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


Channel log from 9 February 2017   (all times are UTC)

00:05ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)
00:25vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
01:38gdi2k has left IRC (gdi2k!~gdi2k@49.151.85.159, Ping timeout: 252 seconds)
02:52gdi2k has joined IRC (gdi2k!~gdi2k@49.151.85.159)
02:54pppingme is now known as pppingme2
02:54pppingme2 is now known as pppingme
04:44zerkalo has left IRC (zerkalo!myricae@ny1.hashbang.sh, Ping timeout: 245 seconds)
04:45zerkalo has joined IRC (zerkalo!myricae@ny1.hashbang.sh)
05:18Freejack has left IRC (Freejack!~Freejack@unaffiliated/freejack, Ping timeout: 260 seconds)
05:23Freejack has joined IRC (Freejack!~Freejack@unaffiliated/freejack)
05:51zama has left IRC (zama!~zama@unaffiliated/stryx/x-3871776, Ping timeout: 240 seconds)
06:00zama has joined IRC (zama!~zama@unaffiliated/stryx/x-3871776)
06:14Statler has joined IRC (Statler!~Georg@p579FE22F.dip0.t-ipconnect.de)
06:32zerkalo_ has joined IRC (zerkalo_!myricae@ny1.hashbang.sh)
06:33zerkalo has left IRC (zerkalo!myricae@ny1.hashbang.sh, Ping timeout: 240 seconds)
06:33zerkalo_ is now known as zerkalo
06:36ricotz has joined IRC (ricotz!~ricotz@p5B2A93D7.dip0.t-ipconnect.de)
06:36ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
06:45
<alkisg>
!cheap-client
06:45
<ltsp>
cheap-client: (#1) http://www.gearbest.com/tv-box-mini-pc/pp_343636.html, or (#2) https://www.aliexpress.com/store/product/New-arrival-Beelink-Pocket-Z83-Windows-10-Mini-PC-Z8300-64bit-1-84GHz-2GB-RAM-32GB/1871240_32640039781.html
07:12
<vlt>
alkisg: Good morning!
07:12
<alkisg>
Good morning vlt :)
07:12
<vlt>
What does that mean “sisfb isn't loaded by default in ubuntu 12.04”?
07:12
<alkisg>
Did you decide to throw away those lame clients yet? :D
07:12
<vlt>
:-D
07:12
Soooooon
07:13
<alkisg>
vlt, it means that if you manage to install xserver-xorg-video-sis to 14.04 or even to 16.04, you will see the available resolutions again
07:13
*all the ...
07:14
vlt, if you plan to throw them away before 2019, why not install ubuntu 14.04, which is supported till then
07:14
And completely ignore 16.04...
07:14
<vlt>
Yes. The -sis pkg seems to be available until 14.04, so that would be an option (for the clients) to get a slightly longer supported system.
07:14
<alkisg>
But take care to use 14.04.1, not e.g. 14.04.3 or anything else, that won't have the driver
07:15
<vlt>
What?!
07:15
They dropped support for -sis within the 14.04 release?
07:16
<alkisg>
That's how lts releases work
07:16
<vlt>
I still don't understand what exactly your sisfb sentence means because yesterday it looked like the -sis driver wasb't used at all.
07:16
<alkisg>
There's lts support for kernels and xorg
07:17
That means that when 14.10 is out, they backport their kernels and xorg to 14.04
07:17
And when 15.04 is out too, and that becomes 14.04.3
07:17
But that's a different line of upgrades
07:17
Those that had the initial kernel and xorg, don't automatically update to the newer stack
07:17
So you'll need to have xserver-xorg installed, not e.g. xserver-xorg-lts-vivid or something
07:18
<vlt>
Ok. In my case it's not this important because for now I'll stick with 12.04 and a SCREEN_07 line for all the clientsm
07:18
-m
07:19
<alkisg>
You have thin clients; you can keep a thin chroot for ages
07:19
And only update the server...
07:19
<vlt>
When everyone is happy with performance again I'll switch to 14.04 clients
07:19
<alkisg>
Just remember, when you'll buy new clients, to select fat ones, not thins :)
07:20
<vlt>
I'll do.
07:21gdi2k has left IRC (gdi2k!~gdi2k@49.151.85.159, Ping timeout: 240 seconds)
07:21
<vlt>
Seems strange now after more than 10 years letting the server do all the work.
07:22
Something in my head tells me that a cheap fat client will never be able to compete with the power of our 12 core 48 GB RAM machine :D
07:22
<alkisg>
Check cpubenchmark and say that again :)
07:22
A monster server scores something like 10.000 passmark
07:22
A modest atom scores 2.000
07:22
So, 5 atoms = 1 monster server
07:22
And if you have 70 clients... do the math
07:23
One i3 scores something like 5.000-6.000
07:23
!cheap-client
07:23
<ltsp>
cheap-client: (#1) http://www.gearbest.com/tv-box-mini-pc/pp_343636.html, or (#2) https://www.aliexpress.com/store/product/New-arrival-Beelink-Pocket-Z83-Windows-10-Mini-PC-Z8300-64bit-1-84GHz-2GB-RAM-32GB/1871240_32640039781.html
07:24
<alkisg>
Even the lamest fat client scores more than 1000: http://www.cpubenchmark.net/cpu.php?cpu=Intel+Atom+x5-Z8350+%40+1.44GHz
07:24
And I've seen expensive Xeon processors scoring less than cheap i3/i5...
07:25
<vlt>
You did the math.
07:26gdi2k has joined IRC (gdi2k!~gdi2k@49.151.85.159)
07:40mikkel has joined IRC (mikkel!~mikkel@mail.dlvs.dk)
07:49
<alkisg>
!server
07:49
<ltsp>
sinisalo.freenode.net
07:50
<alkisg>
!server-ram
07:50
<ltsp>
server-ram: LTSP server RAM *really* depends on the usage. But anyway here's an approximation: Server RAM in MB = 1500 + 30*number-of-fat-clients + 300*number-of-thin-clients
07:50
<alkisg>
!server-cpu
07:50
<ltsp>
I do not know about 'server-cpu', but I do know about these similar topics: 'server-ram'
07:50
<alkisg>
!learn server-cpu as LTSP server CPU *really* depends on the usage. But anyway here's an approximation: Server CPU in http://cpubenchmark.net score: 1500 + 30*number-of-fat-clients + 300*number-of-thin-clients
07:50
<ltsp>
The operation succeeded.
07:52
<alkisg>
vlt: so, according to that ^ formula, for 70 thin clients, you would need a passmark score of 22500, which doesn't exist, and for 70 fat clients, a score of 3600, which is a plain celeron.
08:38markit has joined IRC (markit!~marco@88-149-177-66.v4.ngi.it)
08:39
<alkisg>
markit: try `sudo service network-manager restart` to see if nm comes back online
08:40
It's a known bug with nbd-client
08:41
<markit>
alkisg: hi, today I've re-started the Vm and connectivity works! So strange
08:41
<alkisg>
Nah, it's called "race condition"
08:41
<markit>
alkisg: but the VM that acts as client does not boot, it stops at
08:42
<alkisg>
It works only half of the times
08:42
<markit>
tftp://192.168.0.10//ltsp/amd64/pxelinuux.0..... connection timed out
08:42
alkisg: so better use static network assignment?
08:43
(I'm a fan of /etc/network/interfaces :P)
08:43
<alkisg>
Haha
08:43
Run this: sudo dpkg-divert --divert /etc/nbd-client.init --rename /etc/init.d/nbd-client
08:43
About the connection timed out, what are the contents of /etc/dnsmasq.d/network-manager?
08:45
<markit>
let me check
08:46
only "bind-interfaces"
08:47
<alkisg>
Comment it out and restart dnsmasq
08:48
<markit>
wow, works!
08:48
alkisg for the rescue, as usual
08:48
<alkisg>
Sure, why the "wow"? :P
08:48
:D :D :D
08:49
<markit>
alkisg: ltsp will die if you abandon it
08:49
<alkisg>
Nah, others will pick it up
08:49
<markit>
would love to belive it (just to relieve the pressure on you) but don't think so
08:50
<alkisg>
The problem is with ubuntu there though, not with ltsp
08:50
They decided to abuse dnsmasq to make it a local resolver, without caring that they break a whole lot of other use cases
08:50
<markit>
(at the phone)
08:51
<alkisg>
And the nbd-client issue, is a problem with debian, that it "releases when it's ready"; while they could just set a specific freeze date that aligns with ubuntu, e.g. every 2 years, and then release when they're ready after the freeze date
08:52
(the problem with nbd-client was fixed in time for debian stretch, but not for ubuntu 16.04, and now it's difficult to sync it with 16.04...)
08:52
...bureaucracy...
08:57
<markit>
sorry, back, reading
08:58
I'm with 16.04, classic ltsp worked fine with kde (I've still usual big I/O issues with their damn way of cache icons or other stuff that generates tons of I/O on hte server)
08:58
<alkisg>
Classic you mean with a chroot?
08:59
<markit>
yes sorry :)
08:59
<alkisg>
The network-manager and dnsmasq and nbd-client issues are not related to chroots or pnp
08:59
And neither to desktop environment
08:59
They are there in 16.04 no matter which de you have
08:59
As long as you have those packages installed...
08:59
<markit>
yep, just to notify you that kde works fine with 16.04 (I'm a kde fan, you know)
08:59
<alkisg>
Cool
09:00
<markit>
and since kde people don't care about ltsp, breakage that I could not solve is always a risk for me (I'm not as good as you are, if something very wrong arises I'm lost)
09:01
<alkisg>
Ideally you should be able to reproduce all the issues without ltsp too
09:02
So that you can report to them "it doesn't work in this use case" without even mentioning the word ltsp
09:02
<markit>
I tell you my feeling about GNU desktop development, at least kde
09:02kjackal_ has joined IRC (kjackal_!~quassel@onopfy.static.otenet.gr)
09:02
<markit>
they use ssd, they are "desktop single users" centric
09:03
there are problems with homes shared in nfs or something like that that they never considered
09:04
in my desktop the high I/O is a matter of some seconds, if you have ssd is not a problem, if you have 10 clients booting at the same time is a huge difference instead
09:04
<alkisg>
I don't think kde requires an ssd
09:05
<markit>
I did not say so, just if you are a developer and use ssd, you think that bottleneck is not important
09:05
<alkisg>
Most DEs also care a bit about mobile environments too nowadays
09:05
So they don't want to get the specifications too high
09:05
<markit>
well, probably they find more funny add visual effects than tryint to trubleshoot problems only few people raise
09:06
is a matter of facts that booting kde is slower than other desktops I tested time ago
09:06
and the I/O I measured was 2-3x of gnome at those time
09:07
<alkisg>
Also, if I remember correctly you said that this only happened on ltsp clients, not on standalone workstations
09:07
I.e. there's something there that forces kde to do all those reads
09:07
<markit>
but I'm not skilled enough to pinpoint the problem, and no developer was interested at that time (and they are not very friendly to be sincere)
09:07
<alkisg>
Maybe it's possible to force it to NOT do all those reads, so that's it's fast again
09:07
Then you could switch to another DE
09:08
<markit>
no, it does always all those reads, only is a pain if done through network
09:08
<alkisg>
It always reads e.g. 600 mb for a single login? that sounds lame...
09:08
Someone should pinpoint why that is the case
09:10
<markit>
first time you login does a huge read then puts in /var/tmp, like /var/tmp/kdecache-marco
09:10
the subsequent time it does a better job, but still heavy AFAIR
09:11
<alkisg>
That would mean that "guest login" is very slow, right?
09:11
Because it happens from scratch each time
09:11
That is one way to report it without mentioning "ltsp" anywhere...
09:11
<markit>
I've tried to move that cache to their home dir, so they can find it again, but nfs is not that bright in performances
09:11
<alkisg>
Just tell them that you have a normal, 100 GB rotational hard disk, and it takes a long time to login as guest
09:12
<markit>
alkisg: yep, but they probably tell you that is just a 15 seconds wait, that second time is better, and "who does not have ssd these days?"
09:13
<alkisg>
There's no "second time" for guest
09:13
No sdd => you can point them to 80% of the users that don't have one
09:13
So I can't believe they would say such lame excuses
09:13
<markit>
who cares about "guest" account? :)
09:14
I've heard incredible things in my life as excuses for obvious bugs, and I'm sure you too
09:14
in any case the best answer is "fix it yourself, there is the source code"
09:14
<alkisg>
Where is your official bug report for that?
09:15
Let me read their answers there, in the bug tracker
09:15
Not in irc or forums
09:16
<markit>
urgh, is a long time ago, don't remember, I will have a look and send you an email
09:17
btw, I've created an user, and now I try to login but without success
09:17
Received disconnect from 192.168.0.144 port 42426:11: disconnected by user
09:18
I enter username, the password, then a brief text screen with some messages, then login again
09:18
in the /var/log/auth.log I've traces that it logs, then disconnects
09:19
and btw, how are homes mounted? it depends upon lts.conf as usual?
09:20
<alkisg>
The default is sshfs
09:20
Try with the xterm session
09:20
From the ldm preferences menu
09:25
<markit>
btw, my discussion about kde were almost in chat, and were very frustrating and unfriendly. Finally I contacted (I don't remember how and why) the chief developer that invited me to write to the ml, I did, he replied and I did not advanced my requests since I found how to "make it work decently" and was not able to give more tech details to him
09:25
https://mail.kde.org/pipermail/plasma-devel/2012-September/021607.html
09:26
<alkisg>
Yup, that's not a bug tracker :)
09:26
Bug trackers force developers to reply more formally
09:26
"won't fix" means that they actually don't care
09:26
It's hard to mark a bug "won't fix" if it affects a lot of users
09:26
So, next time use a bug tracker instead
09:28
They also can't reply "fix it yourself" there. They can say "we accept patches", but they cannot close the bug just because you are unable to provide them
09:28
<markit>
xsession: unable to launch "/usr/bin/xterm" X session --- "/usr/bin/xterm" not found; failling back to defaul tsession
09:28
<alkisg>
Install xterm :)
09:28
Or symlink another xterm to that
09:29
<markit>
alkisg: I'm not able to provide them better info, is like "clueless newbie here with a bigh tech problems that points out your desigh is buggy, care to help?" :)
09:29
<alkisg>
I've seen bug reports that start with a whole lot less than what you provided there
09:30
If they want something, they can ask you, and even mention specific steps to get it
09:30
<markit>
mmm dns does not work, I've to try the wiki solution I guess
09:31
<alkisg>
So, a bug report can be: "first login is very slow", and mention that it takes 30 secs in your hardware, while e.g. gnome takes 10 secs
09:31
No mention of ltsp is needed there
09:32
And of course noone will tell you "buy an ssd" there in the bug report
09:33
<markit>
damn, I've applied the fist solution of the wiki and now dnsmasq does not restart
09:33
Job for dnsmasq.service failed because the control process exited with error code. See "systemctl status dnsmasq.service" and "journalctl -xe" for details.
09:34
nss-lookup.target: Job nss-lookup.target/start failed with result 'dependency'.
09:34
what about a reboot? :)
09:34
<alkisg>
Do you have epoptes installed?
09:34
<markit>
no
09:34
<alkisg>
Yup, a reboot would help too
09:35
Or at least restarting network-manager
09:35
<markit>
it restarted fine AFAIK
09:36
ok, now let's try again with xterm
09:38
ehm, probably I've to rebuild the image, right? :)
09:38
<alkisg>
Yup, unless you switch to thin client mode
09:38
Install epoptes too
09:38
You'll probably need it :D
09:39Sameer has joined IRC (Sameer!b649b236@gateway/web/freenode/ip.182.73.178.54)
09:39
<Sameer>
Hi
09:39
<alkisg>
Hello
09:40
<Sameer>
Need to know about LTSP licence
09:40
<alkisg>
It's gpl 2, mostly
09:40
Why?
09:42
<Sameer>
I am asking because we are doing audit for software that we are using in our company. So need to confirm about LTSP Licence.
09:42
Is there any refernce link for it
09:42
<alkisg>
!ltsp-source
09:42
<ltsp>
ltsp-source: at http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/files
09:42
<alkisg>
The files are there,
09:43
and each file states the license in the top of its contents
09:43
<markit>
alkisg: btw, reading my old notes probably sshfs will fail with kde, better try nfs. I will read what you told me time ago about how to pass config files that I usually put in chroot and try
09:43
<alkisg>
Sameer: Most of them say "gpl v2+", so for a general audit, that line should be enough
09:44
<Sameer>
thanks for info. alkisg
09:46
<alkisg>
np
09:46
markit: you enable nfs from lts.conf, not from the chroot
09:47
!nfs
09:47
<ltsp>
nfs: to enable NFS home directories for localapps and fat clients, install nfs-kernel-server etc on your server, and put this inlts.conf: FSTAB_1="server:/home /home nfs defaults,nolock 0 0"
09:47
<markit>
alkisg: I mean I have to set an environment variable in chroot, the client must run it, it was in /etc/X11/Xsession.d/10ltspMM_rm_phonondevicesrc
09:47
(my script) that sets KDEVARTMP=$HOME/.cache
09:48
<alkisg>
I have no idea about phonon
09:48
<markit>
is just the name of the script, there I put everything I need ;P
09:48
<alkisg>
That's to change KDEVARTMP, it's not related to nfs
09:48
<markit>
yep, but without that change you can't login
09:48
<alkisg>
OK, but still not related to nfs that we were talking about
09:50
<markit>
sure, nfs is not a problem, why are we talking about that? My problem is have my script run at client startup
09:50
<alkisg>
No, you wanted it to run before xsession initialization, not at client startup
09:50
<markit>
btw, with xterm sesssion I see the prompt but can't act in it
09:50
<alkisg>
RCFILE is for client startup
09:51
Move the mouse above the prompt
09:51
Then type
09:51
<markit>
alkisg: forgive my ignorance, I just want to replicate the same functionality I had with "normal ltsp" when in "ltsp-pnp", since I've no chroot I don't know how to make my script act
09:52
<vlt>
alkisg: Thank you.
09:52
<alkisg>
No worries; I'm just explaining stuff when you say something that isn't accurate
09:52
(markit)
09:52
<markit>
funny, the mouse is an hourglass in most of the screen but not around the prompt, so I though was frozen, instead works
09:52
alkisg: I know you love me ;P
09:52
<vlt>
The 70 clients are spread among 4 LTSP servers.
09:52* markit takes note and tryes to learn, thanks
09:53
<vlt>
But I will switch to fat clients within the following 4 years.
09:53
<alkisg>
markit: now, run this: grep ^Exec /usr/share/xsessions/*, on your server, and tell us the output
09:54
<markit>
server = client I guess
09:55
<alkisg>
It's the same; I just wanted you to be able to copy/paste more easily...
09:55
Because copy/pasting to web from xterm will be a challenge
09:55
<markit>
:)
09:55
<alkisg>
grep ^Exec /usr/share/xsessions/* | nc termbin.com 9999
09:55
<markit>
2 lines
09:55
[10:53] <alkisg> markit: now, run this: grep ^Exec /usr/share/xsessions/*, on your server, and tell us the output
09:55
ops
09:55
<alkisg>
That is smaller
09:55
<markit>
'/usr/share/xsessions/gnome.desktop:Exec=gnome-session --session=gnome
09:55
'/usr/share/xsessions/plasma.desktop:Exec=/usr/bin/startkde
09:55
<alkisg>
OK I guess you don't want to test gnome,
09:55
so, from that xterm...
09:56
<markit>
really!
09:56
<alkisg>
you do see marco@ltsp123 as the prompt in that xterm, right?
09:56
<markit>
is it a ltsp dependency? I had some problem also with regular ltsp afair
09:56
<alkisg>
...not root or server or something weird, correct?
09:56
<markit>
correct
09:56
<alkisg>
So, from the marco@ltsp123 prompt, run: startkde
09:56
And see if it works, or what error message it produces
09:57
<markit>
wow, works, kde is soooo beautiful ;P
09:57
<alkisg>
If it works, then try to login, and select kde from the ldm preferences menu
09:58
*try to logout, sorry
09:59markus_e92 has left IRC (markus_e92!~markus_e9@91-115-155-174.adsl.highway.telekom.at, Ping timeout: 245 seconds)
09:59
<markit>
works too, is just the "preferred" that is not right
10:00
<alkisg>
Was that an existing user?
10:00
Maybe his .dmrc was wrong
10:00
I.e. a user issue, not an ltsp issue
10:00
<markit>
no, created from scratch (I'm using vms in proxmox)
10:00
alkisg: installing ltsp packages I've seen some "gnome related" packages installed too... could it be an issue?
10:01
<alkisg>
Maybe it tried gnome then
10:01
Could be
10:01
bb later... have fun!
10:01
<markit>
alkisg: thanks so much, now I've to discover how to have kde be the preferred session and I'm almost done, great!
10:01
<alkisg>
LDM_SESSION=plasma
10:09
<markit>
apt purge gnome-session wants to install tons of packages mate related
10:10Statler has left IRC (Statler!~Georg@p579FE22F.dip0.t-ipconnect.de, Remote host closed the connection)
10:23gvy has joined IRC (gvy!~mike@altlinux/developer/mike)
10:23markus_e92 has joined IRC (markus_e92!~markus_e9@62-46-96-227.adsl.highway.telekom.at)
10:32
<alkisg>
markit, put your /var/log/apt/history.log to pastebin
10:33
ldm-server Recommends: gnome-session | mate-desktop-environment | x-session-manager | x-window-manager, ltspfs, libasound2-plugins, xauth
10:33
Maybe that's the issue there
10:34
<markit>
alkisg: btw, if the lock screen starts, I can't login again! in .xsession-errors says "Authentication failure" (don't know if is a kde bug, I remember happend time ago also in my desktop after a day of not usage, but at the first lock, just after a lot of time)
10:34
<alkisg>
Maybe kde doesn't provide x-session-manager, and it's a bug in the kde packaigng
10:34
!hash
10:34
<ltsp>
I do not know about 'hash', but I do know about these similar topics: 'ROOT_PASSWORD_HASH', 'LDM_PASSWORD_HASH'
10:34
<alkisg>
!LDM_PASSWORD_HASH
10:34
<ltsp>
LDM_PASSWORD_HASH: LDM_PASSWORD_HASH=True in lts.conf saves the password hash to /etc/shadow on login, so that the users can unlock the screensaver etc. If they happen to change their password though, that only takes effect until logout.
10:34
<alkisg>
Use that for lock
10:35Statler has joined IRC (Statler!~Georg@mail.lohn24.de)
10:37
<markit>
alkisg: http://paste.ubuntu.com/23959940/
10:37
<alkisg>
markit: yup, sounds like it
10:38
markit, try this: sudo apt install --no-install-recommends choosewm
10:38
Then, remove the gnome packages
10:38Statler has left IRC (Statler!~Georg@mail.lohn24.de, Remote host closed the connection)
10:38
<alkisg>
After that, you should be able to remove them without having to install mate :D
10:39
If things go as I said, then file a bug report for ldm-server to include the plasma de too
10:39
Or, for plasma de to provide x-session-manager
10:39
$ apt-cache show lxde-common | grep Provides
10:39
Provides: x-session-manager
10:40
E.g. that's what lxde does
10:40Statler has joined IRC (Statler!~Georg@mail.lohn24.de)
10:41Sameer has left IRC (Sameer!b649b236@gateway/web/freenode/ip.182.73.178.54, Ping timeout: 260 seconds)
10:43
<markit>
a bug to ltsp bug tracker you mean?
10:43
<alkisg>
!ltsp-bug
10:43
<ltsp>
ltsp-bug: To file a bug report for upstream LTSP, go to https://bugs.launchpad.net/ltsp
10:43
<alkisg>
yup, either the ltsp or the kde tracker
10:43
<markit>
great
10:44
<alkisg>
(the debian kde tracker)
10:44
<markit>
mmm insistes in installing mate stuff
10:46
http://paste.ubuntu.com/23959960/
10:46
<alkisg>
I don't see the command you gave there
10:46
Also, did you install choosewm?
10:47
<markit>
ehm, hell not!
10:47
<alkisg>
:)
10:47
(12:38:39 μμ) alkisg: markit, try this: sudo apt install --no-install-recommends choosewm
10:47
<markit>
choosewm is already the newest version (0.1.6-3)
10:47
ah, yes I did, is the only thing I did
10:48
let me re-read your instructions
10:48
[11:38] <alkisg> markit, try this: sudo apt install --no-install-recommends choosewm
10:48
done
10:48
[11:38] <alkisg> Then, remove the gnome packages
10:49
apt purge gnome-session
10:49
then I stopped before it installed 600MB of poisoning stuff
10:49
<alkisg>
Package: ltsp-server-standalone, Depends: gnome-session | mate-desktop-environment | x-session-manager | x-window-manager, isc-dhcp-server | dhcp3-server | dnsmasq, ldm-server, libasound2-plugins, ltspfs, ltsp-server, nbd-server (>= 1:2.9.25-2), openbsd-inetd | inet-superserver, openssh-server, squashfs-tools, tftpd-hpa | atftpd | dnsmasq (>= 2.41), x11-utils, xauth
10:50
<markit>
I just wondering why a PURGE makes apt install NEW STUFF!
10:50
<alkisg>
To satisfy the ltsp-server-standalone dependencies
10:50
<markit>
(btw, screen lock now works, thanks to LDM_PASSWORD_HASH )
10:51
ah, it installs mate-desktop-environment stuff, now I get
10:51
<alkisg>
Sorry I was wrong choosewm doens't provide x-window-manager
10:51
markit: sudo apt install --no-install-recommends matchbox-window-manager
10:51
Try this instead of choosewm
10:51
Then try to remove gnome
10:51
Cheers!
10:51
bbl
10:52
<markit>
great, works
11:28larryni has left IRC (larryni!~larryni@host31-48-109-69.range31-48.btcentralplus.com, Ping timeout: 258 seconds)
11:41larryni has joined IRC (larryni!~larryni@host31-48-109-69.range31-48.btcentralplus.com)
11:43Faith has joined IRC (Faith!~paty_@unaffiliated/faith)
11:55cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 276 seconds)
12:04kjackal_ has left IRC (kjackal_!~quassel@onopfy.static.otenet.gr, Ping timeout: 240 seconds)
12:13
<vlt>
alkisg: Update: I was curious and booted a freshly installed standard 14.04 image on the client. The sis chip "works" out of the box (as in it runs on 1920x1080 resolution) but image rendering is noticably slower.
12:14
(I tested only in xfreerdp client mdoe)
12:15
*mode
12:17markit has left IRC (markit!~marco@88-149-177-66.v4.ngi.it, )
12:17
<vlt>
Strangely the ltsp image that was build by ltsp-build-client is 14.04.5 while the host is (a debootstrapped 14.04).
12:19
If you’re right, alkisg, then down-grading the client image to 14.04 could help. But w/o the updates I could as well stay on 12.04.
12:20
But still – there must be something present in the 12.04 image that lets the chip work quite ok and I have no idea what o_O
12:33larryni has left IRC (larryni!~larryni@host31-48-109-69.range31-48.btcentralplus.com, Quit: Leaving)
12:35cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
12:35alkisg_web has joined IRC (alkisg_web!4d316cbe@gateway/web/freenode/ip.77.49.108.190)
12:35
<alkisg_web>
vlt, of course 14.04.1 gets all the updates
12:36
You haven't understood the concepts of kernel upgrades
12:36
Read this: https://wiki.ubuntu.com/Kernel/LTSEnablementStack
12:36
An updated 14.04.1 release will be 14.04.5 for everything except kernel and xorg
12:36
Which is exactly what you need
12:37
<vlt>
alkisg_web: You’re right, I absolutely haven’t :D
12:45kjackal_ has joined IRC (kjackal_!~quassel@2a02:587:310e:e800:54ed:4fbb:b4a:ed2e)
12:46
<vlt>
alkisg_web: In this diagram https://wiki.ubuntu.com/Kernel/LTSEnablementStack#Kernel.2FSupport.A14.04.x_Ubuntu_Kernel_Support it looks like the chroot env that says "14.04.5" should have a 4.4 kernel.
12:47
But it’s this instead: initrd.img-3.13.0-108-generic
12:48
<ogra_>
"<alkisg_web> An updated 14.04.1 release will be 14.04.5 for everything except kernel and xorg"
12:48
;)
12:48* ogra_ points to the "except"
12:50* vlt just doesn’t get it :D
12:51
<vlt>
A 14.04.5 should have a 4.4 kernel. Is that correct?
12:56
<ogra_>
if you installed from a 14.04.5 iso image initially, yes
12:57
an upgraded 14.04.1 install will keep the kernel and xorg that was used to install
12:57
https://wiki.ubuntu.com/Kernel/LTSEnablementStack#Ubuntu_14.04_LTS_-_Trusty_Tahr
12:57
you need to perform these steps if you want that upgraded
13:09
<vlt>
Ah ok
13:09
I always use debootstrap.
13:09
<ogra_>
you shouldnt ... the installer seeds a lot of stuff in the debconf db
13:09
stuff that you potentially want
13:10
(for a chroot or such deboostrap is fine, but not really for a real install)
13:10
<alkisg_web>
vlt, the end result is that your 14.04 chroot is exactly what you need, and you can upgrade it without fearing it will lose the sis support
13:11
<vlt>
ogra_: Can you give an example for stuff that’s missing in my debootstrapped systems?
13:12
<alkisg_web>
ltsp-build-client uses debootstrap which installs the 14.04.1 kernel and xorg, along with all the upgrades, which is what you want
13:12
You *don't* want the newer kernel and xorg
13:12
<ogra_>
vlt, you would have to take a look at debian-installer code really
13:12
<vlt>
alkisg_web: Perfect. Then we only have to find out why it behaves noticably slower on 14.04 -sis.
13:13
<alkisg_web>
vlt: is that with the same version of firefox on the server?
13:13
<vlt>
alkisg_web: It’s the very same 16.04 server I connect to via xfreerdp.
13:13
alkisg_web: And it’ſ not specifically firefox.
13:14
<alkisg_web>
Ah, if you're using xfreerdp, then that might be it, the newer version might be using something different
13:14
Try x11perf or gtkperf *locally* to measure the difference between xorg performance
13:14
<vlt>
Every /me wonders why newer means always slower lately o_O
13:14
-Every
13:15
<ogra_>
code grows while it improves :)
13:15
<vlt>
alkisg_web: Ok, I’ll do that. Later. Now I have to configure the 16.04 machine.
13:15* alkisg_web waves :)
13:15alkisg_web has left IRC (alkisg_web!4d316cbe@gateway/web/freenode/ip.77.49.108.190, Quit: Page closed)
13:53lucascastro has joined IRC (lucascastro!~lucas@138.185.108.51)
14:00mikkel has left IRC (mikkel!~mikkel@mail.dlvs.dk, Quit: Leaving)
14:20TatankaT has left IRC (TatankaT!~tim@193.190.253.114, Ping timeout: 252 seconds)
14:20lucascastro has left IRC (lucascastro!~lucas@138.185.108.51, Ping timeout: 255 seconds)
14:27TatankaT has joined IRC (TatankaT!~tim@193.190.253.114)
14:37ben_roose has joined IRC (ben_roose!~roose@roose.cs.wichita.edu)
14:50lucascastro has joined IRC (lucascastro!~lucas@138.185.108.51)
15:03gvy has left IRC (gvy!~mike@altlinux/developer/mike, Quit: Leaving)
15:58markit has joined IRC (markit!~marco@88-149-177-66.v4.ngi.it)
16:11Grembler has joined IRC (Grembler!~Ben@cpc87179-aztw31-2-0-cust6.18-1.cable.virginm.net)
16:13markus_e92 has left IRC (markus_e92!~markus_e9@62-46-96-227.adsl.highway.telekom.at, Ping timeout: 240 seconds)
16:15markus_e92 has joined IRC (markus_e92!~markus_e9@62-46-96-144.adsl.highway.telekom.at)
16:20Grembler has left IRC (Grembler!~Ben@cpc87179-aztw31-2-0-cust6.18-1.cable.virginm.net, Quit: I Leave)
16:27Da-Geek has joined IRC (Da-Geek!Da-Geek@nat/redhat/x-dkyzdolsstmwwkvf)
16:45alkisg_web has joined IRC (alkisg_web!4d316cbe@gateway/web/freenode/ip.77.49.108.190)
17:34markus_e92 has left IRC (markus_e92!~markus_e9@62-46-96-144.adsl.highway.telekom.at, Ping timeout: 240 seconds)
17:36markus_e92 has joined IRC (markus_e92!~markus_e9@91-115-18-4.adsl.highway.telekom.at)
18:01TatankaT has left IRC (TatankaT!~tim@193.190.253.114, Ping timeout: 264 seconds)
18:13kjackal_ has left IRC (kjackal_!~quassel@2a02:587:310e:e800:54ed:4fbb:b4a:ed2e, Quit: No Ping reply in 180 seconds.)
18:14kjackal_ has joined IRC (kjackal_!~quassel@2a02:587:310e:e800:9cb7:ab8d:dd80:84c)
18:19TatankaT has joined IRC (TatankaT!~tim@193.190.253.114)
18:35cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 264 seconds)
18:42cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
18:57cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 256 seconds)
19:31GodFather has joined IRC (GodFather!~rcc@2600:1006:b045:7ce2:c9:b45a:59cb:ea17)
19:36Faith has left IRC (Faith!~paty_@unaffiliated/faith, Quit: Leaving)
19:56bennabiy has left IRC (bennabiy!~bennabiy@unaffiliated/bennabiy, Read error: Connection reset by peer)
20:21alkisg_web has left IRC (alkisg_web!4d316cbe@gateway/web/freenode/ip.77.49.108.190, Quit: Page closed)
20:25Statler has left IRC (Statler!~Georg@mail.lohn24.de, Remote host closed the connection)
20:43markit has left IRC (markit!~marco@88-149-177-66.v4.ngi.it, )
20:50lucascastro has left IRC (lucascastro!~lucas@138.185.108.51, Remote host closed the connection)
20:53gehidore is now known as man
21:10vagrantc has joined IRC (vagrantc!~vagrant@67-5-189-55.ptld.qwest.net)
21:10vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
21:14cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
21:17bennabiy has joined IRC (bennabiy!~bennabiy@unaffiliated/bennabiy)
21:35man is now known as gehidore
22:15kjackal_ has left IRC (kjackal_!~quassel@2a02:587:310e:e800:9cb7:ab8d:dd80:84c, Ping timeout: 255 seconds)
22:43cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 260 seconds)
22:49ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)
22:55cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
22:58ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
23:04ben_roose has joined IRC (ben_roose!~roose@roose.cs.wichita.edu)
23:13Guest71032 is now known as lee
23:19vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
23:19ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)
23:41markus_e92 has left IRC (markus_e92!~markus_e9@91-115-18-4.adsl.highway.telekom.at, Ping timeout: 245 seconds)
23:49markus_e92 has joined IRC (markus_e92!~markus_e9@91-115-157-108.adsl.highway.telekom.at)