Lite Client connection failure
A partition and disk imaging/cloning program
Brought to you by:
steven_shiau
Menu
▾
▴
I'm trying to get a client PC to connect to my Clonezilla Lite Server - running in "interactive-client" mode - and, even though I connect successfully, eventually I get a 404 Not Found error for a file called ocs-client-run.conf [see attached picture] Running current version 3.1.1-27 (Debian). The PCs involved are between 4th-6th generation Intel, using built-in NIC's.
The log file mentioned [/var/log/clonezilla.log] doesn't exist, either.
Last edit: T4T Tech Support 2024-01-18
P.S. I note that the file "ocs-client-run.conf" appears to be deleted in line 90 of this script /usr/sbin/ocs-live-get-img ... I wonder if that's why the error crops up in line 110?
Is your Clonezilla lite server running the IP address as 192.168.168.22?
Make sure you only have one PXE or DHCP server in your LAN.
Also make sure there is no any firewall blocking the http service.
Steven
Yes, that is the server IP address. There is a pre-existing DHCP server so I used the option to keep using that, and allow it to assign the IP address of the Clonezilla lite-server. That aspect of things works well, as can be seen from the picture in the original post: the client is finding the server just fine. The server's telling the client "the file you're looking for isn't here" ... and doing an ls -l on the /tmp directory of the server, it sure isn't there.
In point of fact, I even removed the connection to that DHCP server and made the network truly a closed circuit, making my Clonezilla server the IP address generator. Same exact error, same exact way it happened.
Which brings me again to the /usr/sbin/ocs-live-get-img script. As the picture below shows, this file is deleted in line 90, but it seems that it's not regenerated, so by the time we get to line 110, the client crashes and gives the error message pictured in the original post. What is going on between lines 91-109?
Last edit: T4T Tech Support 2024-01-22
When you encounter this issue, you can enter command line prompt, then run:
wget -P /tmp http://192.168.168.22/ocs-client-run.conf
to see the results.
You can use some of the diagnose tools, e.g., telnet or nc to check the remote service.
Just google to find those tools, and some of them are included in the Clonezilla live, e.g.,
https://www.fosslinux.com/35730/5-ways-to-check-if-a-port-is-open-on-a-remote-linux-pc.htm
Steven
OK. So this time I disconnected the external DHCP server from the switch, resulting in a closed internal LAN. So I set Clonezilla to be the DHCP server, and it auto-created the hostname 192.168.169.250.
The client once again crashed in the exact same way described above. So going to the command line and following your instructions, pictured is what happened.
The client was able to see the server's port 80 as open - but the .conf file in the wget command doesn't exist: 404 Not Found error. Same exact file that line 110 in the script is looking for ... and I just noticed, this is the exact command executed in line 91 of the script - it must be failing there, too, and that's where the error messages are probably coming from.
I even tried to use the wget command with the trailing slash [i.e. "/tmp/"] and got the exact same result.
I did a "find" over the whole server for this .conf file; it's not there anywhere (see IMG_1845.JPG below).
Last edit: T4T Tech Support 2024-01-23
You should be able to locate this file on the server:
/var/www/html/ocs-client-run.conf
In addition, the http service is running as:
root@debian:~# ps -efwww |grep lighttpd
www-data 19786 1 0 12:33 ? 00:00:00 /usr/sbin/lighttpd -D -f /etc/lighttpd/lighttpd.conf
Steven
And. again, the file just isn't there. Here's what happened when I executed the suggested commands:
No idea. So something must go wrong there.
Please use testing Clonezilla live, e.g., 3.1.2-9 or 20240116-noble a try:
https://clonezilla.org/downloads.php
Maybe the results will be different.
Steven
So I did, the latest Debian version, and ... no different. Same exact errors and results as described above. It is definitely not a network connection issue or open port issue. (If port 80 on the server were closed or blocked, I wouldn't be able to boot the client at all, right?) I made sure to even test with an unmanaged switch.
I get the sense that the server is not generating the file the client needs. Would you want me to video-document exactly how I boot up both server and client, or else, to provide diagnostic data of some kind on the server?
Last edit: T4T Tech Support 2024-01-29
That's really weird. We just can not reproduce this issue here.
"I get the sense that the server is not generating the file the client needs. Would you want me to video-document exactly how I boot up both server and client, or else, to provide diagnostic data of some kind on the server?" -> Yes, maybe that helps.
Steven
P.S. Getting the "interactive client" function to work would aid us tremendously in speeding up deployment of refurbished laptops to our clients: we don't always have many of the make, same drive size, and same drive type (SATA or NVMe). Thank you.
Last edit: T4T Tech Support 2024-02-02
OK, using the" interactive client" as a workaround can be a good solution when you have just some clients to be deployed.
Steven
To clarify, this whole issue I'm raising is precisely the failure of the "interactive client" feature to function. Where should I send the video files that I proposed to make?
And should I send a file with technical data of the server and client PC's I'm using in the experiment? If so, what Linux program should I use to generate this output?
You can put the video on cloud drive, like google drive.
Then email me the details to "steven@clonezilla.org".
Steven
Did so yesterday. Now, the video shows me using an existing DHCP server in that network. However, cutting that off and making the Clonezilla lite-server do DHCP, I get the same result.
In "massive deployment" mode, everything works fine. But we really have a need for the flexibility the "interactive client" mode promises.
Oh, after watching the video, I knew you misunderstood the "interactive mode". Sorry for confusing you.
The "interactive mode" is just for you to do something like you boot Clonezilla live from USB flash drive in that machine. The difference is you boot that machine via network.
Hence you have to do something like:
https://clonezilla.org//fine-print-live-doc.php?path=clonezilla-live/doc/01_Save_disk_image
and
https://clonezilla.org//fine-print-live-doc.php?path=clonezilla-live/doc/02_Restore_disk_image
You should not choose "lite-client" in that machine.
Steven
So here's what I was hoping to do: have a Clonezilla server that has multiple images stored, and then each client be able to choose which of those images it should fetch. Ideally, if there's more than one hard drive in the client, that image would be cloned onto all of them.
I had the idea that "interactive client" with the lite-server would allow this, but apparently not. Is this actually feasible with any form of Clonezilla networking? The DRBL server, perhaps?
Sure, it's can be done.
You can edit the netboot file on the server:
/tftpboot/nbi_img/grub/grub.cfg (for uEFI netboot)
/tftpboot/nbi_img/pxelinux.cfg/default (for legacy BIOS netboot)
by referring to these docs:
https://clonezilla.org//fine-print-live-doc.php?path=clonezilla-live/doc/05_Preseed_options_to_do_job_after_booting
https://clonezilla.org//fine-print-live-doc.php?path=./clonezilla-live/doc/99_Misc/00_live-boot-parameters.doc#00_live-boot-parameters.doc
https://clonezilla.org//advanced/reserved-word-ocs-sr.php
For example, you can assign the ocs_live_run parameters using the ocs-sr command with "ask_user" for the image name and destination disk:
ocs_live_run="ocs-sr -g auto -e1 auto -e2 -c -r -j2 -k -p true restoredisk ask_user ask_user"
Steven
If I'm reading correctly: the assignment of the variable ocs_live-run needs to be in this file on the running server: /tftpboot/nbi_img/grub/grub.cfg (since it is supposed to be a UEFI netboot). But I see nowhere in that file to place such a variable (it looks like it just governs the boot menu), and the "preseed" document that you linked to makes reference to other configuration files, wherein that variable and others like it are set. So I'm confused on that point.
As to the server: can it be the "lite-server", or does it have to be the DRBL?
If you choose "interactive-live" in this step:
https://clonezilla.org//clonezilla-live/doc/11_lite_server/images/ocs-08-5-massive-deployment.png
You will see the following in the file "/tftpboot/nbi_img/grub/grub.cfg" on your server. It contains something like this:
The ocs_live_run is that I mentioned. Just refer to my previous post and you should be able to customize that.
Steven
I'm finally able to get back to this....
"Interactive-live": did you mean "interactive-client"? Because that's what the picture shows....
If I understand correctly, I need to substitute the value of the variable "ocs_live_run" from its default [ocs-live-general] to what you mentioned on Feb. 17th [ocs-sr -g auto -e1 auto -e2 -c -r -j2 -k -p true restoredisk ask_user ask_user - or something like it] ? And does the replacement value need to be in quote marks?
Last edit: T4T Tech Support 2024-03-12
[Couldn't find a way to delete this comment]
Last edit: T4T Tech Support 2024-05-14
[Couldn't find a way to delete this comment]
Last edit: T4T Tech Support 2024-05-14
So at last I could get back to this and try to implement it. The test server has one hard drive attached with two partitions: sda1 is the Clonezilla boot drive; sda2 has several already-created images. So on the server, I mount the root of sda2 to /home/partimag.
I put the server into interactive-client mode and then immediately exit to the command line. There I replace the generated /tftpboot/nbi_img/grub/grub.cfg with a version with the modifications to ocs_live_run as follows:
menuentry "Clonezilla PXE" --id clonezilla-live-client {
echo "Enter Clonezilla Live 3.1.2-22-amd64 (Ramdisk)..."
echo 'Loading Clonezilla-live-vmlinuz...'
$linux_cmd Clonezilla-live-vmlinuz initrd=Clonezilla-live-initrd.img boot=live union=overlay username=user config components quiet loglevel=0 noswap edd=on nomodeset enforcing=0 locales=en_US.UTF-8 keyboard-layouts=us ocs_live_run="ocs-sr -r -e1 auto -k0 -scr -e2 -j2 -p choose restoredisk ask_user ask_user" ocs_live_extra_param= ocs_live_batch=no net.ifnames=0 nosplash noeject fetch=http://192.168.169.250/Clonezilla-live-filesystem.squashfs ocs_server="192.168.169.250" vga=788 hostname=lite-client ocs_repository="http://192.168.169.250/home/partimag/"
echo 'Loading initial ramdisk Clonezilla-live-initrd.img...'
$initrd_cmd Clonezilla-live-initrd.img
}
... and then exit the command line and reinitialize the server in interactive-client mode.
I then boot a test client, it gets through the initial part of the boot, then, as pictured below, it eventually gets to an error that "/home/partimag" is not a mounting point .. and I've not yet been able to figure out why that's happening.
Also: is there any way to bypass the need for the client to authenticate to the server?
Last edit: T4T Tech Support 2024-05-14