Hello,
I have been given a couple of Termtek TK-3550 thin clients and I am trying to set them up to open into my Ubuntu 20.04 server at my home over our ethernet LAN. These devices don't have much memory so I have been following the instructions to install the "Classic, NFS- filesystem-based X2Go TCE". I have setup a Virtual Box VM, running Ubuntu 20.04 with a bridged network adapter and a permanent IP address, to be the pxe server. I have followed the instructions as best I was able but I am still having difficulties.
I think one problem is likely to be that our cheap router, which provides our DHCP server, does not allow me to nominate the "next- server" or the filename. In an attempt to address this I have installed dnsmasq on the pxe server VM and am running dnsmasq. The dnsmasq.conf is as follows: port=0 log-dhcp dhcp-range=192.168.1.0,proxy dhcp-boot=pxelinux.0 enable-tftp tftp-root=/srv/tftp (I have also tried tftp- root=/opt/x2gothinclient/chroot on this line)
The thin client is set to boot over LAN and the dhcp-log on the pxe server looks like this when I turn the thin client on:
I'm clearly a complete novice but it seems the thin client is finding its way to the pxe server. Unfortunately the thin client trys to boot until it times out, reporting: "No boot filename received". It then just boots its decades old version of Windows CE. Note that when I changed the tftp-root path in dnsmasq from /opt/x2gothinclient/chroot to /srv/tftp it had some effect on the dhcp log (it nominated its terminal id in a second line) but the nett result did not change. I should add that I have exported both /opt/x2gothinclient/chroot and /srv/tftp with the NFS server. I only switched to /srv/tftp because that's where pxelinux.0 is and I assume that is the file it is looking for and failing to find.
Basically I have been trying to get this working for a couple of days without success, I have decided to seek expert help. I apologise for my ignorance, I understand that hobbists like myself are probably incredibly annoying.
Regards Stuart MacGregor
Hello,
I posted an enquiry quite some time ago but it hasn't attracted anyone's interest. In the meantime I have contiued trying to solve the problem on my own. I moved all my dchp functions from my cheap, rubbish router to my ubuntu 20.04 server. I spent some time fiddling with the settings of isc-dhcp-server. I downloaded ipxe undionly.kpxe and put it in my tftp server (a virual box virtual machine with bridged netwrking also running a ubuntu 20.04 server) at /srv/tftp. When I boot up my Termtek TK3550 Thin Client (AMD GX III LX800 CPU, 256 MB ram, Intel UNDI network card PXE 2.0) undionly.kpxe is successfully booted via tftp and the new bootloder finds pxelinux.0 then just locks up. Nothing happens and it continues to happen indefinitely. I have attached a photo of the boot screen when it locks up.
I have created a directory of the tftp server /srv/tftp/pxelinux.cfg . I copied /srv/tftp/default.cfg into this directory both as default.cfg and as simply default. I also copied x2go-tce-cfg, memtest.cfg and local-boot.cfg into this directory just in case. I am now at a loss. I have no way of checking/altering pxelinux.0 as it is in binary or code of some sort. I have spent hours in my free time to get this far, but in reality the end goal of logging into my linux server on a thin client seems no closer. Any help would be appreciated.
Stuart MacGregor macgregorathome@iinet.net.au
Thank you for the excellent advice. I switched the virtual machine acting as my tftp server to Debian 11 and pxe boot started. I didn't event need to chainload ipxe with undionly.kpxe. Basically using Ubuntu for my tftp server was the problem. Something of note, though. It was necessary create a new directory /srv/tftp/pxelinux.cfg and copy /srv/tftp/default.cfg into it. I discovered this by consulting professor Google. I did not read this in the classic X2go-tce documentation. Perhaps this might be explictly added to this documention to help nubes like myself? I am having more problems, though. The first, red X2go splash screen comes up but I am unable to make any selections, not with the highlighted hotkeys nor with the arrow keys or even TAB. Not a real problem, after timeout it defaults to x2go-tce which is what I want anyway. I just mention it in case it is a symptom of the other problems I'm about to describe. The boot soon fails with the following messages: [ 16.996870] Initramfs unpacking failed : write error[ 17.488013] mce: Unable to init MCE device (rc: -5)[ 19.959922] Failed to execute /init (error -2)[ 19.190964] Kernel panic - not syncing: No working init found. Try passing init =option to Kernel. See Linux Documentation/admin-guide/initrst for guidance.[ 19.961064] CPU: 0 PID : 1 Comm: swapper/0 Not tainted 5.10.0-10-686 #1 Debian 5.10.84-1[ 19.961135] Hardware name: /TK-3550, BIOS 6.00 PG 02/27/2009[ 19.961198] Call Trace:[ 19.961283] dump_stack+0x54/0x68[ 19.961355] ? rest_init+0x30/0x97[ 19.961429] panic+0xa3/0x25b[ 19.961495] ? rest_init+0xe0/0xe9[ 19.961630] ret_from_fork+0x1c/0x28[ 19.961705] Kernel Offset: disabled[ 19.961775] ---[ end Kernel panic - not syncing: No working init found. Try passing init =option to Kernel. See Linux Documentation/admin-guide/initrst for guidance. ]--- Then the whole thing hangs up indefinitely. It seems to me that the old thin client (TK-3550 AMD GX III LX800 CPU, 256 MB ram) needs a legacy BIOS boot and this is unavailable for some reason. Should I redo my X2go-tce setup (x2gothinclient_settings) and use an older version of Debian to create the chroot image? I used the default "Buster". If I need to go back enough versions of Debian will I need to change to archived repositories in the same file (rather than just http://ftp.debian.ors/debian <distr-codename> main)? Do I need to add to the APPEND line of x2go-tce.cfg instead with init=/sbin/.... or some such (professor Google again)? Once again any help here would be appreciated. I recognise that the thin client I am using is rubbish but the price was right and having come so far I'm reluctant to give up now.
On Mon, 2022-01-10 at 10:01 +0300, ovt@zaokknm.ru wrote:
At one time, when, about a year ago, while installing x2go on ubuntu, I faced about the same problem. Nevertheless, in a week I was able to run x2go-server + pxe. Endless poking around in the configuration files gave me the idea to change the distribution kit. Imagine my surprise when, in 30-40 minutes, I installed and configured this bundle on Debian_10 without any problems and have already been using it with 20 clients for a whole year. I also did not find understanding and help in the x2go community. My advice to you is to change your distribution to Debian.
05.01.2022, 15:38, "MacGregor Family" <macgregorathome@iinet.net.au>:
Hello,
I posted an enquiry quite some time ago but it hasn't attracted anyone's interest. In the meantime I have contiued trying to solve the problem on my own. I moved all my dchp functions from my cheap, rubbish router to my ubuntu 20.04 server. I spent some time fiddling with the settings of isc-dhcp-server. I downloaded ipxe undionly.kpxe and put it in my tftp server (a virual box virtual machine with bridged netwrking also running a ubuntu 20.04 server) at /srv/tftp. When I boot up my Termtek TK3550 Thin Client (AMD GX III LX800 CPU, 256 MB ram, Intel UNDI network card PXE 2.0) undionly.kpxe is successfully booted via tftp and the new bootloder finds pxelinux.0 then just locks up. Nothing happens and it continues to happen indefinitely. I have attached a photo of the boot screen when it locks up.
I have created a directory of the tftp server /srv/tftp/pxelinux.cfg . I copied /srv/tftp/default.cfg into this directory both as default.cfg and as simply default. I also copied x2go-tce-cfg, memtest.cfg and local-boot.cfg into this directory just in case. I am now at a loss. I have no way of checking/altering pxelinux.0 as it is in binary or code of some sort. I have spent hours in my free time to get this far, but in reality the end goal of logging into my linux server on a thin client seems no closer. Any help would be appreciated.
Stuart MacGregor macgregorathome@iinet.net.au ,_______________________________________________ x2go-user mailing list x2go-user@lists.x2go.org https://lists.x2go.org/listinfo/x2go-user
Thank you for this info, I've had a little break and now on back on the project. As far as I can see these lines have been added to the x2gothinclient_create script already, perhaps when you reported the problem. I have attached a couple of screen shots showing them. Are these lines now correct or am I missing something?
I recreated the thin client using "stretch" and the boot progressed further, but was still unsuccessful. This time the error message was: [FAILED] Failed to start. Create volatile files and directories
I changed the operating system on the thin client to Tiny Core Linux and it sped up the whole process, however reached the same conclusion.I'm going to try to base the thin client on "Jessie" and change all the repository references to the archived ones, however I don't hold much hope. I looked for an x2goclient package for Tiny Core however it looks like if I want one I need to create the Tiny Core extension myself and this may well be beyond my limited competance. Might give it a go if the all else fails.
Thanks again for replying.
Regards Stuart MacGregor
On Wed, 2022-01-19 at 22:30 +0300, ovt@zaokknm.ru wrote:
I remember this trick with the pxelinux.cfg folder. :-)) If you are at this stage, then this is only the beginning of the journey. I do not know how to find my correspondence in the x2go-user thread, so I quote from my letter -
Without waiting for an answer, I report the errors found in the /usr/sbin/x2gothinclient_create script. Skipped lines (206-207) [ $? -eq 0] && apt-get install gnupg -y [ $? -eq 0] && apt-key adv --recv-keys --keyserver keys.gnupg.net E1F958385BFE2B6E (251)[ $? -eq 0] && systemctl enable x2gothinclient-displaymanager
After adding these lines, x2gothinclient loads fine. Please make these corrections.
P.S.The documentation says that the thin client RAM must be at least 512 MB. I haven't tried it with low RAM, but it makes sense
12.01.2022, 14:12, "MacGregor Family" <macgregorathome@iinet.net.au>:
Thank you for the excellent advice. I switched the virtual machine acting as my tftp server to Debian 11 and pxe boot started. I didn't event need to chainload ipxe with undionly.kpxe. Basically using Ubuntu for my tftp server was the problem.
Something of note, though. It was necessary create a new directory /srv/tftp/pxelinux.cfg and copy /srv/tftp/default.cfg into it. I discovered this by consulting professor Google. I did not read this in the classic X2go-tce documentation. Perhaps this might be explictly added to this documention to help nubes like myself?
I am having more problems, though. The first, red X2go splash screen comes up but I am unable to make any selections, not with the highlighted hotkeys nor with the arrow keys or even TAB. Not a real problem, after timeout it defaults to x2go-tce which is what I want anyway. I just mention it in case it is a symptom of the other problems I'm about to describe. The boot soon fails with the following messages:
[ 16.996870] Initramfs unpacking failed : write error [ 17.488013] mce: Unable to init MCE device (rc: -5) [ 19.959922] Failed to execute /init (error -2) [ 19.190964] Kernel panic - not syncing: No working init found. Try passing init =option to Kernel. See Linux Documentation/admin- guide/initrst for guidance. [ 19.961064] CPU: 0 PID : 1 Comm: swapper/0 Not tainted 5.10.0-10- 686 #1 Debian 5.10.84-1 [ 19.961135] Hardware name: /TK-3550, BIOS 6.00 PG 02/27/2009 [ 19.961198] Call Trace: [ 19.961283] dump_stack+0x54/0x68 [ 19.961355] ? rest_init+0x30/0x97 [ 19.961429] panic+0xa3/0x25b [ 19.961495] ? rest_init+0xe0/0xe9 [ 19.961630] ret_from_fork+0x1c/0x28 [ 19.961705] Kernel Offset: disabled [ 19.961775] ---[ end Kernel panic - not syncing: No working init found. Try passing init =option to Kernel. See Linux Documentation/admin-guide/initrst for guidance. ]---
Then the whole thing hangs up indefinitely. It seems to me that the old thin client (TK-3550 AMD GX III LX800 CPU, 256 MB ram) needs a legacy BIOS boot and this is unavailable for some reason. Should I redo my X2go-tce setup (x2gothinclient_settings) and use an older version of Debian to create the chroot image? I used the default "Buster". If I need to go back enough versions of Debian will I need to change to archived repositories in the same file (rather than just http://ftp.debian.ors/debian <distr-codename> main)? Do I need to add to the APPEND line of x2go-tce.cfg instead with init=/sbin/.... or some such (professor Google again)? Once again any help here would be appreciated. I recognise that the thin client I am using is rubbish but the price was right and having come so far I'm reluctant to give up now.
On Mon, 2022-01-10 at 10:01 +0300, ovt@zaokknm.ru wrote:
At one time, when, about a year ago, while installing x2go on ubuntu, I faced about the same problem. Nevertheless, in a week I was able to run x2go-server + pxe. Endless poking around in the configuration files gave me the idea to change the distribution kit. Imagine my surprise when, in 30-40 minutes, I installed and configured this bundle on Debian_10 without any problems and have already been using it with 20 clients for a whole year. I also did not find understanding and help in the x2go community. My advice to you is to change your distribution to Debian.
05.01.2022, 15:38, "MacGregor Family" < macgregorathome@iinet.net.au>:
Hello,
I posted an enquiry quite some time ago but it hasn't attracted anyone's interest. In the meantime I have contiued trying to solve the problem on my own. I moved all my dchp functions from my cheap, rubbish router to my ubuntu 20.04 server. I spent some time fiddling with the settings of isc-dhcp-server. I downloaded ipxe undionly.kpxe and put it in my tftp server (a virual box virtual machine with bridged netwrking also running a ubuntu 20.04 server) at /srv/tftp. When I boot up my Termtek TK3550 Thin Client (AMD GX III LX800 CPU, 256 MB ram, Intel UNDI network card PXE 2.0) undionly.kpxe is successfully booted via tftp and the new bootloder finds pxelinux.0 then just locks up. Nothing happens and it continues to happen indefinitely. I have attached a photo of the boot screen when it locks up.
I have created a directory of the tftp server /srv/tftp/pxelinux.cfg . I copied /srv/tftp/default.cfg into this directory both as default.cfg and as simply default. I also copied x2go-tce-cfg, memtest.cfg and local-boot.cfg into this directory just in case. I am now at a loss. I have no way of checking/altering pxelinux.0 as it is in binary or code of some sort. I have spent hours in my free time to get this far, but in reality the end goal of logging into my linux server on a thin client seems no closer. Any help would be appreciated.
Stuart MacGregor macgregorathome@iinet.net.au
,
x2go-user mailing list x2go-user@lists.x2go.org https://lists.x2go.org/listinfo/x2go-user