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


Channel log from 8 September 2016   (all times are UTC)

00:14mtcj has joined IRC (mtcj!~AndChat21@75-162-17-214.desm.qwest.net)
00:15
<mtcj>
Is it possible to specify a different nfsroot= option for each arch under /opt/ltsp/$arch ?
00:17
My clients don't know where the nfsroot server is unless I manually modify pxelinux.cfg/default and I'd like to adhere to the advice to modify the client's /etc/ltsp/update-kernels.conf but the options in that file seem to have no effect.
00:17
<vagrantc>
probably the best way is to just use ltsp 5.5.8, which automatically figures that out without having to specify nfsroot= at all.
00:17
it derives it from the boot filename
00:17
<mtcj>
Let me see what version I'm using.
00:17
<vagrantc>
it was just released yesterday :)
00:18
<mtcj>
I see. So it's a new feature?
00:18
<vagrantc>
yes
00:18
<mtcj>
OK I'll see what I can do to get that elegantly installed on my Debian machine. Any new deps you can think of off the top of your head?
00:19
<vagrantc>
mtcj: i would recommend generating a static configuration so that ltsp-update-kernels/update-kernels doesn't clobber your editing of a non-editable file
00:19
pretty minor updates, it's in debian sid
00:20
<mtcj>
Could you expand on that static config advice?
00:20
<vagrantc>
you should be able to remove the pxelinux.cfg/default symlink(s) and replace them with whatever you want.
00:20
<mtcj>
That didn't occur to me
00:20
<vagrantc>
how many architectures are you dealing with?
00:20
<mtcj>
Honesty one.
00:21
<vagrantc>
then why do you nead this at all?
00:21
need
00:21
<mtcj>
I was having issues with different graphics drivers on different machines and was going to try different chroot for each machine
00:22
<vagrantc>
you could also configure the root-path in DHCP per-machine
00:23
lots of options
00:23
if it's a small number of machines, that might be easiest
00:24
you'd probably have to configure a different boot menu per-machine there, too
00:24
<mtcj>
Well I don't think DHCP is the beginning to the solution since currently the pxelinux config isn't pointing to a nfsroot server anyway.
00:25
That honestly was what I was originally trying to solve, when I realized I wasn't "supposed" to fix that manually in each chroot
00:25
<vagrantc>
not sure i follow you ... if the nfsroot isn't defined in DHCP and isn't defined in pxelinux, then where is it coming from?
00:25
<mtcj>
Right so I'll explain like I should've to begin with.
00:26
<vagrantc>
!ltsp-info | echo mtcj
00:26
<ltsp>
mtcj ltsp-info: please run ltsp-info and paste the output to http://ltsp.pastebin.com
00:26
<mtcj>
The boot through TFTP is fine, pxelinux gets sent to the client and vmlinuz and the initrd boot.
00:27
It has a connection refused message however when trying to begin the processes needing NFS. When I manually add the proper IP and patch for the NFS root into pxelinux.cfg/default in the client chroot I can boot using the NFS root
00:28
<vagrantc>
my guess is fixing DHCP would resolve that issue
00:28
there's also ...
00:29
!proxydhcp | echo mtcj
00:29
<ltsp>
mtcj proxydhcp: A proxy DHCP server is defined by the PXE specification as a server which sends auxiliary boot information to clients, like the boot filename, tftp server or rootpath, but leaves the task of IP leasing to the normal DHCP server. More info: https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP
00:29
<vagrantc>
i'd also recommend using ipappend 3
00:29
!ipappend
00:29
<ltsp>
ipappend: o temporarily solve DHCP problems in the initramfs, try putting IPAPPEND 3 after the APPEND line in /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default. More info: http://www.syslinux.org/wiki/index.php/SYSLINUX#IPAPPEND_flag_val_.5BPXELINUX_only.5D
00:29
<mtcj>
I'm currently using dnsmasq for proxy DHCP and I'll paste ltsp-info and dnsmasq.d/pxelinux.conf momentarily
00:30* vagrantc wanders off for a late lunch
00:32
<mtcj>
http://pastebin.com/raw/PaukGH3j and http://pastebin.com/raw/6Cgp6BZA
00:43
<vagrantc>
ok
00:43
so debian jessie?
00:44
<mtcj>
Right
00:44
<vagrantc>
mtcj: do you have IPAPPEND set in /opt/ltsp/amd/etc/ltsp/update-kernels.conf ?
00:45
<mtcj>
No it is currently commented out as value 3
00:45
<vagrantc>
you want IPAPPEND 3
00:46
<mtcj>
OK are you looking specifically at the boot-server-ip bit?
00:46
<vagrantc>
it obviates the need for setting nfsroot=
00:47
oh wait, not with that version
00:47
<mtcj>
Ah OK I'll give that a go then for sure
00:47
<vagrantc>
you still need to set nfsroot=/opt/ltsp/amd64
00:47
but you don't have to specify the ip address
00:48
and if you really need a separate chroot, you can do that with dnsmasq:
00:48
pxe-service=X86PC, "Boot Debian thin-client from network", ltsp/amd64-some-other-config/pxelinux
00:48
rather than using syslinux menu
00:48
oh, but then you can't autodetect it ... right
00:49
you could also configure pxelinux.cfg/* with the mac address and whatnot
00:49
so rather than loading pxelinux.cfg/default, it will load something else based on the machine ...
00:49
<mtcj>
I get the same behavior with IPAPPEND=3 and the addition of nfsroot=foo
00:50
<vagrantc>
did you regenerate the configuration ... e.g. ltsp-chroot /usr/share/ltsp/update-kernels && ltsp-update-kernels ?
00:52
<mtcj>
I had never done it within the chroot
00:52
Now I know
00:52
<vagrantc>
it's not exactly intuitive...
00:53
hmmm... i should probably implement ltsp-update-kernels --regenerate-config or something
00:54
didn't originally support that in order to be able to foreign chroots
00:54
<mtcj>
So it appears nfsroot= is not necessary in my case
00:54
If IPAPPEND=3 is set I have sucess
00:55
IPAPPEND is necessary though
00:55
<vagrantc>
without IPAPPEND 3 in pxelinux.cfg/* it will make another DHCP query, and possibly get the real DHCP server and attempt to mount NFS on your DHCP server.
00:56
<mtcj>
As for the reason for a PXE menu and all that, since nvidia-driver won't work on my machine that benefits from it over PXE, I figured I'd set up a PXE Debian installer to see if it works on a full system
00:57
I bet that's why it didnt work. This is the first time I've tried it with a proxy DHCP setup
00:57
So that was new to me
00:57
<vagrantc>
"ltsp-config dnsmasq" FTW
00:58
will generate a /etc/dnsmasq.d/ config with sane defaults for LTSP and some nicely documented and commented entries
00:59
should actually put a commented example like you have for debian-installer in there
00:59
i didn't figure out till last year that you could have multiple pxe-service entries
00:59
<mtcj>
I didn't know until I saw it in docs
00:59
I thought that was pretty cool
00:59
<vagrantc>
yeah, it's really cool
01:01
if you do need a separate chroot due to driver issues ... i'd recommend creating a /var/lib/tftpboot/pxelinux* in the top-level tftp dir, and then configure it to chain-load the appropriate pxe configs from the appropriate directories
01:02
<mtcj>
I could certainly do that
01:02
<vagrantc>
but hopefully you an figure out a way to avoid that :)
01:02
<mtcj>
It's my hope that after installing Debian on the machine I'll be able to get the driver working, then duplicate that success to a common amd64 chroot without cobflicts.
01:03
That seems to be the most elegant way to do it
01:03
<vagrantc>
which drivers were in conflict?
01:03
nvidia
01:04
nouveau not good enough?
01:04
<mtcj>
Well Cinnamon was in software rendering when using nouveau
01:04
<vagrantc>
that may have as much to do with being a thin client as the driver
01:04
<mtcj>
And there was a performance that maybe wasn't as good as it could have been
01:05
Yeah well see when I get real Debian on there
01:05
<vagrantc>
not using fat clients?
01:06
<mtcj>
I probably could not currently
01:06
Haven't even really got started
01:06
<vagrantc>
ideally need 2GB of ram on the clients, basically
01:07
<mtcj>
I have one powerful machine and two lower end ones that I wanted to see if I could get more out of
01:07
<vagrantc>
performance is generally a lot better, as the graphics can be rendered locally, rather than being sent over the network as X traffic
01:09
<mtcj>
I'll definitely look into it if I'm seeing any problems
01:10
<vagrantc>
if anyone expects things like youtube to work, thin clients aren't going to be a good option
02:02Freejack has left IRC (Freejack!~Freejack@unaffiliated/freejack, Ping timeout: 265 seconds)
02:07Freejack has joined IRC (Freejack!~Freejack@unaffiliated/freejack)
02:51bclub has joined IRC (bclub!484bdf86@gateway/web/freenode/ip.72.75.223.134)
03:01
<mtcj>
Sure thing
03:24
<bclub>
Hi - anyone have a guide for setting up LTSP with single NIC?
03:39
<vagrantc>
!ltsp-pnp | echo bclub
03:39
<ltsp>
bclub 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
03:40
<bclub>
thanks!
03:51* vagrantc waves
03:51vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
03:56bclub has left IRC (bclub!484bdf86@gateway/web/freenode/ip.72.75.223.134, Ping timeout: 264 seconds)
04:21gehidore is now known as ghdr
04:22ghdr is now known as gehidore
04:34mtcj has left IRC (mtcj!~AndChat21@75-162-17-214.desm.qwest.net, Quit: Bye)
05:00alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
05:04alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
06:08ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
06:15mikkel has joined IRC (mikkel!~mikkel@mail.dlvs.dk)
07:01Statler has joined IRC (Statler!~Georg@mail.lohn24.de)
07:35alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
07:38alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
09:44robb_nl has joined IRC (robb_nl!~robb_nl@ip-62-235-225-180.dsl.scarlet.be)
10:29robb_nl has left IRC (robb_nl!~robb_nl@ip-62-235-225-180.dsl.scarlet.be, Ping timeout: 240 seconds)
10:39cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
11:04robb_nl has joined IRC (robb_nl!~robb_nl@ip-62-235-225-180.dsl.scarlet.be)
11:56adrianorg has left IRC (adrianorg!~adrianorg@177.132.220.194, Ping timeout: 276 seconds)
11:58adrianorg has joined IRC (adrianorg!~adrianorg@177.156.63.247)
12:26fnurl has left IRC (fnurl!~paul@36-225-18-102.dynamic-ip.hinet.net, Read error: Connection reset by peer)
12:27fnurl has joined IRC (fnurl!~paul@36-225-18-102.dynamic-ip.hinet.net)
12:53gvy has joined IRC (gvy!~mike@altlinux/developer/mike)
13:10adrianorg has left IRC (adrianorg!~adrianorg@177.156.63.247, Ping timeout: 244 seconds)
13:12adrianorg has joined IRC (adrianorg!~adrianorg@201.22.228.164.dynamic.dialup.gvt.net.br)
13:15robb_nl has left IRC (robb_nl!~robb_nl@ip-62-235-225-180.dsl.scarlet.be, Quit: I'm gone, bye bye)
13:28sbalneav1 has left IRC (sbalneav1!~sbalneav@wnpgmb0311w-ds01-165-236.dynamic.mtsallstream.net, Ping timeout: 240 seconds)
13:29sbalneav1 has joined IRC (sbalneav1!~sbalneav@wnpgmb0311w-ds01-165-236.dynamic.mtsallstream.net)
13:44mikkel has left IRC (mikkel!~mikkel@mail.dlvs.dk, Quit: Leaving)
13:54ben_roose has joined IRC (ben_roose!~roose@roose.cs.wichita.edu)
13:58Grembler has joined IRC (Grembler!~Ben@cpc87179-aztw31-2-0-cust6.18-1.cable.virginm.net)
14:42JuJuBee has joined IRC (JuJuBee!~knichel@24.148.115.153)
14:43
<JuJuBee>
I just upgraded my ltsp server and I think I am running as thin clients instead of fat clients. How to I check? I want FAT not THIN...
14:48Grembler has left IRC (Grembler!~Ben@cpc87179-aztw31-2-0-cust6.18-1.cable.virginm.net, Quit: I Leave)
15:20vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
15:33
<JuJuBee>
I just upgraded my ltsp server and I think I am running as thin clients instead of fat clients. How to I check? I want FAT not THIN...
15:36
<vagrantc>
open a terminal, if it's a thin client, it should say USERNAME@SERVERNAME, if it's a fat-client, it should say something like USERNAME@ltsp123
15:40
<JuJuBee>
vagrantc, thanks. I looked back at my history and i did a ltsp-build-client without the fat-client parameters... I am rebuilding the image now and will check when done.
15:40
<vagrantc>
shouldn't need to rebuild
15:40
ltsp-chroot apt install X Y Z
15:41
also, you might want to explore the ltsp-pnp methods:
15:41
!ltsp-pnp | echo JuJuBee
15:41
<ltsp>
JuJuBee 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
15:41
<vagrantc>
no need to re-run ltsp-build-client, that is ...
15:42
<JuJuBee>
I thought clients had to use same arch as server with that?
15:42
I have some i386 and some amd64 clients now
15:43
<vagrantc>
so use i386 for all of them ? :)
15:43
you can typically get a PAE or amd64 kernel available on an i386 client, which gives many of the benefits of 64-bit
15:43
<JuJuBee>
most of them are i5 quad core with 8G ram
15:44
Id rather ditch the i386 clients actually
15:44
<vagrantc>
fair enough
15:45
though realistically, i doubt there would be much of a performance hit to running i386 as long as you had the appropriate kernels for the machines that support it
15:45
not much, as in, only noticeable in a controlled laboratory experiment :)
15:48
at least, for typical LTSP uses... there are always exceptions, of course :)
15:55robb_nl has joined IRC (robb_nl!~robb_nl@ip-62-235-225-180.dsl.scarlet.be)
15:56eu^17812868191ds has joined IRC (eu^17812868191ds!b28044bf@gateway/web/freenode/ip.178.128.68.191)
15:57eu^17812868191ds has left IRC (eu^17812868191ds!b28044bf@gateway/web/freenode/ip.178.128.68.191, Client Quit)
16:09
<JuJuBee>
vagrantc, I checked the current image and it is thin ... Was there an easy way to convert to fat?
16:11
<vagrantc>
JuJuBee: ltsp-chroot apt-get install PACKAGES
16:11
where PACKAGES are the desktop packages you want
16:11
might be something like FOO-desktop
16:12
<JuJuBee>
Ah, as long as there is a desktop environment then it loads as a fat client?
16:12
<vagrantc>
and has sufficient rAM
16:13
<JuJuBee>
8G should be plenty
16:13
<vagrantc>
you would think, yes. :)
16:13
i think the threshold is maybe 512MB or 800MB or something
16:13kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
16:14
<vagrantc>
whoah, it's only 300MB
16:14
<JuJuBee>
Ok, well i did a ltsp-build-client etc... again so I will just have to wait...
16:15
Is there any reason I *should* switch to pnp instead of using the traditional ltsp server?
16:15
<vagrantc>
we should probably bump that up to at least 512MB...
16:15
JuJuBee: simpler to manage
16:16
JuJuBee: if you're doing fat clients anyways
16:16
<JuJuBee>
vagrantc, I guess I should set it up in my test lab and play with it.
16:16
<vagrantc>
JuJuBee: you just install the packages you want on the clients on the server, and then build the LTSP image out of the server install
16:24
<JuJuBee>
vagrantc, thats what I did with my current install... Then I could chroot into later to install additional packages that I might NOT want on server....
16:27
<vagrantc>
yeah, if you want to make sure there are some things in the LTSP environment that aren't on the server, that method might work better
16:28
although you could add things in ltsp-update-image hooks, if i'm not mistaken
16:42
<JuJuBee>
vagrantc, in my dhcp config file, the line -- filename "/ltsp/amd64/pxelinux.0"; refers to "/var/lib/tftpboot/ltsp/amd64/pxelinux.0" ?
16:44
<vagrantc>
either that, or /srv/tftp/...
16:45
<JuJuBee>
If I created a new images and used the --chroot fat-amd64 parameter, shouldn't i see a folder /var/lib/tftpboot/ltsp/fat-amd64/ ?
16:45
<vagrantc>
does /opt/ltsp/fat-amd64 exist?
16:45
<JuJuBee>
yes
16:46
<vagrantc>
and ltsp-build-client completed without errors?
16:46
which tftp server do you have installed?
16:47
<JuJuBee>
There was an error... http://pastebin.com/K95diMU0
16:48
root 3227 0.0 0.0 15180 268 ? Ss Sep06 0:00 /usr/sbin/in.tftpd
16:48gvy has left IRC (gvy!~mike@altlinux/developer/mike, Quit: Leaving)
16:48
<vagrantc>
dpkg -l '*tftp*'
16:48
JuJuBee: looks like it didn't successfully install
16:49
<JuJuBee>
dpkg-query: no packages found matching tftp
16:49
vagrantc, very strange...
16:49
<vagrantc>
JuJuBee: so, packages can fail for all sorts of reasons... the error you pasted isn't very helpful, likely the real error is higher up
16:49
JuJuBee: or obfuscated entirely
16:50
<JuJuBee>
vagrantc, suggestions?
16:50
<vagrantc>
JuJuBee: honestly, i'd recommend installing a "thin client" first, and then using "ltsp-chroot apt ..." to install the rest of the packages
16:51
then you can isolate the error from the process of building the LTSP environment
16:51
<JuJuBee>
vagrantc, like kubuntu-desktop etc... and other desktop managers
16:52
<vagrantc>
yes
16:52
<JuJuBee>
so if there are multiple folders in /opt/ltsp like amd64 and i386, how should I chroot into just the amd64
16:54
<vagrantc>
ltsp-chroot --chroot or --arch ... i forget off the top of my head
16:55
<JuJuBee>
ltsp-chroot --arch amd64 for example
16:56
<vagrantc>
ltsp-chroot --arch amd64 -- apt install foo bar
16:57
<JuJuBee>
seems to be the ticket... I'll let you know if it worked
17:09
vagrantc, when the install finishes, I still need to ltsp-update-image ? and sshkeys and kernels if applicable.
17:14
<vagrantc>
i *think* ltsp-update-image is all you should need to do ... but you can also manually run ltsp-update-kernels to be sure
17:14
sshkeys won't have changed
17:21Statler has left IRC (Statler!~Georg@mail.lohn24.de, Quit: Leaving)
18:00
<JuJuBee>
vagrantc, looks like I am back to fat clients. Thank you for your help.
18:02
<alkisg>
ltsp-chroot -ma amd64
18:02
-m ==> mount all, package cache for speed and other special dirs for the package installation not to fail
18:19kjackal has left IRC (kjackal!~quassel@97.93.161.2, Ping timeout: 265 seconds)
18:19kjackal_ has joined IRC (kjackal_!~quassel@97.93.161.2)
18:28kjackal_ has left IRC (kjackal_!~quassel@97.93.161.2, Quit: No Ping reply in 180 seconds.)
18:28telex has joined IRC (telex!~telex@freeshell.de)
18:31kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
18:44kjackal has left IRC (kjackal!~quassel@97.93.161.2, Quit: No Ping reply in 180 seconds.)
18:54kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
18:55robb_nl has left IRC (robb_nl!~robb_nl@ip-62-235-225-180.dsl.scarlet.be, Read error: Connection reset by peer)
19:02mgariepy has left IRC (mgariepy!~mgariepy@ubuntu/member/mgariepy, Quit: Leaving)
19:07kjackal has left IRC (kjackal!~quassel@97.93.161.2, Quit: No Ping reply in 180 seconds.)
19:08kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
19:24gehidore is now known as man
19:27man is now known as gehidore
19:43
<JuJuBee>
alkisg, thanks for that.
19:47kjackal_ has joined IRC (kjackal_!~quassel@97.93.161.2)
19:47kjackal has left IRC (kjackal!~quassel@97.93.161.2, Ping timeout: 260 seconds)
19:50vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
20:39kjackal_ has left IRC (kjackal_!~quassel@97.93.161.2, Ping timeout: 244 seconds)
20:40kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
20:51kjackal has left IRC (kjackal!~quassel@97.93.161.2, Quit: No Ping reply in 180 seconds.)
20:57kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
21:02kjackal has left IRC (kjackal!~quassel@97.93.161.2, Ping timeout: 265 seconds)
21:02kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
21:19ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
21:44adrianorg has left IRC (adrianorg!~adrianorg@201.22.228.164.dynamic.dialup.gvt.net.br, Ping timeout: 250 seconds)
21:46adrianorg has joined IRC (adrianorg!~adrianorg@177.204.159.236.dynamic.adsl.gvt.net.br)
22:09kjackal has left IRC (kjackal!~quassel@97.93.161.2, Quit: No Ping reply in 180 seconds.)
22:12kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
23:18kjackal has left IRC (kjackal!~quassel@97.93.161.2, Quit: No Ping reply in 180 seconds.)
23:21kjackal has joined IRC (kjackal!~quassel@97.93.161.2)
23:25ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)
23:38kjackal has left IRC (kjackal!~quassel@97.93.161.2, Quit: No Ping reply in 180 seconds.)
23:39adrianorg has left IRC (adrianorg!~adrianorg@177.204.159.236.dynamic.adsl.gvt.net.br, Ping timeout: 265 seconds)
23:41adrianorg has joined IRC (adrianorg!~adrianorg@177.204.159.246.dynamic.adsl.gvt.net.br)
23:43kjackal has joined IRC (kjackal!~quassel@97.93.161.2)