an alternative (upstream) method to maintain LTSP installations for thin and fat clients that doesn't involve chroots: https://help.ubuntu.com/community/UbuntuLTSP/ltsp-pnp
07:31
<alkisg>
So I would suggest this ^
07:32
vagrantc: let me see if it's possible to use dracut with ubuntu, if so, sure!
07:32
<gehidore>
alkisg: I'll keep playing with that then, it must have been a misunderstanding from me
07:32
<vagrantc>
alkisg: going to a talk later aout the future of initramfs options in debian
07:32
<gehidore>
s/from/by
07:32
<vagrantc>
https://summit.debconf.org/debconf15/2015-08-17/
07:33
<alkisg>
http://askubuntu.com/questions/248970/how-to-install-dracut-on-ubuntu ==> With the current situation, you cannot install dracut without having broken packages.
07:33
<vagrantc>
ouch
07:33
well, if debian abandons initramfs-tools ... maybe ubuntu will reconsider
07:33
<alkisg>
I think after filing a few bug reports, that could be solvable...
07:34
<vagrantc>
well, talks about to begin, need to grab a good seat!
07:35
<alkisg>
Listen well, we'll want you to tell us about it afterwards :)
07:37
<vagrantc>
uh-oh
07:38
alkisg: so i know you wanted to rewrite the overlay(fs)/aufs code ... but is the current code passable for now?
07:38
<alkisg>
vagrantc: I didn't have time to test yet, I was trying to solve the "clients don't shutdown bug" as it seamed more urgent,
07:38
then I was away for a couple of days, I'll catch up this week
07:39* alkisg is planning to put more work in ltsp this week
07:39
<vagrantc>
ok, i've still got time here and there for working on stuff... it looked like everything was working on my end (other than overlay+NFS)
07:40
<alkisg>
We can chat about the other bug reports you've linked, after the initramfs talk...
07:41
<vagrantc>
initramfs isn't for a couple hours ... maddog's (an LTSP advocate on many occasions) giving a talk soonish
07:41
<alkisg>
Did you read the irclogs? I wrote a couple of comments about those bugs
07:41
E.g. how to reproduce one of them
07:41
<vagrantc>
https://summit.debconf.org/debconf15/meeting/326/two-contests-no-waiting/
07:42
don't know if maddog will be talking about LTSP, but maybe
07:42
alkisg: i didn't catch the logs, no
07:43
<alkisg>
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1109029 says dracut should be installable from vivid+... /me tries in wily...
07:44
<vagrantc>
found it in the logs
07:45
i think i tested with /home/Public and /home/SOMEUSER mounted, but not with /home itself as a mounted partition
07:46
when i disabled special handling of /home and removed /home/* from ltsp-update-image.excludes it still, obviously, excluded the directory
07:46
i also experimented with EXCLUDED_DIRS=/sdlkfjdlfjsdlfk
07:46
which should be the same
07:46
<alkisg>
When I wrote that part, I was worried about bugs resulting in deleting things under /home
07:46
<vagrantc>
as disabling it
07:47
<alkisg>
After a few years of real world use, I'm not very worried anymore
07:47
I think we can simply delete that logic
07:47
<vagrantc>
like the deluser stuff?
07:47
<alkisg>
Yes
07:47
<vagrantc>
i think it would be useful to keep the variable, though
07:47
<alkisg>
E.g. suppose someone did userdel --remove-home...
07:47
Sure we can keep it, just make it default to empty
07:48
<vagrantc>
right. ok, will implement
07:48
<alkisg>
Thanks :)
07:48
<vagrantc>
still not sure about the ltsp-update-image.excludes ?
07:48
<alkisg>
Dracut has a very nice cmdline... http://manpages.ubuntu.com/manpages/trusty/man7/dracut.cmdline.7.html
07:48
I think that we need to keep /home/* in .excludes
07:48
Otherwise, by default, all /home will be included
07:49* vagrantc seems to recall reading a bug report on dracut where it handled ip= differently than the kernel :(
07:49
<alkisg>
The kernel handling sucks so I wouldn't mind about that :D
07:49
ip={dhcp|on|any|dhcp6|auto6|ibft}
07:49
ip=<client-IP>:_<server-IP>_:_<gateway-IP>_:_<netmask>_:_<client_hostname>_:_<interface>_:_{none|off|dhcp|on|any|dhcp6|auto6|ibft}_[:[<mtu>][:_<macaddr>_]]
07:50
<vagrantc>
well, things like pxelinux use the kernel commandline for ipappend
07:50
<alkisg>
It covers all of our use cases, afaik
07:50
ipappend included
07:50
<vagrantc>
i hope it works, i didn't test it or follow-up.
07:51
<alkisg>
root=nfs4:[<server-ip>:]<root-dir>[:_<nfs-options>_], root={dhcp|dhcp6}
07:51
netroot=nbd:_<server>_:_<port>_[:_<fstype>_[:_<mountopts>_[:_<nbdopts>_]]]
07:52
root-path=nbd:_<server>_:_<port>_[:_<fstype>_[:_<mountopts>_[:_<nbdopts>_]]]
07:52
(with dhcp)
07:53
The only thing that I see missing, is support for aoe
07:56
<vagrantc>
i guess if the user wants to include some stuff under /home/* , they can tweak ltsp-update-image.excludes ...
07:57
does it only support globbing, and not full regex? is there a negation syntax or smething?
07:57* vagrantc gives dracut a try with stretch
07:58* vagrantc wonders about backports with dracut vs. initramfs-tools
07:58
<alkisg>
I think that we call it with support for regexes
07:59
Ah no, just -wildcards
07:59
-wildcards
07:59
Allow extended shell wildcards (globbing) to be used in exclude dirs/files
07:59
(from man mksquashfs)
07:59
<vagrantc>
right
07:59
i didn't read anything about negation ... so it might be hard to handle shared public dirs and such
07:59
<alkisg>
If dracut uses a different name for the initramfs that it generates, it should be easy to support both of them for a while
08:00* vagrantc doubts it
08:00
<vagrantc>
but maybe it's configurable
08:00
<alkisg>
Why would someone include public shared dirs inside the nbd image?
08:01
As opposed to e.g. mounting them with EXTRA_MOUNTS?
08:01* vagrantc shrugs
08:02* alkisg prefers not to care for cases that he can't find a use for, until someone actually requests to support them... :)
08:02* vagrantc nods
08:09
<vagrantc>
heh. well, apparently installing dracut in stretch removes initramfs-tools and ltsp-client*
08:10
<alkisg>
In Ubuntu it's not installable, it removes ubuntu-minimal etc as well
08:10langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:10
<alkisg>
I tried `dpkg -r --force-all initramfs-tools ltsp-client*`, but that wasn't enough...
08:12langxm has left IRC (langxm!~langxm@218.57.238.129, Read error: Connection reset by peer)
08:12langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:16
<vagrantc>
probably need to allow dracut as an alternative for ltsp-client...
08:16
<alkisg>
I think there's a virtual package for initramfs
08:16
linux-initramfs-tool
08:16
We can depend on that one
08:17
I'm not sure that it's necessary to depend on initramfs though
08:17
ah, we'd need to move the COW part to init-ltsp first
08:17
If we did that, then we wouldn't need the initramfs to have any ltsp-related code...
08:18
I.e. nbd or nfs root could be mounted read only, and the tmpfs+cow part could be done (along with a pivot_root) from init-ltsp
08:19
I managed to get an initramfs with dracut by running dpkg -i --force dracut*.deb
08:19
I guess we could play a bit with it to prepare the ltsp code for the future, when dracut will be installable...
08:20* vagrantc just rebuilds ltsp
08:25langxm has left IRC (langxm!~langxm@218.57.238.129, Read error: Connection reset by peer)
08:26langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:28langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:32langxm has left IRC (langxm!~langxm@218.57.238.129, Read error: Connection reset by peer)
08:33langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:37
<alkisg>
Using upstart, the 15.10 clients do shut down fine...
08:37langxm has left IRC (langxm!~langxm@218.57.238.129, Read error: Connection reset by peer)
08:37
<alkisg>
It's only when using systemd that they don't
08:38langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:39
<alkisg>
vagrantc: about the other bug with sshd, I think that we can just put it in the RM_SYSTEM_SERVICES by default, and tell the users that want it, to put it in KEEP_SYSTEM_SERVICES=sshd
08:39
If that's too much, we can at least do that if we detect an ltsp-pnp environment
08:45langxm has left IRC (langxm!~langxm@218.57.238.129, Read error: Connection reset by peer)
08:46langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:46
<alkisg>
vagrantc: in debian, do you have an ltsp-client-core service file for systemd? Or just the sysvinit one?
08:47langxm has left IRC (langxm!~langxm@218.57.238.129, Read error: Connection reset by peer)
08:48langxm has joined IRC (langxm!~langxm@218.57.238.129)
08:52
<vagrantc>
alkisg: just sysvinit so far, but would like to explore what it can get us
09:04* Hyperbyte <3 systemd
09:05* vagrantc remains ambivalent
09:07
<Hyperbyte>
I've been using it for years now. No more race conditions, parallel startup and startup conditions for services... :-)
09:09
<vagrantc>
i've been using it for about a year and it's worked fine more or less... only had to make a few patches to fix LTSP support
09:11Grembler has joined IRC (Grembler!~Ben@cpc29-aztw22-2-0-cust128.18-1.cable.virginm.net)
09:11
<alkisg>
For starters, it's not defining the RUNLEVEL environment variable, which some sysvinit services use...
09:21fiesh has left IRC (fiesh!~fiesh@hq.wsoptics.de, Ping timeout: 272 seconds)
09:26Grembler has left IRC (Grembler!~Ben@cpc29-aztw22-2-0-cust128.18-1.cable.virginm.net, Remote host closed the connection)
09:41
<alkisg>
...and the second problem is that it's doing `ifdown eth0`, so nbd-disconnect hangs because it cannot run `nbd-cilent -d`
09:42Freejack has left IRC (Freejack!~Freejack@unaffiliated/freejack, Ping timeout: 240 seconds)
09:43
<alkisg>
Btw, nbd-server processes still keep running if nbd-disconnect doesn't run at poweroff, so it's still needed in Wily
09:45Freejack has joined IRC (Freejack!~Freejack@unaffiliated/freejack)
09:52
<alkisg>
...and the third problem is that systemd kills sysvinit jobs that have even a tiny delay like `sleep 1`
09:54
Maybe https://bugs.debian.org/784720 is related...
09:58bojan has joined IRC (bojan!75c1362a@gateway/web/freenode/ip.117.193.54.42)
09:58
<bojan>
I am using thin clients and i want to bock some of the users not to use home space for saving folders or creating folders...Is it possible??
09:58
block
09:59
<alkisg>
So that their /home/username is essentially in RAM and not in the disk?
10:01
bojan, are you still here?
10:02
<bojan>
alkisg:sorry,YES
10:03
<alkisg>
To both questions? :)
10:04
<bojan>
i have some 30 number of users can login in my server and i have created a lvm partition and mounted it on /mnt and kept link on all the users desktop
10:04Grembler has joined IRC (Grembler!~Ben@cpc29-aztw22-2-0-cust128.18-1.cable.virginm.net)
10:04
<bojan>
But the thing is people are not only saving their files in that particular luink which is mounted on /mnt and they are also saving in desktop
10:05
I want my users not to have any files or folders saved in desktop
10:05
how to restrict them??
10:05
<alkisg>
Why don't you symlink /home/username/Desktop to /mnt/username/Desktop ?
10:05
So that when they save things on desktop, it will go to /mnt ?
10:06
<bojan>
Yes i can but here we have multiple projects with different permission for the users
10:06
<alkisg>
You can also just use chmod to restrict access to their own desktop folder... that way they'll get access denied
10:06
<bojan>
can you tell me how to do that??
10:06
MEans how to restrict access ??
10:06
am little confused on that??
10:06
<alkisg>
The command is chmod -r, it's not related to ltsp
10:07
Sorry, chmod -w
10:07
chmod -w folder means "don't allow the users to write to that folder"
10:07
<bojan>
chmod -w /home/username/Desktop??
10:07
<alkisg>
Yup
10:07* vagrantc would expect that to issue warnings
10:08
<bojan>
How to make them activate again??
10:08
+w??
10:08vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
10:08
<alkisg>
Yes
10:09
<bojan>
thank you
10:10
<alkisg>
I think it would make more sense to symlink Desktop to /mnt/somedir though
10:18bojan has left IRC (bojan!75c1362a@gateway/web/freenode/ip.117.193.54.42, Quit: Page closed)
10:21
<alkisg>
The fourth problem with systemd is that it deactivates nbd-swap without calling nbd-client -d
10:27
<Hyperbyte>
alkisg, why don't you make a systemd ltsp service that is dependant on network target, and put all this stuff there?
10:28
Then it gets run only after network is available on startup and on shutdown will run before network is brought down
10:28
That's the beauty of systemd. You can tell it "this service MUST but shutdown before you bring down the network"
10:28
<alkisg>
Hyperbyte: and if I want to make sure that it runs before swap is deactivated as well, I would also depend on the swap target?
10:29
And, which one is the swap target? I can only see it hardcoded in swap.c...
10:29
Ah, it still shows in targets though
10:29
Yup, that could work, if it works as advertised...
10:29
<Hyperbyte>
To me, it sounds like most problems you're having with it is because you're using a sysvinit approach to systemd. It's completely new. You should forget about runlevels and stuff, and depend on targets.
10:30
<alkisg>
It's supposed to have a compatibility module for sysvinit scripts
10:30
So one approach is to consider the problems a bug in their compatibility layer
10:30
Another is what you suggested, which may be the easiest way to do it...
10:31
...and we're probably going to write a systemd service anyway, so maybe now is the time to do it
10:32
<Hyperbyte>
I don't think compatibility layer was made for production use, more as fallback method.
10:32
For compatibility with old packages that don't support systemd
10:32
<alkisg>
Hyperbyte, have you created any systemd targets of your own? Do you think you can propose one for ltsp?
10:33
*services... my systemd terminology is probably wrong
10:34
<Hyperbyte>
I did once, not sure where it went
10:34
<alkisg>
cyberorg, warren, maybe you already have a systemd service for ltsp?
10:35
<Hyperbyte>
But it's pretty easy, systemd scripts are easier than init scripts
10:35
Do you have access to a machine with systemd?
10:35
<alkisg>
Yes, but I'd need to read up on those services, targets, wants etc
10:35
<Hyperbyte>
ls /usr/lib/systemd/system/
10:36fiesh has joined IRC (fiesh!~fiesh@hq.wsoptics.de)
10:36
<Hyperbyte>
Just open one .service file in a text editor and you'll be like "omg that's all?!"
10:36
<alkisg>
There are .service, .target, .timer, .slice etc files there
10:36
<Hyperbyte>
And you'll also immediately know how to write ltsp .service files.
10:36
<alkisg>
And some .wants directories
10:36
So it's not as simple as it sounds...
10:36
The sysinit.target is probably part of the compatibility layer
10:37
and it has a full .wants/ dir...
10:37
<Hyperbyte>
You don't need all that
10:37
You can just create a .service file and do systemctl daemon-reload
10:37
<alkisg>
I would need to depend on swap, right? Isn't that done via the wants directory?
10:37
<Hyperbyte>
No
10:38
http://patrakov.blogspot.nl/2011/01/writing-systemd-service-files.html
10:39
Pretty sure the .wants directory is maintained by the daemon itself
10:39
Either way, it's just symlinks.
10:39
<alkisg>
The files in /lib are parts of packages
10:39
I.e. they're included in the .deb files
10:39
Processes don't write there
10:40
<Hyperbyte>
You're right, it should all go in /etc/systemd/system/
10:40
But for examples it's a good place to look.
10:40
<alkisg>
Thanks Hyperbyte, I'll read up on debconf etc too, to do it properly in the debian way
10:40
<Hyperbyte>
Right, I'm speaking from Centos 7. :-)
10:40
<cyberorg>
alkisg, https://sourceforge.net/p/kiwi-ltsp/code/HEAD/tree/trunk/kiwi-ltsp/ltsp/suse-13.2/root/usr/lib/systemd/system/ltsp-init.service
10:40
<alkisg>
I.e. that will probably be a debian/ltsp-client-core.service file that will get installed in the proper places
10:40
Thank you cyberorg, I'll check that after lunch
10:40* alkisg waves :)
10:41
<Hyperbyte>
Enjoy!
10:41alkisg is now known as work_alkisg
10:42
<cyberorg>
work_alkisg, not sure what exactly you are looking for, most stuff is handled by init=/sbin/init-ltsp
11:10vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
11:17AlexPortable has joined IRC (AlexPortable!uid7568@gateway/web/irccloud.com/x-uhqiochhicmamvyn)
11:17Grembler has left IRC (Grembler!~Ben@cpc29-aztw22-2-0-cust128.18-1.cable.virginm.net, Quit: I Leave)
11:28
<Hyperbyte>
cyberorg, that .service file has double After= lines in [Unit]. Are you sure that works? To me it'd seem only the last After= line gets processed.
11:29
Also, out of curiousity, why does it have "xdm.service" as alias?
11:31
<cyberorg>
Hyperbyte, yeah, looks like leftover from file it was copied from, LDM replaces XDM so it is left there for other services dependent on xdm.service
11:32
<Hyperbyte>
Ah, makes sense.
11:35danau11 has joined IRC (danau11!~durban@12.197.179.122)
11:37vsuojane1 is now known as vsuojanen
11:37danau11 has left IRC (danau11!~durban@12.197.179.122)
11:43
<muppis>
!chromium
11:43
<ltsp`>
Error: "chromium" is not a valid command.
11:44
<muppis>
!chrome
11:44
<ltsp`>
Error: "chrome" is not a valid command.
11:44
<muppis>
Meh.
11:44Faith has joined IRC (Faith!~paty@unaffiliated/faith)
11:52Freejack has left IRC (Freejack!~Freejack@unaffiliated/freejack, Remote host closed the connection)
11:52Freejack has joined IRC (Freejack!~Freejack@unaffiliated/freejack)
12:11
<vagrantc>
!chrom
12:11
<ltsp`>
Error: "chrom" is not a valid command.
12:19adrianorg has left IRC (adrianorg!~adrianorg@177.132.217.210, Ping timeout: 244 seconds)
12:20adrianorg has joined IRC (adrianorg!~adrianorg@179.180.175.106)
12:28ogra_ has left IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de, Ping timeout: 272 seconds)
12:29ogra_ has joined IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de)
13:57ben_roose has joined IRC (ben_roose!~roose@roose.cs.wichita.edu)
13:58vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Ping timeout: 244 seconds)
14:05vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
14:45work_alkisg is now known as alkisg
15:02
<alkisg>
vagrantc: have you guys ever talked about "what exactly is the ltsp-client-core service"?
15:02
I'm trying to write a systemd service for it
15:04
E.g. what does `stop ltsp-client` mean? Does it mean to stop the screen scripts?
15:04
(I mean, in the ideal case...)
15:04
<vagrantc>
i think it's primarily for starting screen scripts
15:05
<alkisg>
So, `stop ltsp-client`, doesn't mean that the user wants to poweroff the client, right?
15:07
<vagrantc>
sure
15:08
<alkisg>
From lightdm.service: "Conflicts=getty@tty7.service"
15:09
I wonder what's the proper way to handle all those screens in the systemd era...
15:09
Maybe init-ltsp.d should dynamically create one .service file for each screen
15:10
The default display-manager is now defined with the display-manager.service symlink in /etc
15:10
<vagrantc>
possibly to create .service files for each screen script
15:10
in init-ltsp.d phase or similar
15:10
<alkisg>
Since each screen script can be ran multiple times, or none at all, maybe one for each vt
15:10
with possible copies + sed's
15:11
(conflit=getty@ttyX.service...)
15:11
systemd does have 7 different services, one for each tty...
15:12
And we'd also need a documentation page for the Documentation:man:ltsp-client(1) line
15:13
Ah, each screen script should have its own man page there
15:21
One idea is to have
15:34
>> A unit configuration file whose name ends in ".target" encodes information about a target unit of systemd, which is used for grouping units and as well-known synchronization points during start-up.
15:34
That might match the "ltsp-client" concept better... while ltsp3.service is more like tty3.service
15:43telex has left IRC (telex!teletype@freeshell.de, Remote host closed the connection)
15:44telex has joined IRC (telex!teletype@freeshell.de)
15:48Phantomas has joined IRC (Phantomas!~ftsamis@2a02:582:1081:8f00:f81a:911e:6835:161c)
15:48Phantomas has joined IRC (Phantomas!~ftsamis@ubuntu/member/phantomas)
15:54vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
15:55mikkel_ has left IRC (mikkel_!~mikkel@mail.dlvs.dk, Quit: Leaving)
16:00vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
16:05
<alkisg>
This works for emulating SCREEN_02=shell in systemd: https://wiki.gentoo.org/wiki/Automatic_login_to_virtual_console
16:31* vagrantc has only be half-present
16:50vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
17:45
<alkisg>
https://github.com/systemd/systemd/blob/3a3701f03360f9a292dcab31fda1f6e0e0629c64/units/getty%40.service.m4
17:48
This generates multiple getty@ttyX services... that's probably what we should do with ltsp screens as well
17:48
I hope we don't have to implement systemd.generator. Hyperbyte, nope, systemd is not as easy as you made it sound :D
17:49
It's a good project with good documentation, but it does have its learning curve
17:51
I think the easiest and most compatible way to support all init systems, would be to have init-ltsp.d rename agetty to agetty.real, and create a wrapper script there instead :D
17:51
Really lame, really easy and compatible with everything :D
17:58
<quinox>
sounds like there are no downsides!
18:07
<gehidore>
as a systemd user I say ... as long as it works, I don't care how you did it :P
18:10* alkisg waves!
18:10alkisg is now known as work_alkisg
19:32mip5 has joined IRC (mip5!~chatzilla@208.72.71.69)
19:37
<mip5>
Hi - updated an ubuntu 12.04.5 to 14.0.3, and rebuilt ltsp clients. Trying to use amd64 this time (rather i386) as arch. Edited /etc/ltsp/dhcpd.conf to reflect new arch. The build process appears to go ok, but the clients hang after "Trying to load pxelinux.cfg/default....ok. I get "Negotiation: ..Error: Server closd connection. Exiting." Then I get dropped to a busybox. Any ideas?
19:39
<gehidore>
did you update which pxe image is being served?
19:40
<mip5>
I'm not sure what you mean. I moved the old one (ltsp-old), before I built the new one, and edited /etc/ltsp/dhcpd.conf and replaced references to i386 with amd64. Have I missed something?
19:43
I ran "service nbd-server restart" but that hasn't helped.
20:00
<gehidore>
for the tftp server...
20:00
it's prolly still pointed at the i368 image
20:00
which doesn't exist
20:04ricotz has left IRC (ricotz!~rico@ubuntu/member/ricotz, Quit: Ex-Chat)
20:05
<mip5>
Where would else would I specify it to use amd64? FWIW - looking around in busybox shell I seen lib64, and other stuff that looks like the client is getting at least some of what it needs.
20:09
<gehidore>
mip5: it's going to be specific to the distro I suspect, I just know I ran into the same issue when I was tinkering with it on a debian VM and switched from 386 to amd64
20:11
<mip5>
gehidore: Thanks. other errors I forgot to mention: mounting /dev/nbd0 on /root failed: Invalid argument. So you're right, it's not getting the image it needs.
20:12
gehidore: I'm using ubuntu, though tempted to try debian.
20:14
ls
20:15
<gehidore>
mip5: mostly I was doing research to see how everything needed to flow so I could establish a system on arch
20:26
<mip5>
gehidore: thanks. I'll keep poking around
20:36Ark74 has joined IRC (Ark74!~Ark74@189.220.255.72.cable.dyn.cableonline.com.mx)
21:04Faith has left IRC (Faith!~paty@unaffiliated/faith, Quit: Saindo)
21:05
<work_alkisg>
mip5: what's the output of `cat /proc/cmdline` on the client initramfs?
21:07
<mip5>
work_alkisg: ro initrd=initrd.img root /dev/nbd0 init=/sbin/init-ltsp quiet splash plymouth:force-splash vt.handoff=7 nbdroot=:ltsp_amd64 BOOT_IMAGE=vmlinux
21:08
<work_alkisg>
mip5: the command line is wrong, remove nbdroot from there
21:08
(it's the old style)
21:09
also, check /etc/nbd-server/conf.d/ltsp_amd64.conf, that it has [/opt/ltsp/amd64] as its header
21:09
finally, verify that your dhcp server sends ROOTPATH=/opt/ltsp/amd64
21:10
(the command line is edited in /var/lib/tftpboot/ltsp/amd64/pxelinux.cfg/default)
21:12
<mip5>
work_alkisg: ok - making changes now
21:16
work_alkisg: Same error.
21:17
<work_alkisg>
mip5: and nbdroot=:ltsp_amd64 now is not in the kernel command line on the client/
21:17
?
21:18
<mip5>
work_alkisg: though cat /proc/cmdline looks good now - nbdroot entry is *not* there
21:18
<work_alkisg>
OK, next step, try this:
21:18
!nbd-client
21:18
<ltsp`>
nbd-client: To try mounting the NBD image from the client initramfs: nbd-client 192.168.67.1 -N /opt/ltsp/i386 /dev/nbd0
21:18
<work_alkisg>
change the ip and i386
21:20
<mip5>
I get: Negotiation: .. size = 580 MB bs = 1023 sz=608448512 bytes
21:21
<work_alkisg>
OK, now try this: `ipconfig -n eth0` again in the client initramfs
21:21
Check rootpath there, is it /opt/ltsp/amd64?
21:21Ark74 has left IRC (Ark74!~Ark74@189.220.255.72.cable.dyn.cableonline.com.mx, Ping timeout: 250 seconds)
21:21gbaman has joined IRC (gbaman!~gbaman@104.40.144.249)
21:22
<mip5>
rootpath is /opt/ltsp not /opt/ltsp/amd64
21:22
<work_alkisg>
Fix your dhcp
21:23
<mip5>
thanks - rebooting client now after restarting dhcp service
21:24
client boots! Yay! but screen is shifted far to the left. Can only see "ntu" and beneath that, ername and box
21:24
<work_alkisg>
(12:09:45 πμ) work_alkisg: finally, verify that your dhcp server sends ROOTPATH=/opt/ltsp/amd64
21:25
same advice 12 minutes ago :)
21:25
<mip5>
;-)
21:25
<work_alkisg>
Is it a graphics screen, LDM, or a text screen?
21:26
<mip5>
mostly white, not sure whether that's called graphics or LDM
21:27
<work_alkisg>
There's a bug in which the clients think they have 2 monitors, and the screen is centered between them, even if one of them doesn't really exist
21:27
Do you think that description fits what you're seeing?
21:28
<mip5>
Could be - the monitor is pretty wide.
21:28
it could be the right half of a dual monitor setup where the login was spread across both
21:28
<work_alkisg>
Then try with SCREEN_07=xterm in lts.conf,
21:29
ah no easier way
21:29
just try to login blindly
21:29
type username/enter/password/enter
21:29
then open a terminal and type xrandr and see the outputs there
21:29
finally, create an xrandr command line to put it in XRANDR_COMMAND_01 in lts.conf
21:29
<mip5>
it centers when I enter the credentials - logging in ow
21:30* work_alkisg needs to go, good luck
21:30
<mip5>
thanks!
21:40gbaman has left IRC (gbaman!~gbaman@104.40.144.249, Ping timeout: 240 seconds)
22:11
<mip5>
work_alkisg: I was able to login and open display preferences, and see that client "thinks" there are 2 monitors. I couldn't fix it using xrandr, but I probably didn't use the proper syntax for that line in lts.conf. Also - the machines are super, super slow. I'll try to work on this more tomorrow. Thanks *so* much for your help. I figured dhcpd.conf was wrong, but obviously didn't quite get...
22:11
...it right until you helped.
22:16Phantomas has left IRC (Phantomas!~ftsamis@ubuntu/member/phantomas)
22:25Cramer has joined IRC (Cramer!42fff48c@gateway/web/freenode/ip.66.255.244.140)
22:28telex has left IRC (telex!teletype@freeshell.de, Remote host closed the connection)
22:30telex has joined IRC (telex!teletype@freeshell.de)
22:32
<mip5>
work_alkisg: I found another way to disable the 2nd monitor. I added video=LVDS-1:d to /pxelinux.cfg/default - just added video=LVDS-1:d
22:36mip5 has left IRC (mip5!~chatzilla@208.72.71.69)
23:38AlexPortable has left IRC (AlexPortable!uid7568@gateway/web/irccloud.com/x-uhqiochhicmamvyn, Quit: Connection closed for inactivity)
23:42ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)