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

Request log from specific day:

Channel log from 17 November 2023 (all times are UTC)

00:34	vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20, Quit: leaving)
02:22	XYZ_ has joined IRC (XYZ_!~XYZ@37-48-1-27.nat.epc.tmcz.cz)
04:49	fiesh has left IRC (fiesh!~fiesh@2003:fb:1018::21, Server closed connection)
04:50	fiesh has joined IRC (fiesh!~fiesh@2003:fb:1018::21)
05:37	hark has left IRC (hark!~hark@94.142.244.66, Server closed connection)
05:37	hark has joined IRC (hark!~hark@94.142.244.66)
06:29	book` has left IRC (book`!~book`@static.38.6.217.95.clients.your-server.de, Quit: Leaving)
06:38	book` has joined IRC (book`!~book`@static.38.6.217.95.clients.your-server.de)
06:58	ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
08:16	alkisg has left IRC (alkisg!~Thunderbi@2a02:587:7441:6200:56e0:4cff:fe68:d, Server closed connection)
08:16	alkisg has joined IRC (alkisg!~Thunderbi@2a02:587:7441:6200:56e0:4cff:fe68:d)
08:40	woernie has joined IRC (woernie!~werner@2001:9e8:e25c:3c00:7872:6d1f:a035:b94a)
08:47	XYZ_ has left IRC (XYZ_!~XYZ@37-48-1-27.nat.epc.tmcz.cz, Remote host closed the connection)
08:51	XYZ_ has joined IRC (XYZ_!~XYZ@37-48-1-27.nat.epc.tmcz.cz)
08:58	woernie has left IRC (woernie!~werner@2001:9e8:e25c:3c00:7872:6d1f:a035:b94a, Ping timeout: 260 seconds)
12:29	XYZ_ has left IRC (XYZ_!~XYZ@37-48-1-27.nat.epc.tmcz.cz, Ping timeout: 246 seconds)
12:42	XYZ_ has joined IRC (XYZ_!~XYZ@78-80-96-74.customers.tmcz.cz)
12:43	wyre is back
12:51	wyre is now away: Auto away at Fri Nov 17 12:51:24 2023 UTC
12:58	wyre is back
13:13	wyre is now away: Auto away at Fri Nov 17 13:12:49 2023 UTC
13:39	woernie has joined IRC (woernie!~werner@2001:9e8:e25c:3c00:7872:6d1f:a035:b94a)
15:37	XYZ_ has left IRC (XYZ_!~XYZ@78-80-96-74.customers.tmcz.cz, Read error: Connection reset by peer)
15:41	XYZ_ has joined IRC (XYZ_!~XYZ@78-80-96-74.customers.tmcz.cz)
16:01	Anon443 has joined IRC (Anon443!~Anon443@50.231.145.86)
16:06	woernie has left IRC (woernie!~werner@2001:9e8:e25c:3c00:7872:6d1f:a035:b94a, Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
16:07	<Anon443> Two questions about ltsp.conf
16:07	If I have 2 lines I need to place in the lightdm.conf do I have 2 LIGHTDM_CONF lines or do I put both lines in a single LIGHTDM_CONF entry?
16:07	Second question. For IMAGE_TO_RAM does it stay compressed on the Client or should I expect the RAM utilization to be the fully decompressed size of the squashfs image?
16:10	<quinox> it's a bunch of Bash variables under water, so setting the same thing twice will only make the last one stick
16:11	some of the variables are really prefixes, allowing you to set VARNAME_01 VARNAME_02 etc. which does what you expect
16:13	about the RAM usage: mmm I don't know, easy enough to test. I think what happens: Linux mounts the squashfs, that option triggers a "look at all bytes on the filesystem" which effectively loads the image into RAM. Thus, I suppose it laods the squashed bits into memory, not the decompressed ones
16:13	<Anon443> I don't think I saw that logic "_01, _02, etc" in the underling script for the LIGHTDM_CONF entry but may have missed it. I can test but only have limited window to do so when it wont impact others. Was just hoping to have a good idea before my window opened up and save some time.
16:14	<quinox> yeah if the manual now specifies LIGHTDM_CONF I don't think it uses the _01 _02 setup
16:15	CRONTAB_x <-- that one is a prefix fe.
16:15	<Anon443> yeah
16:20	<quinox> the LIGHTDM_CONF variable does support files, so if the configuration can't parse "option1=val1; option=val2" you can put it in a file insteead
16:24	<Anon443> Interesting... I'll have to play with that.
16:25	<quinox> for testing during office hours I modify my dnsmasq.conf to point my custom Mac address to a different boot image and then boot a VirtualBox with an ipxe.iso file using that Mac address
16:25	it can't test stuff like nvidia videocard drivers but the rest is all the same
16:26	<Anon443> Thanks for the hint there save me some late nights.
16:27	<quinox> another way: modify the ipxe menu to add another option. might be nicer than dnsmasq if you don't mind people seeing your play stuff
16:28	last tip: you can boot from a raw disk image directly, no need to build an image first
16:30	<Anon443> Oh that one is actually huge. Save a lot of time processing it for testing.
16:31	<quinox> for sure
16:32	I assume it's slower / less stable than using the squashfs image but so far I've never noticed any issues
16:34	ltsp.conf DEBUG_SHELL is also nice
17:03	<Anon443> Looks like IMAGE_TO_RAM=1 does keep it compressed. Which obviously takes a little longer during boot to transfer, unexpectedly though logging in with an ldap user, first login is very slow (like 1-2 minutes), home folder creation was my first thought since I mount home as tmpfs and has to be created fresh, however the home folder is only 6mb. Weird
17:03	that login is almost instant when IMAGE_TO_RAM=0
17:06	<alkisg> Anon443: variable="line1<literal line break>line2"
17:06	I.e. you hit enter in the middle of the string, without closing the quotes
17:07	<Anon443> Thank you!
17:07	<alkisg> IMAGE_TO_RAM keeps it compressed, yeah. First login should be <30seconds, if it's not, you have some other issue
17:07	The home creation is on sshfs/nfs; image_to_ram shouldn't be related to that
17:08	Big delays are some times connected to bad dns entries
17:08	You can set a /usr/share/xsession/*.desktop => Exec=xterm line, to login with xterm
17:08	Then you run `mate-session` or whatever else you want from that xterm
17:09	And you notice the output, the delays, the error messages et c
17:10	<Anon443> Yeah interesting that IMAGE_TO_RAM being the only change, affects user first login so much. Haven't started hunting that down, but that makes sense to see where the slow down is. Not a huge deal either way just not expected.
17:11	<alkisg> Which distro/version/desktop environment is that?
17:12	<Anon443> xubuntu 22.04.3 xfce4
17:12	<alkisg> Dunno, I haven't played much with XFCE
17:21	<Anon443> Doesn't seem the be the window manager. switched to tty1 (ctrl+alt+F1) cli login still fast on IMAGE_TO_RAM=0 but doesn't work (times out?) on 1 it just goes back to the login prompt. Switching back and forth with that being the only change. It isn't a big deal for me I don't mind running with it as 0 and if I needed it to be fully loaded in RAM I
17:21	can just tell the users to wait.
17:24	Side note. I have FSTAB_HOME="tmpfs /home tmpfs default,mode=0755 0 0" during login should I still see ".Pamltsp failed to mount home SSHFS:" during login? it does endup creating the directory fine on the tmpfs mount. Just wondering if that was expected behavior.
17:26	Saw that in regular ubuntu and xubuntu as well as cli login
17:27	<alkisg> You want ldap users and non-persistent home?
17:28	<Anon443> Yes that is what I am doing.
17:30	Basically a custom thinclient image.
17:31	<alkisg> After you boot a client, what are the contents of /usr/share/pam-configs/ltsp ?
17:33	Also is that with IMAGE_TO_RAM=1, where / is tmpfs?
17:36	Finally, try setting SSHFS=0 in ltsp.conf, under [clients]
17:37	<Anon443> Name: SSH based authentication for LTSP
17:37	Default: yes
17:37	Priority: 0
17:37	Auth-Type: Additional
17:37	Auth:
17:37	optional pam_exec.so expose_authtok seteuid stdout quiet /usr/share/ltsp/client/login/pamltsp pam_auth
17:37	Session_Interactive-Only: yes
17:37	Session-Type: Additional
17:37	Session:
17:37	optional pam_exec.so seteuid stdout quiet /usr/share/ltsp/client/login/pamltsp pam_session
17:38	Anon443 has left IRC (Anon443!~Anon443@50.231.145.86, Quit: Client closed)
17:39	Anon443 has joined IRC (Anon443!~Anon443@50.231.145.86)
17:39	<alkisg> ...use a pastebin :D
17:39	Otherwise libera kicks you out
17:39	<Anon443> lol
17:39	<alkisg> OK so it's Additional, i.e. LDAP was detected
17:39	Try setting SSHFS=0 in ltsp.conf, under [clients]
17:39	Then ltsp initrd, then reboot clients
17:40	<Anon443> I get the error with IMAGE_TO_RAM=0 (normal quick login) I can't tell for sure on the RAM=1 need the cli login to work to see the error.
17:45	Yes still get the same error. It does proceed past that error just fine.
17:49	Anon443 has left IRC (Anon443!~Anon443@50.231.145.86, Quit: Client closed)
17:49	<alkisg> Strange, oh well if it works no point in debugging
17:49	Anon443 has joined IRC (Anon443!~Anon443@50.231.145.86)
18:14	vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20)
18:17	<Anon443> With IMAGE_TO_RAM=1 logging in (slowly) through XFCE then inside a terminal running su - <user> (same user that is already logged in, so for sure not a home folder thing) it is still very slow and triggers the ".Pamltsp failed to mount home via SSHFS: ." then immediately gives the shell prompt.
18:21	Not a huge deal for what I am doing. Just wanted to provide the datapoint and see if I was missing something. Thank you for the suggestions and information. As well as the software to begin with. I am going to step away for a while.
18:28	One last datapoint on the slow login with IMAGE_TO_RAM=1 a wrong ldap password fails fast but correct is slow to login.
19:19	vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20, Quit: leaving)
19:28	<alkisg> The IMAGE_TO_RAM code was added by someone to suit his own needs; personally I've no use for it so I haven't spent any time locating or fixing bugs related to it
19:32	Anon443 has left IRC (Anon443!~Anon443@50.231.145.86, Quit: Client closed)
19:37	vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20)
21:09	XYZ_ has left IRC (XYZ_!~XYZ@78-80-96-74.customers.tmcz.cz, Read error: Connection reset by peer)
21:17	XYZ_ has joined IRC (XYZ_!~XYZ@78-80-96-74.customers.tmcz.cz)
23:15	ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
23:23	Vercas has left IRC (Vercas!~Vercas@gateway/tor-sasl/vercas, .net .split)