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


Channel log from 24 March 2022   (all times are UTC)

01:14vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20, Quit: leaving)
01:17Vercas has joined IRC (Vercas!~Vercas@gateway/tor-sasl/vercas)
02:04Vercas has left IRC (Vercas!~Vercas@gateway/tor-sasl/vercas, Ping timeout: 240 seconds)
02:14vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20)
02:18Vercas has joined IRC (Vercas!~Vercas@gateway/tor-sasl/vercas)
02:47vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20, Quit: leaving)
05:32lucascastro has left IRC (lucascastro!~lucascast@192-140-51-239.static.oncabo.net.br, Ping timeout: 260 seconds)
06:20lucascastro has joined IRC (lucascastro!~lucascast@192-140-51-239.static.oncabo.net.br)
07:17ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
07:53wyre has left IRC (wyre!~wyre@user/wyre, Quit: ZNC 1.8.2 - https://znc.in)
07:55wyre has joined IRC (wyre!~wyre@user/wyre)
08:15Hyperbyte has left IRC (Hyperbyte!~jan@middelkoop.cc, Ping timeout: 240 seconds)
09:18ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
09:19ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
10:24wyre has left IRC (wyre!~wyre@user/wyre, Quit: ZNC 1.8.2 - https://znc.in)
10:25wyre has joined IRC (wyre!~wyre@user/wyre)
11:51lucascastro has left IRC (lucascastro!~lucascast@192-140-51-239.static.oncabo.net.br, Remote host closed the connection)
11:53lucascastro has joined IRC (lucascastro!~lucascast@192-140-51-239.static.oncabo.net.br)
16:17vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:7:77:0:20)
16:47lucascastro has left IRC (lucascastro!~lucascast@192-140-51-239.static.oncabo.net.br, Ping timeout: 260 seconds)
20:00lucascastro has joined IRC (lucascastro!~lucascast@192-140-51-239.static.oncabo.net.br)
20:29
<quinox>
alkis: no hurry (I'm not at work so I can't do anything until tomorrow), but do you perhaps know what's wrong when my LTSP22-with-proxydhcp setup works great when I test it with a virtualbox-machine-with-network-in-bridgemode, both legacy BIOS boot and UEFI boot, but fails on real hardware: initial DHCP works, PXE message shows up, iPXE menu shows up (and works), loading of vmlinux works, ltsp.img
20:29
works, initrd.img works, I start to see kernel messages but then DHCP client fails with "No broadcast interfaces found" over and over again
20:29
I tried it on 2 different computers (I buy single computers whenever I need one, so they all differ a bit), and these machines currently happily boot LTSP5
20:30
I think it happens right before it tries to mount /
20:30
https://ghost.qtea.nl/tmp/ltsp22-boot-no-broadcast-interfaces.txt
20:32
I tried disabling IPv6 in my own DHCP server but that didn't seem to fix it
20:33
the network cards are standard on-board thingies
20:34
10.50.0.12 is the LTSP22 server
20:34
IPv4 firewall on it is fully open
20:39
I fully accept it's a problem in my own setup somewhere, I just don't know where to look right now :)
20:39
and booting a VM using this LTSP setup works flawlessly without changing anything, so that's a bit odd
20:42
my LTSP5 setup uses Ubuntu18.04, for LTSP22 I used Ubuntu20.04 as client
21:05
<alkisg>
quinox: I don't think I've seen "no broadcast interfaces" before, but it sounds like it's missing network drivers. E.g. maybe you somehow have kernel 5.13 in the image, but 5.4 in the tftp, so the modules aren't there?
21:06
You can set DEBUG_SHELL=1 in ltsp.conf and get a shell, and run `ip a` to see if it does see an ethernet card, `uname -r` to see the kernel version etc,
21:06
and in the off-case where the card is there and the problem in somewhere in DHCP, you could try KERNEL_PARAMETERS='ip=${ip}:${srv}:${gateway}:${netmask}:${hostname}:' in ltsp.conf
21:14
<quinox>
thanks
21:14
does it still process ltsp.conf on the client if it can't mount / ?
21:15
it comes from ltsp.img, so I'm guessing yes
21:17
I found some post on the internet where this happened for people with dual-stack, but supposedly that was fixed in DHCP Client 4.4.0 and I see 4.4.1 being used
21:17
and I disabled RA in my own DHCP server which didn't fix it
21:20
<alkisg>
Yes, it comes from ltsp.img
21:21
The second DHCP request (ipxe and then initramfs) in ltsp clients isn't very important, so you can solve it with the KERNEL_PARAMETERS that I mentioned above; it goes under [common] in ltsp.conf and you need to run `ltsp ipxe` afterwards
21:21
<quinox>
awesome, thanks
21:23
the UEFI boot seems to do even yet another DHCP request, I see the PXE message twice
21:23ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
21:32
<alkisg>
At most there are three requests: BIOS or UEFI, then iPXE, then initramfs
21:33
If one uses iPXE directly (e.g. some NICs have it; or VM client; or iPXE on disk/CD/win32-loader/grub-ipxe etc), and passes the KERNEL_PARAMETERS, then it's only a single DHCP request
21:33
I asked the iPXE developers if we could reuse the DHCP packet, and they've had good progress in implementing it; I need to see if it's ready for us to use it
21:35
Unfortunately pxelinux/syslinux doesn't support UEFI well so we couldn't keep using it, we did have to switch to iPXE. And grub didn't have broad netbooting support.