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


Channel log from 13 January 2020   (all times are UTC)

00:35douglas_br has left IRC (douglas_br!b1dcb028@177.220.176.40, Remote host closed the connection)
04:12gvy has left IRC (gvy!~mike@altlinux/developer/mike, Ping timeout: 265 seconds)
04:12gvy has joined IRC (gvy!~mike@altlinux/developer/mike)
06:01statler has joined IRC (statler!~Georg@p5B30EE2F.dip0.t-ipconnect.de)
07:04linux_user has joined IRC (linux_user!~Roger@71-38-204-217.lsv2.qwest.net)
07:05
<linux_user>
alkisg: I am here
07:05
I have installed all the packages
07:06
can I do a test (without the GUI) of the installation so far with a second VM booting via PXE?
07:07
this failed:
07:07
root@LTSP:/home/stuart# ltsp dnsmasq
07:07
Failed to get global data: Unit dbus-org.freedesktop.resolve1.service not found.
07:07
LTSP command failed: systemd-resolve --status
07:07
Installed /usr/share/ltsp/server/dnsmasq/ltsp-dnsmasq.conf in /etc/dnsmasq.d/ltsp-dnsmasq.conf
07:07
Restarted dnsmasq
07:09
<alkisg>
(will be back in 10 mins; supporting schools atm..)
07:09
<linux_user>
ok
07:09
I understand!
07:09
time for a snack anyway!
07:10
<alkisg>
!printer
07:10
<ltspbot>
printer: Quick how-to: RCFILE_01="/usr/sbin/jetpipe /dev/usb/lp0 9100 &" in lts.conf, reboot client, then go to server's add printer dialog, and specify network printer → jetdirect → ltsp123.local
07:16
<alkisg>
linux_user: that warning shouldn't matter, it's a problem with debian dependencies in non-gui installations
07:17
They have systemd-resolve but they don't use it or something like that, if I remember correctly
07:17
You may proceeed with the rest
07:17
<linux_user>
ok
07:17
cool
07:18
do I skip managing a client image?
07:26
<alkisg>
linux_user: run `ltsp image /` once, as a test, to see if your clients will be able to netboot into debian without gui
07:26
Once you have that, you move to netbooting isos
08:02
vagrantc: now that Fotis isn't around to complain :P, I'm pondering whether to bump the epoptes version to 20.01, to match ltsp, sch-scripts etc
08:02
If you're even a slightest bit in favor of this, I'm doing it :)
08:03
<vagrantc>
alkisg: i like dateversions :)
08:03
<alkisg>
Yey!
08:03
OK I'll push epoptes today or tomorrow, and I'll leave ltsp for a bit later as I'm working on rpi support
08:03
If we can have them both in unstable or testing until the end of the month, it'll be great
08:04
<vagrantc>
by the end of the month?
08:04
<alkisg>
I mean the upload, not the migration
08:04
<linux_user>
alkisg: Well, do I need to start over? as I did ltsp ipxe and ltsp initrd
08:04
<alkisg>
vagrantc: I think debian import freeze is mid february, so there'll be enough time for migration
08:04
<vagrantc>
alkisg: should be doable if i can get to it this week, which i should
08:04
<alkisg>
linux_user: nah, just rerun ltsp ipxe after ltsp image /
08:04
vagrantc: great :)
08:05
<vagrantc>
alkisg: the new stuff is a lot easier to test :)
08:05
<alkisg>
True; I no longer fear when they'll tell me "this doesn't work in mint or kubuntu"
08:05
I just use an .iso as the image
08:07
<linux_user>
do I rerun the initrd also?
08:08
<alkisg>
linux_user: no
08:09
<linux_user>
ok
08:09
<vagrantc>
alkisg: heh.
08:21vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
08:31statler has left IRC (statler!~Georg@p5B30EE2F.dip0.t-ipconnect.de, Remote host closed the connection)
08:59kjackal has joined IRC (kjackal!~quassel@2a02:587:3115:3000:2dfe:a98:f7:8100)
09:23statler has joined IRC (statler!~Georg@gwrz3.lohn24.de)
09:58markit has joined IRC (markit!~marco@mail.ammdomus.it)
09:59
<markit>
Hi alkisg :) Still experimenting with Kubuntu 20.04 -64 that I want to use in a local school as an "upgrade" from 14.04 - 32 bit. I've seen that for each login 2 .fusehidden* files, 128K, are created and never deleted
10:00
<alkisg>
markit: where?
10:00
<markit>
in the home
10:00
<alkisg>
(hello :))
10:00
What are the exact names?
10:00
<markit>
something along with /home/ltsp101/.fuse_hidden0000009500000003
10:01
# lsof /home/ltsp101/.fuse_hidden0000009500000003
10:01
sftp-serv 1225 ltsp101 35u REG 8,1 131072 655529 /home/ltsp101/.fuse_hidden0000009500000003
10:01
<alkisg>
https://askubuntu.com/questions/493198/what-is-a-fuse-hidden-file-and-why-do-they-exist
10:01
This says they're deleted when the process stops using them
10:01
Do you still have them after the process finished?
10:02
<markit>
if I choose to logout (K menu -> poweroff) they are still there
10:02
<alkisg>
But is the process finished ?
10:02
Because you might logout and still have processes running
10:02
<markit>
the client is turned off...
10:03
<alkisg>
Can I see?
10:03
!vnc-edide
10:03
<ltspbot>
vnc-edide: To share your screen with me, open Epoptes → Help menu → Remote support → Host: srv1-dide.ioa.sch.gr, and click the Connect button
10:03
<markit>
mmmm I should check the server side, but if I shutdown the server, the files still exist... well, let's try
10:03
ok
10:03
it's a VM
10:03
<alkisg>
OK
10:05
<markit>
the first 2 are from a previous session
10:05
<alkisg>
kwin_x11 1419 ltsp101 DEL REG 0,56 123 .fuse_hidden0000007b00000003
10:05
plasmashe 1431 ltsp101 DEL REG 0,56 138 .fuse_hidden0000008a00000004
10:10
<markit>
so seems that if you K -> logout, fuse are deleted, if I K -> poweroff they are not?
10:11
(it's a test VM, user ltsp101, pwd the same)
10:15
<alkisg>
markit: sshfs automatically deletes them when they're no longer in use, but it appears that after logout, kde uses them for several seconds at least,
10:15
so, if you poweroff, sshfs doesn't have a chance to delete them in time
10:17
<markit>
I see... so I need a script to delete the old one before the new one are created, if possibile?
10:17
<alkisg>
markit: I guess you can delete them on login
10:18
<markit>
or could be also at logout, I think rm .fuse* will fail if they are still in use
10:18
<alkisg>
Yeah
10:18
<markit>
I don't care have 2-4 of them in each home, but don't want a ton of them after 300 user have logghed in/out for an year :()
10:19
<alkisg>
A cron job on the server could probably work too
10:20
<markit>
if I want to use ltsp.conf instead, which could be the "best place"?
10:21
<alkisg>
ltsp.conf currently doesn't have hooks for login or logout
10:21
You need a session script for that
10:22
<markit>
ok, I'll use crontab until a better solution is found :) Thanks a lot
10:24
<alkisg>
np
10:26
<markit>
btw, I'm a bit confused... I've changed something in /etc on the server, to have clients access these modifications, is enough an 'ltsp initrd' ?
10:27
<alkisg>
No, you need `ltsp image /`. But /etc/ltsp is a special case, as that one goes via ltsp.img, so it doesn't need `ltsp image /`.
10:28
So if you change things in /etc/ltsp, then you only need: ltsp initrd
10:29
<markit>
thanks :) A little "bug" or I'm missing a package, if I issue "ltsp" the writing is
10:29
[...] To get help with applets and their options, run `man ltsp applet` or `ltsp --help applet`.
10:29
but if I type
10:30
<alkisg>
applet means any ltsp command
10:30
So you'd do: man ltsp dnsmasq
10:30
<markit>
AH, my bad! I feel soooo dumb sometime
10:30
<alkisg>
applet isn't the best name...but I didn't like "subcommands" either...
10:31
(busybox also uses the "applets" word)
10:55
<markit>
let's recap... if I want a file in the DIRECTORY /etc/xdg of the client, I have to create under /etc/ltsp/xdg directory?
11:11
mmm seems not (I've issued ltsp initrd but in the client /etc I've not the new files)
11:22section1 has joined IRC (section1!~section1@178.33.109.106)
11:24douglas_br has joined IRC (douglas_br!bb5f6554@187.95.101.84)
11:25
<douglas_br>
hello!
11:27
Mini PC athlon 200GE with 4 GB Ram DDR4 is it good to be LTSP client?
11:30
<markit>
ok, about my question, let's try "POST_INIT" stuff
11:32
douglas_br: gbit ethernet? Sounds good, PassMark index for that CPU is 4957, better than my current desktop :)
11:33
<douglas_br>
markit yes, it has gigabit
11:34
(y)
11:38
<themeoizr>
douglas_br: 4gb would work, though my preference is 8
11:39
there's no network swapping, so 4gb is pushing it
11:48
<markit>
themeoizr: urgh...not good, I've clients with 3GB that I have to migrate to the new ltsp
12:11
alkisg: in ltsp.conf, under [clients], I've something like: POST_INIT_KDECONF_01="cp -a /etc/ltsp/xdg/* /etc/xdg"
12:12
(yes, I did a ltsp initrd and rebooted the client)
12:13
<alkisg>
(12:55:06 PM) markit: let's recap... if I want a file in the DIRECTORY /etc/xdg of the client, I have to create under /etc/ltsp/xdg directory?
12:13
==> one way is to create /etc/ltsp/file and then have POST_INIT_BLAH="ln -s /etc/ltsp/file /etc/xdg/file"
12:14
douglas_br: yeah sounds fine
12:14
<markit>
alkisg: no cp ? I want to override if a file already exists (due to image creation)
12:15
<alkisg>
themeoizr: oh it's very easy to enable network swapping. The very difficult part is to make swapping work nice in Linux; this won't happen in the next 10 years; kernel developers are preparing early OOM daemons though which might help
12:15
markit: ln -sf overwrites
12:15
ln saves a bit of ram, compared to cp
12:16
<markit>
alkisg: maybe you missed my second question of 13:11, cp did not worked and I'm wondering why
12:17
<alkisg>
markit: run that command from epoptes and see if it works
12:17
(true, I missed that question)
12:18
<markit>
local root you mean?
12:18
<alkisg>
Yes
12:20
<markit>
my bad, I mistyped a name :(
12:22
<themeoizr>
markit: youll be fine if you dont have 30 browser tabs
12:22
alkisg: what are the actual problems with network swapping atm?
12:59douglas_br has left IRC (douglas_br!bb5f6554@187.95.101.84, Remote host closed the connection)
13:05woernie has joined IRC (woernie!~werner@p50867A20.dip0.t-ipconnect.de)
13:25douglas_br has joined IRC (douglas_br!bb5f6554@187.95.101.84)
13:26douglas_br has left IRC (douglas_br!bb5f6554@187.95.101.84, Remote host closed the connection)
13:36themeoizr is now known as ENOKITTIES
13:57
<alkisg>
ENOKITTIES: the main drawback is that if one client really needs swapping, it'll take all of the network bandwidth and much of the server CPU, making the rest of the client unresponsive
13:57
It was the same in LTSP5 as well, it's not a new problem
13:58
<ENOKITTIES>
no way to QoS it down?
13:58
<alkisg>
The secondary problem is that even if a client has local swap etc, i.e. it's not even an LTSP client, and really needs swapping... again, it'll just pretty much hang
13:58
I.e. swap is not really useful in Linux in general atm
13:58
<ENOKITTIES>
so realistically the trajectory is RAM upgrades?
13:58
<alkisg>
So there's no point in trying to solve the first problem, as the second problem means that swap is not useful anyway
13:59
The kernel developers, google, facebook etc, are preparing "early OOM", which means that they'll e.g. kill the browser earlier
13:59
At *that* point, swap will be useful, so we'll be able to use network swap as well, without having to solve the first problem
14:00
For now yeah, if someone needs swap (which means the computer hangs due to lack of memory), yeah, he should buy RAM
14:02statler has left IRC (statler!~Georg@gwrz3.lohn24.de, Remote host closed the connection)
14:04
<ENOKITTIES>
i've observed that on about half of my stations
14:04
and upgrading 4 to 8gb solved it
14:04
but there's a specific profile to this: they tend to have multiple browser tabs open due to how their crm is constructed
14:04
<alkisg>
Were you using NBD_SWAP? This has additional problems, that's why I completely avoided it in the new LTSP
14:04
<ENOKITTIES>
I never used the old ltsp
14:04
<alkisg>
OK
14:05
You could also test this, if you want to: put a local disk or even a usb stick, formatted as linux-swap, which will be automatically used,
14:05
and see if that local linux-swap along with 4 GB RAM solves the issues
14:05
In my tests, that *didn't* work; the crashes persisted
14:10
<ENOKITTIES>
actually my environment would be crap for this because some of the crashes are caused by extremely crappy ethernet wall sockets
14:11
(the idiot contractor ripped them out during renovation and then put them back in half assedly)
14:11
<alkisg>
Ouch, that needs network quality checking devices
14:16
<ENOKITTIES>
no, contractor stabbing devices
14:18
<alkisg>
:D
15:00
<markit>
ENOKITTIES: just yesterday I read an article about how much better is firefox compared to chrome if you have multiple tabs, regarding resources
15:03
<linux_user>
alkisg: everything seems to have finished
15:03
what steps might I wish to take to test it?
16:22woernie has left IRC (woernie!~werner@p50867A20.dip0.t-ipconnect.de, Remote host closed the connection)
16:52douglas_br has joined IRC (douglas_br!bb5f6554@187.95.101.84)
16:55markit has left IRC (markit!~marco@mail.ammdomus.it, )
16:57
<douglas_br>
hello! when client side enter in protect screen if try comeback it does not accept client password. How to fix it?
17:09
<alkisg>
douglas_br: at that point, do you see the client in epoptes, or is it hanged?
17:10
linux_user: you'd just pxe boot a client, see the ltsp.org section about netbooting clients
17:11
<douglas_br>
alkisg I am sorry, its ok
17:11
<alkisg>
np
17:39
<linux_user>
ok
17:47
alkisg: it did not work
17:48
alkisg: vmware fusion said: PXE-E55: ProxyDHCP service did not reply to request on port 4011
17:48
PXE-M0F: Exiting Intel PXE ROM.
17:48
Operating System not Found
17:49
<alkisg>
If you're using VMs, make sure that you use bridged networking and that the VM has full access to ports, even restricted ones
17:49
Or set up internal networking and use a virtual client
18:04kjackal_v2 has joined IRC (kjackal_v2!~quassel@athedsl-238118.home.otenet.gr)
18:04kjackal has left IRC (kjackal!~quassel@2a02:587:3115:3000:2dfe:a98:f7:8100, Ping timeout: 248 seconds)
18:13shored has left IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi, Ping timeout: 268 seconds)
18:14shored has joined IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi)
18:26shored has left IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi, Ping timeout: 265 seconds)
19:01
<linux_user>
alkisg: I used bridged network but I will check and be sure its all correct.
19:01
The client VM did DHCP an IP address it just did not get any PXE responses I think
19:01
brb
19:01
yeah its all setup for bridged network properly since I see it got a DHCP address from the real network 192.168.70.x
19:02section1 has left IRC (section1!~section1@178.33.109.106, Quit: Leaving)
19:04
<alkisg>
linux_user: in the VM, did you run `ltsp dnsmasq`?
19:05
<linux_user>
I think so, let me check
19:05
<alkisg>
Also, run `journalctl -fu dnsmasq` and see if it receives the client DHCP request
19:05
<linux_user>
ok
19:05
1 sec
19:05
<alkisg>
And of course, the VM must have an 192.168.70.x IP then
19:08shored has joined IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi)
19:08
<linux_user>
The VM with LTSP has a 192.168.70.x address, my router runs Gargoyle (which uses dnsmasq also), the VM (client0 got a 192.168.70.x address
19:08
-- Logs begin at Sun 2020-01-12 22:52:53 PST. --
19:08
Jan 13 11:05:34 LTSP dnsmasq[10798]: started, version 2.80 DNS disabled
19:08
Jan 13 11:05:34 LTSP dnsmasq[10798]: compile time options: IPv6 GNU-getopt DBus i18n IDN DHCP DHCPv6 no-Lua TFTP conntrack ipset auth DNSSEC loop-detect inotify dumpfile
19:08
Jan 13 11:05:34 LTSP dnsmasq-dhcp[10798]: DHCP, proxy on subnet 192.168.70.0
19:08
Jan 13 11:05:34 LTSP dnsmasq-dhcp[10798]: DHCP, IP range 192.168.67.20 -- 192.168.67.250, lease time 12h
19:08
Jan 13 11:05:34 LTSP dnsmasq-tftp[10798]: TFTP root is /srv/tftp
19:08
Jan 13 11:05:34 LTSP systemd[1]: Started dnsmasq - A lightweight DHCP and caching DNS server.
19:08
Jan 13 11:05:50 LTSP dnsmasq-dhcp[10798]: PXE(ens33) 00:0c:29:b9:55:9b proxy
19:09
Jan 13 11:05:52 LTSP dnsmasq-dhcp[10798]: PXE(ens33) 00:0c:29:b9:55:9b proxy
19:09
Jan 13 11:05:56 LTSP dnsmasq-dhcp[10798]: PXE(ens33) 00:0c:29:b9:55:9b proxy
19:09
Jan 13 11:06:04 LTSP dnsmasq-dhcp[10798]: PXE(ens33) 00:0c:29:b9:55:9b proxy
19:09
that is from the journalctl -fu output
19:09
not sure what the 192.168.67 nonsense is about though, I have no such range on my network at all
19:09
<alkisg>
!paste
19:09
<ltspbot>
paste: To avoid channel flooding, please upload text longer than 3 lines to http://paste.debian.net. Don't forget to paste the resulting URL here.
19:09
<alkisg>
OK so the server receives the request and replies
19:10
Can you test the client with a recent ipxe.iso?
19:10
E.g. boot.ipxe.org/ipxe.iso
19:10
<linux_user>
sure
19:10
1 sec
19:10
<alkisg>
192.168.67.x is the special ltsp subnet, ignore it, you don't use it
19:36woernie has joined IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de)
19:50GodFather has left IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net, Ping timeout: 268 seconds)
19:54douglas_br has left IRC (douglas_br!bb5f6554@187.95.101.84, Remote host closed the connection)
20:05shored has left IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi, Ping timeout: 240 seconds)
20:20shored has joined IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi)
20:41woernie has left IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de, Remote host closed the connection)
21:04kjackal_v2 has left IRC (kjackal_v2!~quassel@athedsl-238118.home.otenet.gr, Remote host closed the connection)
21:18shored has left IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi, Ping timeout: 265 seconds)
21:20shored has joined IRC (shored!~shored@87-92-122-167.bb.dnainternet.fi)