Is there a way I can install hiveos-0.6-222-stable instead of ubuntu20? The AMD driver doesn’t play well with the BC-250.
Which one for example?
I’m having trouble with the BC-250. It works with the following but it will not init the GPU on the version used in diskless PXE:
5.15.0-hiveos #110
Kernel Version
A22.20.5 (5.18.2301)
BC-250 × 1
Yes the BC-250 has issues (amdgpu failed to read reg:ec48) I’ve tried modify the deploy_pxe file to force install other amd gpu versions but none work and all give the same error
Is there a repo with the older diskless kernels so that I can have it pull down one of those versions? For instance, they work fine with the Hive on Ubuntu 18. If the diskless kernel for that one was available I could just change the repo url and have it install that one instead.
1 Like
Anyone want to get BC-250 working ive made a modified version.
TheJames is it possible to make a pxe boot version of the beta image hiveos? there are as few coins that require it but it impossible to do a hive replace beta on pxe, thanks.
I’ve tried hacking around but can’t quite figure out how to send the hostname to the dhcp server… I suspect that i can’t because we get the ip address from the initial boot image…
However i was under the impression that i could trigger a dhcp renew once booted and send the hostname that way… except it looks like hive has everything setup as static ip…
lts ubutnu 20:
pxe doesn’t boot error build kerneals first.
go into defuatl fodler path
per github readme:
run build command
errors:
issue post made on github:
may be time for a refresh…
Due to hive lack of a response, I hope they fix their s^l*t ! Expected better…
in the mean time here is a system that should help clients build and use PXE diskless:
Just tried to revive my old Hive OS PXE server and tried to update it.
The update, on the PXE server itself, was successful.
However, when I try to boot my client off of it, it is looking for http://<<IP_address>>/boot/ubuntu20.img which doesn’t exist in /path/to/pxe/server/boot as I only have an EFI folder in there, and inside that, there is an ubuntu folder, not ubuntu20.
The the HiveOS PXE server is great…when it works.
(It used to work fantastically for me and I had always recommended this to people rather than having to deal with flashing HiveOS onto a USB stick.)
edit
I did revive my old HiveOS PXE boot server rather than trying to deploy/upgrade to the new one, and that one worked.
I was able to update the client software, but my first attempt to update the HiveOS PXE server software produced the same issue/error, but when I ran it a second time (destroyed the LXC container and cloned it again from my previous “backup”), I choose “no” when it asked me if I wanted to update the HiveOS PXE Server package, but I choose “yes” to update the HiveOS (client) OS. That worked.
I didn’t try @bmmbmm01’s solution, primarily due to laziness (my old Ubuntu based HiveOS PXE server worked, so it was easier for me to just deploy that as a LXC via Proxmox than it was for me to try and set up a new Debian LXC, to run @bmmbmm01’s scripts).
So it’s back up and running now.
If people want a copy of my old HiveOS PXE server, please feel free to either reply to this and tag me in it or PM me. I think that I should still have a sanitised version of it floating around somewhere.