LTSP 5 is in minimal maintenance mode
The new LTSP is hosted at https://ltsp.github.io

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


Channel log from 29 July 2019   (all times are UTC)

03:07shored has left IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi, Read error: Connection reset by peer)
03:07shored has joined IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi)
04:54gdi2k has joined IRC (gdi2k!~gdi2k@58.69.160.28)
05:20statler has joined IRC (statler!~Georg@p54897D17.dip0.t-ipconnect.de)
07:28othin has joined IRC (othin!c367adf3@195.103.173.243)
07:59woernie has joined IRC (woernie!~werner@p578bb7b6.dip0.t-ipconnect.de)
08:05
<othin>
Hi guys! there is a way to put a crontab in a linux thinclinet pro? when i do a reboot, reset itself
08:25othin has left IRC (othin!c367adf3@195.103.173.243, Remote host closed the connection)
09:09statler has left IRC (statler!~Georg@p54897D17.dip0.t-ipconnect.de, Remote host closed the connection)
09:22alexxtasi[m] has left IRC (alexxtasi[m]!alexxtasim@gateway/shell/matrix.org/x-yxhrkrxhrehjhsjr, Write error: Connection reset by peer)
09:31alexxtasi[m] has joined IRC (alexxtasi[m]!alexxtasim@gateway/shell/matrix.org/x-xequehrsfepufixs)
09:49pavars has joined IRC (pavars!5509db4e@85.9.219.78)
09:56statler has joined IRC (statler!~Georg@gwrz.lohn24.de)
10:30pavars has left IRC (pavars!5509db4e@85.9.219.78, Ping timeout: 260 seconds)
10:38vsuojanen has left IRC (vsuojanen!~vsuojanen@cable-hml-585686-205.dhcp.inet.fi, Ping timeout: 268 seconds)
11:20pavars has joined IRC (pavars!5509db4e@85.9.219.78)
11:57pavars has left IRC (pavars!5509db4e@85.9.219.78, Remote host closed the connection)
11:58section1 has joined IRC (section1!~section1@178.33.109.106)
11:59vsuojanen has joined IRC (vsuojanen!~vsuojanen@cable-hml-585686-205.dhcp.inet.fi)
13:33ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
13:37
<section1>
hello all...have some questions ..like..my clients are getting black screen on vt7 where should be showing ldm login screen... i run a ps auc in tty1 and i see running xorg and ldm ...not sure what can be the problem.
13:38
the server need to run Xorg ?
14:05woernie has left IRC (woernie!~werner@p578bb7b6.dip0.t-ipconnect.de, Remote host closed the connection)
14:09
<section1>
this is a debian server.
14:09
without xorg running but have xorg installed
14:10
<mwalters>
I think modern debian defaults to the wayland renderer? that might be the issue? I'm just guessing based on what I picked up from other recent debian issues.
14:12
<section1>
mm maybe ...well im trying with ubuntu...either way clients will be ubuntu..so more easy ..really new at ltsp..
14:12
thanks for the hint mwalters
14:13
<mwalters>
Hmm, I think if wayland was the issue, you'd see the LDM screen, but it'd break after logging in... so I'm not really sure at all ;)
14:14
but yeah, 18.04 should "work fine" ;)
14:18
<section1>
yeah really i dont see anything when the client start..i have to switch to tty1
14:18
but yeah i go with 18.04 and test..
14:19
<mwalters>
You could set LDM_SYSLOG=true in lts.conf... that'd get you a root login on tty1
14:19
from there you could check the syslog for clues
14:20
(just remove the line when you "go live"... having an open root login on tty1 is definitely a security issue)
14:20
<section1>
trying that..
14:20
<mwalters>
ops, there's another line needed
14:20* mwalters digs
14:21
<section1>
when i change lts.conf i need to update the image ?
14:21
<mwalters>
no, you'll need something else too
14:21
<section1>
with LDM_SYSLOG=true
14:21
<mwalters>
I"m digging
14:21
<section1>
ok
14:21
<mwalters>
that gets LDM to log to syslog, I'm getting my config options confused
14:22
!root
14:22
<ltsp>
I do not know about 'root', but I do know about these similar topics: 'ROOT_PASSWORD_HASH'
14:23
<section1>
i put a root passwd in the image
14:23
so i can login in tty
14:24
<mwalters>
oh ok
14:24
that makes it easy, then... just login as root ;)
14:25
but yeah, you don't need to update image after editing lts.conf
14:32
<section1>
ok don't see any error about ldm ..but i see an error on xorg not sure is fatal..says: modeset(0): failed to se mode: invalid argument
14:33
well now i try ubuntu.
14:35
<alkisg>
section1: if you press alt+ctrl+f8/f7 etc etc, can you find xorg running?
14:35
<section1>
don't see anything... if i do a ps aux i see xorg running and ldm running
14:35
don't see anything... if i do a ps aux i see xorg running and ldm running
14:35
ops
14:36
<alkisg>
Do you see the same xorg process after 3 seconds, or different?
14:36
The same process id
14:37
<section1>
yeah i think its different now its not running at all
14:37
sure its something about videocar and xorg
14:37
<alkisg>
That probably means it can't initialize xorg and fails. Does this command work? xinit
14:37
Just type xinit without options etc, do you get a terminal at the top left, where you can move the mouse and type?
14:38
<section1>
yeap itts run a xterm
14:38
<alkisg>
So it sounds like a misconfiguration, if xorg is able to run
14:38
<section1>
yeah strange
14:38
<alkisg>
In that xterm,if you run `getltscfg -a`, does it return anything?
14:39
<section1>
and the root xorg.log i don't see that error about modeset
14:39
yes
14:39
de lts.conf
14:39
the*
14:40
<alkisg>
Can you pastebin it? E.g. this sends it to termbin: getltscfg -a | nc termbin.com 9999
14:40
(if you have working internet)
14:40
<section1>
only 3 ...LDM_DIRECTX LOCALDEV LDM_SYSLOG
14:40
maybe a ineed one for a resolution ?
14:41
<alkisg>
Try without an lts.conf
14:41
If you have something wrong there, it might cause crashes
14:41
<section1>
the same happens..
14:41
<alkisg>
OK. How did you build the chroot? Maybe you're missing a package
14:42
<section1>
ltsp-build-client --mount-package-cache
14:43
then chroot and i add a root pass
14:43
and update the image
14:43
<alkisg>
And you tried that in debian buster?
14:43
<section1>
debian 9
14:43
<alkisg>
Which one is that?
14:44
stretch?
14:44
<section1>
yes
14:44
<alkisg>
cat /etc/os-release
14:44
OK
14:44
<section1>
PRETTY_NAME="Debian GNU/Linux 9 (stretch)"
14:44
<alkisg>
So it's a "thin stretch chroot"
14:44
What are you clients, how much ram, which cpu?
14:46
<section1>
2gb ram and i3
14:46
trying with X_MODE_0 = 1024x768 and XRANDR_DISABLE=True
14:46
too se what happen
14:46
<alkisg>
Then there's no reason to use a thin chroot
14:47
<section1>
yeah i know...but this is only a test
14:47
<alkisg>
I mean, you try with the very hard way, instead of the easy way
14:47
And the hard way gives you a thin chroot, while the easy way gives you the correct fat chroot
14:47
So it's both hard to do, and bad results
14:47
Just go with the easy/good results method
14:48
I.e. install a buster desktop server or a 18.04 server, and then follow this:
14:48
!install
14:48
<ltsp>
install: http://wiki.ltsp.org/wiki/Installation/Ubuntu for Ubuntu, or http://wiki.ltsp.org/wiki/Installation for other distributions
14:48
<alkisg>
Use the first link there for debian too, except for the part that mentions a ppa
14:49
Use the debian-live.iso or the ubuntu-desktop.iso. Don't use ubuntu-server.iso to install the server.
14:49
<section1>
oka
14:49
alkisg, i need to run xorg in the server ?
14:49
or only in text mode is sufficient ?
14:49
<alkisg>
section1: it's much easier if the server is a template of the clients
14:50
<section1>
yeah
14:50
<alkisg>
So, if you can do that, do it. Then you can disable xorg if you want
14:50
<section1>
oka
14:50
<alkisg>
I.e. if you can install graphically, install the ubuntu-desktop.iso
14:50
<section1>
try ubuntu 18.04
14:50
i will try*
14:50
thankyou guys.
14:50
<alkisg>
OK. Then follow the steps on that page exactly. You should be ready in 1 hour, from the time you download the .iso, to the time the client boots.
14:51
<section1>
i will report back up.success or fail :D
14:51
<alkisg>
I.e. 1 hour to install the server, run ltsp-update-image -c /, and boot client
14:51
OK
14:51
<section1>
ok
14:51
<alkisg>
Good luck :)
14:51
<section1>
ty :)
15:43GodFather has joined IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com)
15:46GodFather_ has joined IRC (GodFather_!~rcc@d53-64-7-141.nap.wideopenwest.com)
15:59shored has left IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi, Read error: Connection reset by peer)
15:59shored has joined IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi)
16:14section1 has left IRC (section1!~section1@178.33.109.106, Remote host closed the connection)
16:19section1 has joined IRC (section1!~section1@178.33.109.106)
17:29shored has left IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi, Read error: Connection reset by peer)
17:29shored has joined IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi)
17:51
<section1>
when i run add-apt-repository --yes ppa:ts.sch.gr from the guide i get Cannot add PPA: 'ppa:~ts.sch.gr/ubuntu/ppa'. -> Cannot add PPA: 'ppa:~ts.sch.gr/ubuntu/ppa'.
17:51
ERROR: '~ts.sch.gr' user or team does not exist.
17:53
<alkisg>
section1: did you use copy/paste?
17:53
You probably misttyped something
17:53
<section1>
mm reading some other post maybe a issue with proxy.
17:53
<alkisg>
section1: did you put ~
17:53
It doesn't have ~
17:53
<section1>
yeah that shows the console
17:54
<alkisg>
Are you using a proxy?
17:54
<section1>
yeah exits a proxy in the network
17:54
<alkisg>
Does `apt update` work?
17:55
sudo apt update
17:58woernie has joined IRC (woernie!~werner@p57A0EC92.dip0.t-ipconnect.de)
17:59
<section1>
yeap
17:59
<alkisg>
What's the output of: env | grep -i proxy
18:00
Also, you're using 18.04, not 19.xx, right?
18:00
<section1>
yes 18.x
18:00
<alkisg>
18.04 or 18.10?
18:00
18.10 isn't supported
18:01
<section1>
http_proxy=http://192.168.0.1:3128/
18:01
.04
18:01
Ubuntu 18.04.2 LTS
18:01
strange i don't see any connection on the gw
18:01
proxy
18:02
<alkisg>
It sounds very strange that sudo apt works but sudo add-apt-repository doesn't
18:02
Are you sure you're not using e.g. "sudo -i" in the second case, thus losing the proxy?
18:02
try: sudo -E add-apt-repository ppa:ts.sch.gr
18:02
<section1>
yeah i think i now what is
18:03
dnsmasq fuckup the dns
18:03
<alkisg>
You installed dnsmasq before the ppa?!
18:03
<section1>
yea
18:03
<alkisg>
It's important to follow the steps in order :)
18:03
<section1>
its working now
18:04
yes i fuckup in that
18:04
<alkisg>
Do (a), chrootless
18:05shored has left IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi, Read error: Connection reset by peer)
18:05
<section1>
sure ?
18:06
ok
18:06
<alkisg>
Yes, chrootless is stable and easier
18:06
<mwalters>
as someone who didn't... do (a), chrootless ;)
18:06shored has joined IRC (shored!~shored@87-92-94-52.bb.dnainternet.fi)
18:06
<section1>
oka i go with that
18:07
that takes the info for the client with /
18:07
?
18:07
<alkisg>
Yes
18:07
<section1>
from the root server
18:07
ok
18:07
<alkisg>
While removing sensitive data
18:07
Like passwords, accounts etc
18:07
<section1>
thx guys
18:07
<alkisg>
np
18:07
<section1>
nice
18:17
ok now i booted and i can login fine in ubu
18:17
for example i open firefox i see it running in the client output of ps aux
18:17
that is because is running in the client ?
18:18
<mwalters>
yes
18:18
<section1>
maybe because its detect that have much ram ?
18:18
how i disable that so i can test the apps on the server
18:18
<mwalters>
!thin
18:18
<ltsp>
I do not know about 'thin', but I do know about these similar topics: 'thin-clients-bandwidth'
18:19
<mwalters>
I think... set LTSP_FATCLIENT=False in lts.conf?
18:19
<section1>
ok i gogole that
18:19
thanks mwalters
18:19
<mwalters>
np :)
18:19
most settings will be in lts.conf, the man page is very long, but helpful: `man lts.conf`
18:20
most ltsp specific settings, anyways ;)
18:21
note: even when we used thin clients back on 14.04, we still ran the browser on the client. Anything that redraws frequently will run pretty poorly unless run on the client
18:24
<section1>
yeah
18:24
<alkisg>
section1: if your client is i3, why would you want to run apps on the server?!
18:24
<section1>
i want to probe to my boss that running browser in the server don't is a good idea
18:24
<alkisg>
prove, ok
18:24
<section1>
no this test box is an i3
18:25
<alkisg>
!flash
18:25
<ltsp>
flash: Yes, flash sucks. An HD full screen 30 fps video needs 2.5 Gbps bandwidth (1920×1080×4×30)! Make sure you have LDM_DIRECTX=True in your lts.conf file, or if it's just youtube you're after, try some flash replacing plugin like http://linterna-magica.nongnu.org
18:25
<section1>
the real thin client i don't know the hw will have...sorry about my english
18:25
<alkisg>
Tell him that he'll need 2.5 gbps per client :P
18:25
!cheap-client
18:25
<ltsp>
cheap-client: https://www.gearbest.com/tv-box-c_11262/?attr=2081-1279
18:25
<section1>
alkisg, lol
18:25
<alkisg>
Even cheap clients like that are better in fat mode than thin clients
18:25
<mwalters>
hey, 10gbe switches are... still super expensive, just kidding ;)
18:26
<section1>
yeah
18:26
spend all the money in switch and server hardware...non sense
18:26
<alkisg>
Thin clients will be fine again in 10 years or so, when google stadia will have convinced nvidia to give us hardware encoding of multiple hdmi streams :D
18:27
Even with 10 gbps switch and a monster server, fat clients still go better for normal tasks like web surfing
18:27
Of course thin clients do still have uses, but they're not appropriate for normal use
18:39vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
18:49* alkisg managed to get go-md2man to work properly with option and indentation etc; the solution was that "a newer go-md2man package is needed" :D https://github.com/cpuguy83/go-md2man/issues/27#issuecomment-515856608
18:50
<mwalters>
lol
18:50
debian and ubuntu packages out of date? no way! ;)
18:52
<alkisg>
Oh well we'll have lower-quality ltsp5 and ltsp19 man pages for a couple of more years
18:52statler has left IRC (statler!~Georg@gwrz.lohn24.de, Remote host closed the connection)
19:11
<section1>
this is crazy but if a shutdown ubuntu client shutdows the server :D
19:11
<alkisg>
section1: thin client? which desktop environment?
19:12
<section1>
yes thin client..and gnome
19:12
LTSP_FATCLIENT=False the only option on the lts
19:12
conf
19:12
<alkisg>
You can configure it not to allow remote clients to shut down the server
19:12
other environments have it as the default
19:13
Gnome had its own thoughts :)
19:13
There's an option in policykit about it
19:13
<mwalters>
oh, lol
19:16
<vagrantc>
using xdmcp or what?
19:16
<alkisg>
Normal LTSP thin clients
19:17
<vagrantc>
i don't see how it could even shut down the server
19:17
<alkisg>
GNOME UI => select shutdown
19:17
systemd picks up the user request, shut downs the server
19:17
*shuts down
19:18
<section1>
yeah
19:18
<alkisg>
Thin client, uses ldm, logs in to gnome, the desktop runs on the server; even opening a terminal and running "shutdown" without root shuts down the server
19:19
policykit there has options to allow remote or only local users to do that
19:19
So it needs to be enabled only for local users
19:20
<section1>
and the thin clients how is supposed to shutdown when the person don't need it anymore ?
19:20
btw
19:20
<alkisg>
In the mate that the page proposes, shutdown logs out the thin client, then he can shut down from LDM
19:20
<section1>
the thin client if i open somethin i get a black window
19:21
<vagrantc>
oh ... something used to block that from working...
19:21* alkisg really stopped caring about thin clients :) xfreerdp/x2go should be more than enough for this.
19:22
<alkisg>
sunweaver even adds upstream support to mate for things like that, over x2go
19:24
BTW, in ltsp19 I'm defaulting to --cleanup=1 in ltsp-update-image, does anyone think this is a bad idea?
19:24
<vagrantc>
not i!
19:25
<alkisg>
I.e. if someone is using a VM as the source, the users will be omitted from squashfs; but if someone adds a user in a chroot on purpose, he'd need to pass --cleanup=0
19:25
Great :)
19:25
<vagrantc>
the list of excluded files is configurable?
19:25
<alkisg>
Yes, but it's a regex now
19:25
<vagrantc>
aslong as there's a way
19:25
<alkisg>
/etc/ltsp/image.excludes was wildcard based, now it's regex, to allow us e.g. to avoid all the logs.gz.[0-9]
19:28Faith has joined IRC (Faith!~Paty_@unaffiliated/faith)
19:36
<alkisg>
Does it make sense to allow ltsp-update-image --revert (or) --revert=1 (or) --revert=0 ? Or those options that default to false, should not allow =x?
19:36
While e.g. now that --cleanup is the default, it's required to allow: --cleanup=0
19:42* vagrantc never liked 0/1 for true/false because of shell vs. all other languages basically being opposite
19:43
<alkisg>
If one starts the true/false things though, he gets into yes/no, on/off, and people even start asking for localization of these :D
19:43
And code gets complicated; other packages sourcing our .confs don't always have functions similar to "boolean_is_false"
19:44
<vagrantc>
but i don't think the answer is to use numeric booleans
19:45* vagrantc is partial to true/false exactly and no case variations and ... no localization
19:45
<alkisg>
systemd uses yes/no in its configs
19:46
<vagrantc>
that's fine by me
19:47
<alkisg>
And if a user misstypes "Yes", we consider it "no"? :)
19:50
https://yaml.org/type/bool.html
19:50
<vagrantc>
ideally, you'd error out
19:50
never liked implicit booleans
19:50
<alkisg>
Regexp: y|Y|yes|Yes|YES|n|N|no|No|NO |true|True|TRUE|false|False|FALSE |on|On|ON|off|Off|OFF
19:51
<vagrantc>
as far as i'm concerned, yes != Yes != "i like feta"
19:52
<alkisg>
That's why I preferred 0/1, I can't imagine someone reading "-b, --backup=0|1" and thinking "what does 0 there mean?"
19:52
<vagrantc>
0 means success in shell.
19:53
<alkisg>
Error codes are not boolean
19:53
It's an exit code, an integer
19:53
<vagrantc>
when i see 0 and 1 in configuration files, i always have to look up the documentation
19:53
because i've seen 0 mean both yes and i've also seen it mean no
19:54
does systemd support all the above cases?
19:54
<alkisg>
I think so
19:54
<vagrantc>
hrm.
19:54
<alkisg>
xorg supported even more
19:54
including 0/1
19:57
<section1>
i can say that running the thin client with default ubuntu desktop(gnome) don't work ..i get black windows..maybe something with render or don't know but i try and install xfce and works slow but works.
19:57
thanks guys for the help...very awesome channel
19:58
<alkisg>
yw
20:05
<section1>
byes
20:05section1 has left IRC (section1!~section1@178.33.109.106, Quit: Leaving)
20:08woernie has left IRC (woernie!~werner@p57A0EC92.dip0.t-ipconnect.de, Remote host closed the connection)
20:32
<alkisg>
vagrantc: you mentioned you didn't like "applets", any alternative suggestions before it's ...too late? :D ltsp tools?
20:38
<vagrantc>
heh
20:38
modules?
20:38
subcommand?
20:40
<alkisg>
"the `ltsp image` subcommand generates a squashfs image..."
20:40
"the `ltsp image` module generates a squashfs image..."
20:40
"the `ltsp image` applet generates a squashfs image..."
20:40
I think a native speaker like you should decide, I don't have the "feeling" to do it
20:41
<vagrantc>
not that git is to be emulated for user interface issues, but it just calls them "commands"
20:42
<alkisg>
git has the benefit that it's actually using verbs there, while in our case, they're not
20:42
<vagrantc>
some others call them subcommands
20:42
<alkisg>
git commit => is a verb, ltsp ipxe => isn't
20:42Faith has left IRC (Faith!~Paty_@unaffiliated/faith, Quit: Leaving)
20:42
<alkisg>
Can we call them commands if they're not verbs? Then ok, commands sounds find to me...
20:43
_COMMAND="ipxe" in the code though, sounds like something that we want to execute, not the name of the running "applet"...
20:44
<vagrantc>
subcommand is also used by a few things, it's just a bit long
20:44
<alkisg>
Works for me...
20:44
<vagrantc>
since these are all commandline commands, i treat the whole thing "ltsp image" as a single command in my mind
20:45
it's just an implementation detail of PATH ... it could just as easily be ltsp-image and ltsp-ipxe
20:46
<alkisg>
Actually, ltsp has its own man page and its own options, and each subcommand its own man page and its own options, e.g.
20:47
ltsp --base-dir=/opt/ltsp --overwrite dnsmasq --dns=1 --proxy-dhcp=0
20:47
The "overwrite" there is a global flag, it can't go after "dnsmasq"
20:47
And it only appears in the ltsp man page
20:47
<vagrantc>
sure
20:48
that's fine; i'm not suggesting to change that ... just suggesting about avoiding getting too caught up in the terminology around the verb-fulness of "command" vs. "subcommand" etc.
20:49
<alkisg>
Gotcha. I'll see if changing them all to "commands" is readable enough...
20:52
<vagrantc>
sometimes non-native speakers have a better eye for what makes sense, too, since native speakers often just fuzzy match without thinking too hard
20:56ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
21:31
<alkisg>
The manpages look more beautiful when read via `man`, than online as .md: https://github.com/eellak/gsoc2019-ltsp/blob/master/man/ltsp-image.8.md
21:32
Can someone try to read the "IMAGE TYPES" and "VM IMAGES" sections in the URL above, and tell me if he can understand how to use virtual machines as ltsp sources?
21:44GodFather has left IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com, Ping timeout: 248 seconds)
21:44GodFather_ has left IRC (GodFather_!~rcc@d53-64-7-141.nap.wideopenwest.com, Ping timeout: 244 seconds)
21:52GodFather_ has joined IRC (GodFather_!~rcc@d53-64-7-141.nap.wideopenwest.com)
21:53GodFather has joined IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com)
22:28GodFather has left IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com, Ping timeout: 258 seconds)
22:28GodFather_ has left IRC (GodFather_!~rcc@d53-64-7-141.nap.wideopenwest.com, Ping timeout: 245 seconds)
22:55GodFather_ has joined IRC (GodFather_!~rcc@d53-64-7-141.nap.wideopenwest.com)
22:55GodFather has joined IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com)