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


Channel log from 16 August 2019   (all times are UTC)

03:33pppingme has left IRC (pppingme!~pppingme@unaffiliated/pppingme, Ping timeout: 258 seconds)
03:36pppingme has joined IRC (pppingme!~pppingme@unaffiliated/pppingme)
06:39kjackal has joined IRC (kjackal!~quassel@adsl-221.109.242.68.tellas.gr)
08:04statler has joined IRC (statler!~Georg@gwrz3.lohn24.de)
08:59
<alkisg>
!root
08:59
<ltsp>
I do not know about 'root', but I do know about these similar topics: 'ROOT_PASSWORD_HASH'
08:59
<alkisg>
!ROOT_PASSWORD_HASH
08:59
<ltsp>
ROOT_PASSWORD_HASH: To be able to login as root in vt1, with password "qwer1234", put this in lts.conf: INIT_COMMAND_ROOT_HASH="sed 's!^root:[^:]*:!root:\$6\$p2LdWE6j\$PDd1TUzGvvIkj9SE8wbw1gA/MD66tHHlStqi1.qyv860oK47UnKcafSKqGp7cbgZUPlgyPv6giCVyCSCdJt1b0:!' -i /etc/shadow"
10:16ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
10:51kjackal has left IRC (kjackal!~quassel@adsl-221.109.242.68.tellas.gr, Ping timeout: 272 seconds)
11:14kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-77.home.otenet.gr)
11:52gbaman has joined IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
11:54
<gbaman>
So am slowly getting somewhere with this new LTSP and PiNet. Now got it booting and logging in. Oddly something has broken *all* the icons with the OS and login screen though...
11:55
Am I right in saying alkisg that the usernames available are actually contained within the ltsp.img file though?
12:10section1 has joined IRC (section1!~section1@178.33.109.106)
13:07gbaman has left IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com, Remote host closed the connection)
13:07gbaman has joined IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
13:12gbaman has left IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com, Ping timeout: 268 seconds)
13:23gbaman has joined IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
13:25
<alkisg>
gbaman: yes, `ltsp initrd` copies the server's /etc/passwd and group into ltsp.img, which then goes to the client is is merged there
13:26
At that point, the PWMERGE_* ltsp.conf parameters can control which users will get migrated and which not
13:27
(in case someone wants to limit some clients to only a few users)
13:29bengoa has left IRC (bengoa!~alberto@194.50.55.200, Remote host closed the connection)
13:30kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-77.home.otenet.gr, Remote host closed the connection)
13:32kjackal has joined IRC (kjackal!~quassel@ppp-94-66-220-77.home.otenet.gr)
14:55* alkisg just uploaded a new builds in the PPA that migrates most of the old init-ltsp.d
15:18
<gbaman>
Yeah, so it isn't really going to be feasible to have the ltsp.img file on the SD card then
15:21
<alkisg>
gbaman: how so?
15:21
It's the same as the kernel/initrd
15:22
<gbaman>
Yeah, but it needs updated every time a new user is added
15:22
<alkisg>
Yup; that can be handled either on boot, like KERNEL_DEVICE, or on login
15:22
<gbaman>
(where as kernel/initrds only need updated every time a new one is out, which is more like every 2-3 months)
15:22
<alkisg>
The tftp dir is exported over nfs
15:22
Synching ltsp.img from there would take like 0.1 secs
15:23
So it can be done in any phase of the boot process
15:23
<gbaman>
Yeah, I guess but then a fiddle to get that merged in without doing a reboot
15:23
<alkisg>
It's just one pwmerge command
15:23
See ./client/init/55-pam.sh
15:25
Having the users geometry on the clients was necessary for pamltsp to work at all
15:25
<gbaman>
What benefits are there doing it this way vs previous way?
15:25
<alkisg>
PAM
15:25
<gbaman>
Ahh
15:25
<alkisg>
Otherwise we needed our own DM, which broke all over the place
15:25
<gbaman>
I guess the real answer is not having stuff on the card :)
15:25
<alkisg>
And didn't list users and didn't allow passwordless logins etc etc
15:25
No
15:26
<gbaman>
(well, bar the bootcode.bin)
15:26
<alkisg>
Updating the users list should be supported by diskless clients too
15:26
And updating ltsp.conf as well
15:26
And, there are non-raspi cases where local kernel/initrd is also nice to have
15:26
E.g. NICs with weird drivers that ipxe doesn't support, or booting 32bit os under uefi etc
15:27
<gbaman>
Perhaps, but for my case, I should really be going down the bootcode.bin route
15:27
<alkisg>
There ARE some parts thought that I postpone for after beta, like this ^ user live-resynching, the swap (which I think crashes clients with NBD and hurts the server), jetpipe which might be replaced etc
15:28
Other than those hard parts, the rest should be done within the next week
15:28
<gbaman>
Still doesn't get around the annoying fact the Pi doesn't nicely handle 2 initrds
15:28
<alkisg>
cat initrd.img ltsp.img > initrd-new.img
15:28
That should be all; in the synching process
15:29
<gbaman>
Yeah, but that defeats the entire point of keeping it split :)
15:29
I could just nicely ask RPF to add support for it
15:29
<alkisg>
Of course, that would be ideal
15:29
I think that new ltsp.img design is very very efficient
15:30
Now in the development, I was able to test and upgrade and debug things 5 times more efficiently than with LTSP5
15:30
Literally `ltsp initrd`; netboot any VM/iso
15:30
<gbaman>
Speed wise?
15:30
<alkisg>
Without having to ltsp-update-image 10 different isos for each little ltsp client code change
15:30
<gbaman>
Yeah, does make sense
15:31
<alkisg>
E.g. "did I correctly disable autologin in gdm3? let's test those 5 VMs" ==> is 5 minutes with the new LTSP, and like 2 hours with LTSP5
15:34
And it's not only speed wise, now the code tell me itself in which part broke, as I'm using some re (run-or-exit, but report what failed) shell functions
15:34
<gbaman>
Oooo, fancy
15:56
Soo here is another interesting one, after further digging, my test machine here seems unhappy with mime data. Potentially unrelated to LTSP. But if I log in, run "update-mime-database /usr/share/mime", then on re-logging in, it is fine. Reboot obviously results in it being lost and seems to make no difference if run in the chroot itself...
15:56
(all icons are dead otherwise)
15:57
Any thoughts alkisg?
15:59vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
16:02mmarconm has joined IRC (mmarconm!~mmarconm@unaffiliated/mmarconm)
16:11gbaman has left IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com, Read error: Connection reset by peer)
16:11mmarconm has left IRC (mmarconm!~mmarconm@unaffiliated/mmarconm, Quit: Leaving)
16:12gbaman has joined IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
16:44kjackal has left IRC (kjackal!~quassel@ppp-94-66-220-77.home.otenet.gr, Quit: No Ping reply in 180 seconds.)
16:46kjackal has joined IRC (kjackal!~quassel@62.74.20.190)
16:49kjackal has left IRC (kjackal!~quassel@62.74.20.190, Remote host closed the connection)
16:54gbaman has left IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com, Remote host closed the connection)
16:55gbaman has joined IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
16:59gbaman has left IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com, Ping timeout: 258 seconds)
17:13statler has left IRC (statler!~Georg@gwrz3.lohn24.de, Remote host closed the connection)
17:27gbaman has joined IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com)
17:35
<alkisg>
gbaman: sounds unrelated to ltsp, yeah
17:36
Try to boot the chroot with kvm, and see if the problem is there without ltsp at all
17:42spectra has left IRC (spectra!~spectra@debian/developer/spectra, Ping timeout: 248 seconds)
17:48
<alkisg>
!kvm
17:48
<ltsp>
kvm: Virtual thin client: kvm -m 256 -vga vmware -ctrl-grab -no-shutdown -net nic,model=virtio -net user,tftp=/var/lib/tftpboot,bootfile=/ltsp/i386/pxelinux.0
17:48
<alkisg>
Hmm no, not that
17:48
!kv
17:48
<ltsp>
Error: "kv" is not a valid command.
17:48
<alkisg>
!nfs
17:48
<ltsp>
nfs: to enable NFS home directories for localapps and fat clients, install nfs-kernel-server on your server, nfs-common on your client (don't forget ltsp-update-image), and put this in lts.conf: FSTAB_1="server:/home /home nfs defaults,nolock 0 0"
17:48
<alkisg>
Meh, nevermind; i mean, boot it with kvm and nfs rw root
17:49
(or directly from an rpi client with nfs rw root)
17:55
<gbaman>
Beyond that, I now have it at the point where it is basically working...
17:55
(on all versions of Pis, including Pi 4)
17:56
Of course the code needs written to reproduce it all, but the proof of concept works
18:04spectra has joined IRC (spectra!~spectra@debian/developer/spectra)
18:50shored has left IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi, Read error: Connection reset by peer)
18:52shored has joined IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi)
20:05section1 has left IRC (section1!~section1@178.33.109.106, Quit: Leaving)
20:35
<alkisg>
gbaman: btw if you only want to make a couple of users available, the ltsp.conf parameter is: PWMERGE_SUR="^(administrator|ltsp)"
20:35
SUR means "source users (passwd) regexp"
22:23adrianor1 has joined IRC (adrianor1!~adrianorg@186.215.20.35)
22:26adrianorg has left IRC (adrianorg!~adrianorg@179.187.25.90.dynamic.adsl.gvt.net.br, Ping timeout: 248 seconds)
22:43ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
22:48
<gbaman>
Interesting, probably for simplicity sake, probably going to have them all enabled alkisg
23:40vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
23:54gbaman has left IRC (gbaman!~gbaman@host81-142-46-233.in-addr.btopenworld.com, )