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


Channel log from 9 August 2019   (all times are UTC)

00:07vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
05:06ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
05:27os_a has joined IRC (os_a!~Thunderbi@195.112.116.22)
05:28statler has joined IRC (statler!~Georg@p548978DB.dip0.t-ipconnect.de)
08:07shored has left IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi, Read error: Connection reset by peer)
08:07shored has joined IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi)
08:17adrianor1 has joined IRC (adrianor1!~adrianorg@187.113.244.70)
08:20adrianorg has left IRC (adrianorg!~adrianorg@179.177.213.74.dynamic.adsl.gvt.net.br, Ping timeout: 268 seconds)
08:26kjackal has joined IRC (kjackal!~quassel@ppp-94-66-223-118.home.otenet.gr)
09:54yanu has left IRC (yanu!~yanu@178-116-55-218.access.telenet.be, Remote host closed the connection)
09:58shored has left IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi, Read error: Connection reset by peer)
09:58shored has joined IRC (shored!~shored@87-92-64-8.bb.dnainternet.fi)
10:24statler_ has joined IRC (statler_!~Georg@gwrz.lohn24.de)
10:41kjackal has left IRC (kjackal!~quassel@ppp-94-66-223-118.home.otenet.gr, Ping timeout: 248 seconds)
11:13kjackal has joined IRC (kjackal!~quassel@2a02:587:9a0e:ef00:9178:8d0f:d9b7:4f16)
11:47woernie has joined IRC (woernie!~werner@2001:638:408:2013:b428:a14b:222:ec78)
11:49Faith has joined IRC (Faith!~Paty_@2001:12d0:2080::231:49)
11:49Faith has joined IRC (Faith!~Paty_@unaffiliated/faith)
11:54pavars has joined IRC (pavars!5509db4e@85.9.219.78)
11:58os_a has left IRC (os_a!~Thunderbi@195.112.116.22, Remote host closed the connection)
11:59os_a has joined IRC (os_a!~Thunderbi@195.112.116.22)
11:59
<pavars>
Hi, we're having issues with LTSP. I did a fresh install of Xubuntu (proved to be faster than Mate), installed ltsp, configured it as chrootless and thin clients (old hw so fat clients is not an option, server is beefy enough). Also these are ldap users not local and we're using LVM for storage management. The problem that we have - as soon as more
11:59
than 13 people log on they start experiencing issues. Some cannot open chrome or firefox or any other app - getting error "Resource temporary unavailable" and some couldn't log in - getting "ssh broken pipe" error. Also when executing commands on server after 13 people log in, we're also seeing root@ltsp-riga ~ # htop-bash: fork: retry: Resource t
11:59
emporarily unavailable-bash: fork: retry: Resource temporarily unavailable
12:00
htop looks good load average with everyone logged in was 4.5 (from 24 available threads)
12:00
and memory usage was 8/32 Gb
12:00
and memory usage was 8 of 32 Gb
12:01
some errors from .xsession-errors
12:01
[23870:23902:0809/125206.962071:ERROR:platform_thread_posix.cc(129)] pthread_create: Resource temporarily unavailable (11)
12:02
really banging my head against a wall here
12:05pavars_ has joined IRC (pavars_!~pavars@85.9.219.78)
12:30section1 has joined IRC (section1!~section1@178.33.109.106)
13:04
<pavars>
am I missing some vital step here?
13:09
<mwalters>
pavars: what's the output of `ps -elfT | wc -l` ?
13:09
on the server
13:09
also: `ps -elfT | grep firefox | wc -l` and `ps -elfT | grep chrome | wc -l`
13:10
(the second two will get you threadcounts for the browsers themselves)
13:10
I think you're going to find that the counts are HUUGUGE
13:15pavars_ has left IRC (pavars_!~pavars@85.9.219.78, Remote host closed the connection)
13:15pavars has left IRC (pavars!5509db4e@85.9.219.78, Remote host closed the connection)
13:15
<mwalters>
`ulimit -u` will show you your max thread count
13:16pavars has joined IRC (pavars!~pavars@85.9.219.78)
13:16
<mwalters>
assuming that you're hitting your max thread count... back when I did thin clients, we'd use ltsp-localapps for browsers. even with old hw the browser will likely run better on the client
13:18
also, post the HW specs for the clients... if they were built in the last 10 years... they'll probably work fine as fat clients.
13:19
<pavars>
mostly we have Fujitsu S550 as thin clients
13:20
<mwalters>
sempron! wow!
13:20
but yeah, those might be pushing it for fat clients ;)
13:21
<pavars>
yeah, it's a single core cpu which is pretty much the bottleneck
13:21
opening firefox on a fat client takes ages and cpu maxes out to 100%
13:21
<mwalters>
We had a few intel atom boxes @ 1.6ghz (I think) left when we moved from thin to fat... they were pretty bad ;)
13:22
<pavars>
root@ltsp-riga /home # ulimit -u
13:22
127991
13:22
root@ltsp-riga /home # ps -elfT | wc -l
13:22
1021
13:22
I'm logged in with only 1 user
13:22
and the thread count raised by like 300
13:22
with a lot of tabs open in both chrome and ff
13:22
<mwalters>
yeah, you'll want to do the thread count when you're encountering the issue
13:23
chrome/firefox seem to create them expoentially ;)
13:24
<pavars>
ok, so I can increase the thread count by raising ulimits?
13:24
<mwalters>
yes... I've never had to venture there, so you'll be googling on your own ;)
13:24
<pavars>
in this fiile /etc/security/limits.conf
13:25
<mwalters>
but it certainly sounds like you're hitting thread limits
13:25
and firefox and chrome sure do like their threads... wonder what they're weaving ;)
13:26
<pavars>
:D
13:26
weaving your data that is
13:27
would it be the culprit for also broken pipe error when trying to log in?
13:27
<mwalters>
It certainly could be. Look for sshd errors in syslog when that happens. I bet it can't create the process for the user
13:28
#linux or (I'm assuming...) #ubuntu should be able to help more with upping the thread count and provide better insight. This isn't nessecarily an ltsp problem, specifically
13:33
<pavars>
makes sense
13:35
Aug 9 12:48:35 ltsp-riga sshd[21471]: Accepted password for jrudzite from 10.4.1.32 port 58304 ssh2
13:35
Aug 9 12:48:35 ltsp-riga sshd[21471]: pam_unix(sshd:session): session opened for user jrudzite by (uid=0)
13:35
Aug 9 12:48:35 ltsp-riga sshd[21471]: fatal: fork of unprivileged child failed
13:41
<alkisg>
xfce faster than mate? Do you mean for thin clients only? Did you have disabled composition in both of them, or just one of them?
13:50yanu has joined IRC (yanu!~yanu@178-116-55-218.access.telenet.be)
13:54
<pavars>
# Don't encrypt X traffic. Gives much faster graphics at a security cost.
13:54
LDM_DIRECTX=True
13:54
do you mean ?
13:55
for thin clients it is faster, the interface is much smoother and loads faster than mate, also no ghosting when switching tabs, windows, apps etc unlike Mate
13:55os_a has left IRC (os_a!~Thunderbi@195.112.116.22, Quit: os_a)
13:55
<alkisg>
pavars: no, I mean this: $ gsettings list-recursively | grep -i composi
13:55
org.gnome.metacity compositing-manager true
13:55
This is an option in the GUI too
13:56
So if you enable or disable compositing, the performance is a lot different
13:56
Usually in thin clients I disable compositing
14:01
<pavars>
root@ltsp-riga /home # gsettings list-recursively | grep -i comp
14:01
org.gnome.software compatible-projects ['GNOME', 'KDE', 'XFCE']
14:01
org.onboard layout 'Compact'
14:01
org.mate.engrampa.general compression-level 'normal'
14:01
<alkisg>
In xfce they might call it another way, or it might be always enabled
14:02
But, to test mate in similar terms, you'd need to disable it
14:13pavars has left IRC (pavars!~pavars@85.9.219.78, Remote host closed the connection)
14:15pavars has joined IRC (pavars!~pavars@85.9.219.78)
14:18Geckopia has joined IRC (Geckopia!addaf389@173-218-243-137.pkbgcmtk01.res.dyn.suddenlink.net)
14:18
<Geckopia>
I have a quick question. first time here.
14:18
<alkisg>
Hello then :)
14:20
<Geckopia>
Thank you. I'll straight forward. I use LTSP in my home. Life is easier that way :P Anyways, I have an odd situation. If one of my clients shuts down the computer without logging out, I have to reset their password on the server before they can log in again. I realized they need to log out propertly, but any clue why?
14:20
<alkisg>
Thin or fat clients?
14:20
<Geckopia>
Fat
14:21
<alkisg>
Did the user change the password when he logged in?
14:21
<Geckopia>
no
14:21
<alkisg>
Which desktop environment, e.g. mate, and which distro/version?
14:22
<Geckopia>
they log in. If they log out, all is well. If the cat jumps up on the computer and hits the ^&**()(*& power button, that's what's been happening. I'm running Ubuntu/Epoptes/Mate 64bit
14:24
<alkisg>
Ah
14:24
Hmmm no, fat clients, that's not it... Hmm...
14:27
Geckopia: try to shutdown the client to reproduce the problem; then try "sudo killall -u theuser" to kill his processes; then boot and try to login without changing password
14:28
It sounds very strange, but maybe you have something that keeps some process in use, I can't imagine anything else on fat clients...
14:36pavars has left IRC (pavars!~pavars@85.9.219.78, Remote host closed the connection)
14:42pavars has joined IRC (pavars!~pavars@85.9.219.78)
14:48pavars has left IRC (pavars!~pavars@85.9.219.78, Remote host closed the connection)
14:53kjackal has left IRC (kjackal!~quassel@2a02:587:9a0e:ef00:9178:8d0f:d9b7:4f16, Remote host closed the connection)
14:55kjackal has joined IRC (kjackal!~quassel@2a02:587:9a0e:ef00:15e4:a34b:3780:8936)
15:23woernie has left IRC (woernie!~werner@2001:638:408:2013:b428:a14b:222:ec78, Remote host closed the connection)
15:28woernie has joined IRC (woernie!~werner@194.95.66.31)
15:39pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
15:43pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Ping timeout: 245 seconds)
16:05woernie has left IRC (woernie!~werner@194.95.66.31, Remote host closed the connection)
16:17woernie has joined IRC (woernie!~werner@194.95.66.31)
16:34woernie has left IRC (woernie!~werner@194.95.66.31, Remote host closed the connection)
16:36statler_ has left IRC (statler_!~Georg@gwrz.lohn24.de, Remote host closed the connection)
16:45alexxtasi[m] has left IRC (alexxtasi[m]!alexxtasim@gateway/shell/matrix.org/x-xequehrsfepufixs, Write error: Connection reset by peer)
16:55alexxtasi[m] has joined IRC (alexxtasi[m]!alexxtasim@gateway/shell/matrix.org/x-wpauwvtkboyepxyj)
17:28kjackal has left IRC (kjackal!~quassel@2a02:587:9a0e:ef00:15e4:a34b:3780:8936, Ping timeout: 252 seconds)
17:39ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
19:01vsuojanen has left IRC (vsuojanen!~vsuojanen@cable-hml-585686-205.dhcp.inet.fi, Ping timeout: 245 seconds)
19:29statler has left IRC (statler!~Georg@p548978DB.dip0.t-ipconnect.de, Remote host closed the connection)
19:40
<Geckopia>
alkisg, Not sure how to do that, since I'm running chrooted clients. from which terminal am I running this command, the main server?
19:41
<alkisg>
Geckopia: yes, you run it on the server to ensure that there are no processes still running for that user
19:54section1 has left IRC (section1!~section1@178.33.109.106, Quit: Leaving)
20:01
<Geckopia>
No my other client doesn't do that. I read somewhere that if the fat client was built before the account was made, I may have to just rebuild the client. Is there any truth to that?
20:02
<alkisg>
None at all
20:02
<Geckopia>
client image*
20:02
https://sourceforge.net/p/ltsp/mailman/message/35091197/
20:02
granted, a different issue I think
20:30kjackal has joined IRC (kjackal!~quassel@2a02:587:9a0e:ef00:15e4:a34b:3780:8936)
20:48Faith has left IRC (Faith!~Paty_@unaffiliated/faith, Quit: Leaving)
21:04vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
21:21statler has joined IRC (statler!~Georg@p548978DB.dip0.t-ipconnect.de)
21:31statler has left IRC (statler!~Georg@p548978DB.dip0.t-ipconnect.de, Remote host closed the connection)
21:49pavars has joined IRC (pavars!~pavars@balticom-198-107.balticom.lv)
21:54pavars has left IRC (pavars!~pavars@balticom-198-107.balticom.lv, Ping timeout: 245 seconds)
23:20kjackal has left IRC (kjackal!~quassel@2a02:587:9a0e:ef00:15e4:a34b:3780:8936, Ping timeout: 252 seconds)
23:35Geckopia has left IRC (Geckopia!addaf389@173-218-243-137.pkbgcmtk01.res.dyn.suddenlink.net, Remote host closed the connection)