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 25 October 2019   (all times are UTC)

05:20ltsp has joined IRC (ltsp!bot@ltsp.org)
05:20
<alkisg>
!bot are you alive now? :)
05:20
<ltsp>
Error: "bot" is not a valid command.
05:34
<alkisg>
Let's see if restoring the ltsp5 wiki to a two-years-old backup will work... before we dis-bundle it :P
05:35
!learn html-to-markdown as HTML to markdown converter: https://euangoddard.github.io/clipboard2markdown/
05:35
<ltsp>
The operation succeeded.
05:40os_a has left IRC (os_a!~Thunderbi@195.112.116.22, Quit: os_a)
05:54os_a has joined IRC (os_a!~Thunderbi@195.112.116.22)
06:04woernie has joined IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de)
06:05
<os_a>
Hi! Does LTSP work with UEFI terminals?
06:05
<alkisg>
os_a: the new ltsp, yes
06:05
The old one, no
06:05
!install
06:05
<ltsp>
install: To install LTSP19+: https://github.com/ltsp/ltsp/wiki/installation. To install LTSP5: http://wiki.ltsp.org/wiki/Installation/Ubuntu for Ubuntu, or http://wiki.ltsp.org/wiki/Installation for other distributions
06:05
<alkisg>
!forget install
06:05
<ltsp>
The operation succeeded.
06:06
<alkisg>
!learn install as To install LTSP19+: https://ltsp.github.io/docs/installation. To install LTSP5: http://wiki.ltsp.org/wiki/Installation/Ubuntu for Ubuntu, or http://wiki.ltsp.org/wiki/Installation for other distributions
06:06
<ltsp>
The operation succeeded.
06:07
<os_a>
ok, merci!
06:10ltsp has joined IRC (ltsp!bot@ltsp.org)
06:19ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
06:33vsuojanen has joined IRC (vsuojanen!~vsuojanen@cable-hml-585686-205.dhcp.inet.fi)
06:39georgeneophytou has joined IRC (georgeneophytou!~georgeneo@217.27.33.213)
06:42kjackal has joined IRC (kjackal!~quassel@2a02:587:3107:2e00:e50f:a11:3274:b4f9)
06:45
<georgeneophytou>
morning, alkisg, I have provisioned a new LTSP but when generating the image, initrd etc, I get the following error: Could not locate vmlinuz and initrd.img in /srv/ltsp/images/x86_64.img
07:07
<alkisg>
Hi georgeneophytou
07:08
Does it have a separate boot partition?
07:08
If so, check this out: https://github.com/ltsp/ltsp/issues/43
07:09
https://github.com/ltsp/ltsp/issues/43#issuecomment-541388632
07:27
<georgeneophytou>
yes it does
07:27
I will check it thanks
07:37
<alkisg>
!learn stargazers as If you're using LTSP, star it on github: https://github.com/ltsp/ltsp/stargazers
07:37
<ltsp>
The operation succeeded.
07:38
<georgeneophytou>
I am very confused what this command is doing:
07:38
mount --bind / /mnt
07:38
cp -a /boot/. /mnt/boot/
07:38
it's copying from and to the same place of the filesystem?
07:39
can you explain this a bit alkisg?
07:40
"copy the contents of the /boot partition to the /boot folder of the / partition, before calling ltsp image /"
07:40
<alkisg>
georgeneophytou: if you have a file /boot/something, and then you mount something over /boot, you can't see the file /boot/something under the mount
07:40
<georgeneophytou>
there are already all the files in /boot
07:40
<alkisg>
So we want to copy the boot files from the boot partition to the root partition
07:40
Your boot partition shows in /boot
07:40
Your root partition in /
07:40
But you can't copy from /boot to /boot
07:41
So by bind-mounting, you get another VIEW of /, in /mnt
07:41
So then the /mnt/boot folder is the contents of the boot folder of the REAL root, not of the boot partition
07:41
<georgeneophytou>
ohhhhhhhhh
07:41
alright, wow, now I get it, thanks! :D
07:41
<alkisg>
:)
07:42
I hope I'll have some time to fix that in a few days
07:42
<georgeneophytou>
that really hurt my brain haha
07:42
<alkisg>
So that you won't need the workaround
07:42
Haha
07:42
<georgeneophytou>
thanks that would be great :)
07:42
<alkisg>
Are you using chrootless?
07:42
<georgeneophytou>
yes
07:42
<alkisg>
OK
07:43
<georgeneophytou>
I have followed your advice and installed the OS to RAID5
07:43
I have a 4th drive just for the boot partition
07:43
3x3tb disks took almost 4 hours to recover :D
07:44
but, I had to install ubuntu server first and then ubuntu desktop after, the ubuntu desktop installer could not see a previously created raid5 array
07:45
<alkisg>
You just needed to apt install dmraid from the live cd
07:45
But ok this will also do
07:48
<georgeneophytou>
I wish I knew this yesterday... lol
07:48
you are a real guru alkisg :)
07:50
<alkisg>
Eh everyone is a guru in a small little area where he focuses his knowledge...
07:50
<georgeneophytou>
true true :)
07:50
are you also on the epoptes team?
07:52
<alkisg>
Yup, the only developer left there :/
07:52
*except for vagrant which takes care of uploading things to debian
07:54statler_ has joined IRC (statler_!~Georg@gwrz.lohn24.de)
07:54
<georgeneophytou>
oh wow, okay, because I have been very keen to get involved with epoptes
07:57
<alkisg>
You'd be very welcomed
07:59
<georgeneophytou>
alright, I was start looking into this
08:05
where can you suggest I write feature requests for epoptes?
08:06
<alkisg>
georgeneophytou: I hope you mean "implement" and not "ask" features, right? :D
08:06
https://epoptes.github.io/documentation/gsoc/
08:07
If you just want to ask features, you'd do it in the epoptes issue tracker
08:07
https://github.com/epoptes/epoptes/issues
08:07
You may also check the issues, find something that interests you, and provide a solution
08:07
(with or without code)
08:08
<georgeneophytou>
okay
08:20georgeneophytou has left IRC (georgeneophytou!~georgeneo@217.27.33.213, Quit: My MacBook has gone to sleep. ZZZzzz…)
08:31georgeneophytou has joined IRC (georgeneophytou!~georgeneo@217.27.33.213)
09:25
<meo>
so I brought my first ltsp19 setup up and its working
09:25
v. nice
09:26
<alkisg>
Yey! ;)
09:27
Btw if anyone knows Jekyll and wants to help with putting irclogs to github: https://github.com/ltsp/irclogs/tree/master
10:37statler has left IRC (statler!~Georg@p5B30EE2F.dip0.t-ipconnect.de, Remote host closed the connection)
10:46os_a has left IRC (os_a!~Thunderbi@195.112.116.22, Quit: os_a)
11:50Faith has joined IRC (Faith!~Paty_@unaffiliated/faith)
12:05
<georgeneophytou>
alkisg, is it possible at all for clients to change their password from the client machine?
12:20
<alkisg>
georgeneophytou: currently, no,
12:20
ltsp5 had a "remote apps" feature, that one could use to run a password changing app on the server, and have it displayed locally,
12:20
that can't possibly work with wayland, as ssh -X doesn't work on wayland,
12:21
so while we can reimplement remoteapps, or even just write a small script that uses passwordless ssh to substitute remoteapps, it'll only work on xorg,
12:24
...to sum up: (1) in the long future, ssh -X or something else will work with both wayland and xorg; we'll have to wait for that one;
12:24
(2) now, one can create a small script to run remote apps on ltsp19,
12:24
and finally (3) one could write a GUI to change the app, which again would use ssh to change the pass on the server
12:24
*change the pass
12:24
But, no, no out of the box solution; unless you use ldap, webmin etc etc
12:32
<georgeneophytou>
alright, thanks for the answer!
12:34
<alkisg>
Hrm, better yet: since we already have pam hooks, we could use those to change the pass on the server, with the existing gui tools, without the user realizing what's happenning under the hood
12:34
That deserves a bug report
12:34section1 has joined IRC (section1!~section1@178.33.109.106)
13:01meo_ has joined IRC (meo_!~systemdju@unaffiliated/mikeseth)
13:02meo has left IRC (meo!~systemdju@unaffiliated/mikeseth, Ping timeout: 265 seconds)
13:23os_a has joined IRC (os_a!~Thunderbi@195.112.116.22)
13:57os_a has left IRC (os_a!~Thunderbi@195.112.116.22, Quit: os_a)
14:08GodFather has joined IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com)
14:13
<georgeneophytou>
I have another suggestion alkisg, currently, it is possible for the same user to login to multiple clients at the same time, perhaps add a features to terminate all but the newest session
14:16
<alkisg>
georgeneophytou: what would "terminate" mean over nfs?
14:16
There are no processes on the server
14:17
<georgeneophytou>
if a user logs into client A, and then also on client B, we need to terminate the session on client A
14:17
<alkisg>
georgeneophytou: I mean, how would the server know?
14:18
There's no persistent ssh process like in ltsp5
14:18
The server is only used for authentication, or sshfs
14:18
So if one uses nfs, the server doesn't know at all if a client is logged in in A
14:18
And it's also possible that the user would want that; many apps support it
14:20
<georgeneophytou>
sure they could want it, however, in my use case, business environment, a user should only be logged into 1 client at a time. its not really a major issue, I just think that in the case where the user locks their session and a different user logs in on the same client, then we have more and more users logged in on the same client. happens a lot here at the office where employees move around the office
14:21
<alkisg>
georgeneophytou: and what would be the correct way out of it? To kill the user processes without allowing the user to save them?
14:21
<georgeneophytou>
well :) it would be their fault for not saving, computers are rebooted once per day anyways
14:22
<alkisg>
Then wouldn't it be better to just not allow multiple logins?
14:22
And ask the next user, "do you want to terminate the old user processes?"
14:22
<georgeneophytou>
yes, but how does one go about asking that question?
14:23
as we said, the server is not aware, so somehow have to make the clients aware
14:23
<alkisg>
Much easier than terminating processes from *another* pc, that's for sure :)
14:23
I'm thinking of implementing an "ltsp session" script; one that would run right after log in
14:24
So at that point, you could run a little script, that would show that dialog, and if the user presses yes, it'd terminate the old user processes
14:24
<georgeneophytou>
i thought of using session files, ie, when a user logs into the client, a session file is created with a timestamp, if said user logs into another client, this new session file is detected from the old client and issues a termination command on the old client
14:24
well there you go, we thought the same :)
14:24
more or less..
14:25
ideally, the new client session should receive a message "you are already logged into $hostname, do you want to terminate that session?" if yes, terminate old session, if no, terminate new session
14:26
<alkisg>
I'm talking about a different thing
14:26
user1 logs in to client1
14:26
user2 tries to login to client1
14:26
He's asked: terminate processes from user1?
14:27
While, if user1 logs in to client2, there wouldn't be any notifications
14:27
As there's no easy way to know that user1 is logged in to client1
14:27
Suppose client1 hard crashes; even lock files would be in use for a looong time
14:28
So user1 would need to wait for that long time, to be able to login to the next client
14:28
<georgeneophytou>
oh no no, I dont think user2 should be able to terminate user1's session, i was thinking more like this:
14:28
<alkisg>
I'm talking about what can be done
14:28
You're talking about what you'd like :)
14:28
Which isn't easy, technically
14:28
<georgeneophytou>
user1 logs into client1, session file is created on user1's /home
14:29
user1 then logs into client2 and is asked the question because the session file is detected
14:29
<alkisg>
And he says "yes, log me out". What happens then?
14:29
How can the server affect processes running in client1?
14:29
It doesn't have reverse ssh
14:30
What you're saying is possible, but with a huge framework that coordinates everything
14:30
It can't be done with little scripts, like ltsp is
14:30
<georgeneophytou>
exactly, they will have to return to client1 and either log out first to be able to use client2, or continue working on client1
14:30
i see
14:30
<alkisg>
And what if client1 has hanged?
14:30
Again the session file would be there
14:30
<georgeneophytou>
well you're right
14:30
session files can be bad
14:30
unless
14:31
<alkisg>
Lock files are supported by the kernel
14:31
They're a bit better; some ever work over nfs
14:31
But again they're not really reliable
14:31
So there would have to be constant client<=>server communication for this to work
14:31
Like, an ltsp daemon network process
14:31
Which is not what ltsp is, currently
14:31
<georgeneophytou>
yes
14:31
true
14:32
<alkisg>
Mayyybeeee in 10 years, ltsp can grow network daemons etc; but atm, epoptes is closer to that, than ltsp
14:32
<georgeneophytou>
well....we could use some epoptes functionality here, as it already has all the requirements ready
14:32
for example, sending messages to clients, viewing sessions etc
14:33
<alkisg>
Currently epoptes-client can't notify the server though; only the server can request things
14:33
So that doesn't really fit the requirements either
14:33
You could implement a separate application, that registers log ins to a network server
14:33
and does all that you ask, unrelated to ltsp or epoptes
14:34
For now, what ltsp could do easily, is what I described initially; user2 would kill locked user1's processes
14:34
<georgeneophytou>
do you think this is something worth looking into? I mean, I could look into it
14:34
would user2 have privileges for that though?
14:34
<alkisg>
I think there are other tasks that are more significant for the majority of users; e.g. the one you said about password
14:35
You can ping ltsp to do that; or use a setuid binary
14:35
<georgeneophytou>
could you explain?
14:36
<alkisg>
E.g. /usr/sbin/terminate-other-users can be a script that accepts no parameters and just terminates processes from all other users (non system ones)
14:37
<georgeneophytou>
that is one way yes
14:40
anways, I will think about this over the weekend, time to go home now, have a good weekend and thanks for all the help
14:40
<alkisg>
Bye, have a nice weekend too
14:43georgeneophytou has left IRC (georgeneophytou!~georgeneo@217.27.33.213, Quit: My MacBook has gone to sleep. ZZZzzz…)
15:19kjackal has left IRC (kjackal!~quassel@2a02:587:3107:2e00:e50f:a11:3274:b4f9, Remote host closed the connection)
15:24meo_ has left IRC (meo_!~systemdju@unaffiliated/mikeseth, Quit: Changing server)
15:24meo has joined IRC (meo!~systemdju@85.196.137.173)
15:24meo has joined IRC (meo!~systemdju@unaffiliated/mikeseth)
17:09Ark74 has left IRC (Ark74!~Luis@177.238.145.140, Ping timeout: 240 seconds)
17:41statler_ has left IRC (statler_!~Georg@gwrz.lohn24.de, Remote host closed the connection)
17:59woernie has left IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de, Remote host closed the connection)
18:34statler has joined IRC (statler!~Georg@p5B30EE2F.dip0.t-ipconnect.de)
19:36Ark74 has joined IRC (Ark74!~Luis@177.238.145.140)
19:54section1 has left IRC (section1!~section1@178.33.109.106, Quit: Leaving)
20:39Faith has left IRC (Faith!~Paty_@unaffiliated/faith, Quit: Leaving)
21:10ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
23:13jgee has left IRC (jgee!~jgee@190.159.118.121, Remote host closed the connection)
23:23jgee has joined IRC (jgee!~jgee@190.159.118.121)