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


Channel log from 28 January 2022   (all times are UTC)

02:33vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100e)
03:06vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100e, Ping timeout: 240 seconds)
03:42vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100e)
04:36vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100e, Quit: leaving)
05:58quinox has left IRC (quinox!~quinox@ghost.qtea.nl, Quit: WeeChat 3.4)
06:01quinox has joined IRC (quinox!~quinox@ghost.qtea.nl)
06:07Hyperbyte has left IRC (Hyperbyte!~jan@middelkoop.cc, *.net *.split)
06:12Hyperbyte has joined IRC (Hyperbyte!~jan@middelkoop.cc)
06:31woernie has joined IRC (woernie!~werner@p5b296cc2.dip0.t-ipconnect.de)
06:42spectra has left IRC (spectra!~spectra@63.142.251.150, Quit: ZNC - http://znc.sourceforge.net)
07:24woernie has left IRC (woernie!~werner@p5b296cc2.dip0.t-ipconnect.de, Remote host closed the connection)
07:47
<sfxworks[m]>
So I can use the nfs command but im assuming I don't want to pull the actual image....... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/8ad22fe1637e04c37862ae8aa13945c6b812cbb4)
07:47
re: rockpro64 ltsp+pxe booting
07:48
Would I pull intrid or something?
07:49
<alkisg>
sfxworks: you need the kernel and initrd, yes
07:50
But I've no idea about uboot, so I can't provide details :/
07:51
It's probably better to search the net for rockpro64 pxe booting first, and AFTER you found how to do it in general, then I might be able to help you fill the missing bits for LTSP
07:52
Otherwise, try to find vagrantc , he might have some experience in the matter
07:53
<sfxworks[m]>
My friend was able to pxe boot with the setup here https://forum.pine64.org/showthread.php?tid=6814 and flashed the spi for me. I tried flashing the spi with the later build and succeeded but it didn't follow the tftp server addr
07:53
a later build* specifically,
07:54
https://github.com/sigmaris/u-boot/releases
07:54
I might just do a two nic system real quick to replicate his setup...
07:59ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
08:54
<alkisg>
Hi dfanp, currently epoptes isn't scriptable, so you'd need to write a python-twisted program to directly access the epoptes daemon, which isn't trivial
08:55
The epoptes daemon keeps communicating with the clients, so it knows when they go offline
08:59
If you're up to it, you may even make a more generic "epoptes-cli" command out of it :)
09:19woernie has joined IRC (woernie!~werner@p200300cf0714320004f145cbfb748e7a.dip0.t-ipconnect.de)
09:39woernie has left IRC (woernie!~werner@p200300cf0714320004f145cbfb748e7a.dip0.t-ipconnect.de, Ping timeout: 250 seconds)
10:15woernie has joined IRC (woernie!~werner@p200300cf0714320004f145cbfb748e7a.dip0.t-ipconnect.de)
10:19woernie has left IRC (woernie!~werner@p200300cf0714320004f145cbfb748e7a.dip0.t-ipconnect.de, Ping timeout: 250 seconds)
11:51danboid has joined IRC (danboid!~dan@portal.salford.ac.uk)
11:51
<danboid>
Happy Friday LTSPers!
11:54
alkisg, OK so I think this is my final LTSP hurdle, I'm SURE! With just ssh, logins are instant. Just LDAP, logins are snappy. ssh + LDAP = 20+ second delay on ssh logins. Have you seen this before? The internet hasn't, according to my searches. I've done all the ssh server tweaks a man can
11:56
I've tried disabling ssh keys and all sorts. As soon as I disable nslcd, ssh logins are instant so its a LDAP and or PAM thing but the logs aren't helping me much.
11:57
I've posted this problem to the nslcd mailig list but no response yet
11:59
https://lists.arthurdejong.org/nss-pam-ldapd-users/2022/msg00010.html
11:59Faith has joined IRC (Faith!~Faith@user/faith)
12:00
<danboid>
If anyone here can spot anything in there that looks off then I'll love you forever!
12:55
<Hyperbyte>
danboid, are you sure it's not a hostname resolution problem?
12:56
Your LTSP server or clients may be using nameservers that don't properly resolve the internal hostnames.
12:56
<danboid>
Hyperbyte, I've almost fixed it
12:56
The delay is only whe using SSH keys
13:00
Hmm. 14 second login whe using a key isn't almost fixed
13:01
Its fine when I disabled keys and just use passwords, after I rebooted
13:01
ssh had this problem:
13:01
systemd[1]: ssh.service: Found left-over process 1835582 (bash) in control group while starting unit.
13:02
Which was solved by a reboot and I think I've found a fix for that here
13:02
https://serverfault.com/questions/967325/sshd-starts-two-processes-but-systemd-stops-only-one
13:58Nils[m]1234 has joined IRC (Nils[m]1234!~ravagexen@2001:470:69fc:105::1:5087)
15:54woernie has joined IRC (woernie!~werner@p5b296cc2.dip0.t-ipconnect.de)
16:25ltspbot has joined IRC (ltspbot!~supybot@devs.ts.sch.gr)
16:32vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100b)
17:44ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
18:16lynchmike77[m] has joined IRC (lynchmike77[m]!~lynchmike@2001:470:69fc:105::1:8354)
18:17
<lynchmike77[m]>
Does dnsmasq support different subnets? I can't get boot file when client is on a different subnets.
18:18eu^1607223110lig has joined IRC (eu^1607223110lig!~eu^160722@160.72.23.110)
18:18
<eu^1607223110lig>
Does dnsmasq support different subnets? I can't get boot file when client is on a different subnets.
18:23
<alkisg>
lynchmike77: eu^1607223110lig: sure, what's your dnsmasq.conf like?
18:24
Wait, do you mean that your dhcp server is on a different subnet than the dhcp client, or that you have multiple NICs and multiple subnets on one server?
18:30
<lynchmike77[m]>
pxe server works on it's subnet but it doesn't work on different subnet. Cisco switch is providing DHCP info & I have ip helper address.
18:32
dhcp-range=set:proxy,10.122.68.0,proxy,255.255.254.0
18:32
dhcp-range=set:proxy,10.122.72.0,proxy,255.255.254.0
18:33
the first entry works but the second one doesn't
18:33
<alkisg>
Run `journalctl -b -fu dnsmasq`; this will show you the log file. Reboot a client in the other subnet. If you don't see anything in the logs, then the request never reaches dnsmasq.
18:33
That would mean that the other parts are misconfigured, not dnsdmasq
18:34
Jan 29 01:32:57 ltsp-server dnsmasq-dhcp[44898]: no address range available for DHCP request via enp3s0.4
18:34
Jan 29 01:34:02 ltsp-server dnsmasq-dhcp[44898]: DHCPDISCOVER(enp3s0) 00:02:c9:ed:8c:30
18:35
These are examples of dhcp requests that actually reach dnsmasq; in the first one, dnsmasq won't reply on purpose, in the second it replies with an offer
18:35
If you don't see anything at all, it means that your other parts are not configured properly to propagate the needed information
18:48
<lynchmike77[m]>
I didn't see any logs when I reboot the client on a different subnet
18:48
wireshark didn't see any boot file name entry from pxe server
18:50
<alkisg>
Right, so the rest of your network is misconfigured, it doesn't pass the dhcp request to the dnsmasq server
18:51
What software are you using for "ip helper"?
18:51
<lynchmike77[m]>
Cisco c9500 switch
18:52
<alkisg>
Try to ask in cisco forums etc, for how to properly configure it; unfortunately I've never seen the web (or cli) interface of such switches...
18:52
If you can't do that, consider providing a local proxydhcp server
18:52
(to the other subnet)
18:53
<lynchmike77[m]>
I had a cisco tac engineer capture packets from both end and no such luck
18:54
what do you mean by providing a local proxydhcp server?
18:54
for each different subnets, just have another pxe server?
18:54
<alkisg>
You'd setup just dnsmasq in any kind of pc or openwrt router etc, and use it as a local realdhcp or proxydhcp there
18:55
That would then point to the ltsp server for tftp
18:55
(=pxe)
18:55
dnsmasq is very light so you can run it in an rpi-zero or an old router or anywhere...
18:58
<lynchmike77[m]>
Thank you
18:58
<alkisg>
You're welcome
18:59
<lynchmike77[m]>
so dnsmasq works in different subnets in your environment?
18:59
<alkisg>
The meaning of subnets for me is different than yours
18:59
<lynchmike77[m]>
10.122.68.0 is a subnet
18:59
<alkisg>
E.g. one server can properly serve 10 different subnets, if it has 10 NICs or 10 IPs
18:59
<lynchmike77[m]>
10.122.72.0 is a different subnet
18:59
<alkisg>
There's no DHCP relaying involved, like in your case
19:00
If you put two NICs to your LTSP server= dnsmasq server, one .68 and one .72, you don't need a helper
19:00* vagrantc forgot to upload ltsp to debian :)
19:00
<alkisg>
So there's nothing to configure in the switches at all
19:01
<lynchmike77[m]>
I have tons of subnets
19:01
<alkisg>
You can add tons of IPs in a single dnsmasq NIC
19:01
Are you using vlans, or just subnets?
19:02
<lynchmike77[m]>
vlans
19:02
<alkisg>
You can add e.g. 100 vlans = 100 IPs in your LTSP server, each one for a vlan/subnet
19:02
Then if you run `ltsp dnsmasq`, it'll probably be configured as a proxydhcp in all these subnets
19:03
<lynchmike77[m]>
do you have a sample config and which file do I modify those entries? thanks.
19:03
<alkisg>
In my example above, `no address range available for DHCP request via enp3s0.4`, that .4 is the 4th vlan; I don't want it to be a proxydhcp there, but that's just a line in dnsmasq.conf
19:04
Personally I'm using network manager to add vlans+IPs to a single NIC
19:04
I find it easier than the netplan/networkd-systemd equivalents
19:04
<lynchmike77[m]>
I'm using ltsp-dnsmasq.conf
19:04eu^1607223110lig has left IRC (eu^1607223110lig!~eu^160722@160.72.23.110, Quit: Client closed)
19:04
<alkisg>
Edit connections (nm-connection-editor), add vlan, enter the IP, that's all
19:05
After you have added the vlans and IPs in your ltsp server, you just run `ltsp dnsmasq` without any options
19:05
The configuration you need is in network-manager, not in dnsmasq/ltsp
19:05
You just need multiple IPs, one for each subnet
19:06
Plenty of time till Ubuntu feature freeze, no hurry! Also upload epoptes though :D
19:06
Thanks vagrantc !
19:07
I deployed these versions to schools and businesses that I could reach, I heard of no issues so far
20:24vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100b, Ping timeout: 240 seconds)
20:31vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100b)
20:32Faith has left IRC (Faith!~Faith@user/faith, Quit: Leaving)