Proxmox VE 8.0 released!

LiloBzH said:
Hi !

i have connection probleme since upgrade proxmox 7.4 > 8.0

Code: 
2023-06-24T17:23:07.249280+02:00 proxmox kernel: [ 3309.419261] r8169 0000:02:00.0 enp2s0: rtl_ephyar_cond == 1 (loop: 100, delay: 10).
2023-06-24T17:23:07.249285+02:00 proxmox kernel: [ 3309.420982] r8169 0000:02:00.0 enp2s0: rtl_ephyar_cond == 1 (loop: 100, delay: 10).
2023-06-24T17:23:07.253285+02:00 proxmox kernel: [ 3309.422741] r8169 0000:02:00.0 enp2s0: rtl_ephyar_cond == 1 (loop: 100, delay: 10).
2023-06-24T17:23:07.281274+02:00 proxmox kernel: [ 3309.452104] r8169 0000:02:00.0 enp2s0: rtl_eriar_cond == 1 (loop: 100, delay: 100).
2023-06-24T17:23:07.313268+02:00 proxmox kernel: [ 3309.482136] r8169 0000:02:00.0 enp2s0: rtl_eriar_cond == 1 (loop: 100, delay: 100).
2023-06-24T17:23:07.341302+02:00 proxmox kernel: [ 3309.512139] r8169 0000:02:00.0 enp2s0: rtl_eriar_cond == 1 (loop: 100, delay: 100).

no crash. Just no connection ! if i reboot, it's ok but connection is lost after some times
Click to expand...

I have exactly the same issue, but i didn’t found a solution so far.
Tonight I will try a clean install with V8 and look if the problem is solved then, if not I will reinstall 7.4.
 
Rijswijker said:
I have exactly the same issue, but i didn’t found a solution so far.
Tonight I will try a clean install with V8 and look if the problem is solved then, if not I will reinstall 7.4.
Click to expand...
other users have same problem with this card, seem that last driver installed with dkms is fixing it.
So you need to wait that proxmox team release a new kernel with the fixed drivers.
 
spirit said:
other users have same problem with this card, seem that last driver installed with dkms is fixing it.
So you need to wait that proxmox team release a new kernel with the fixed drivers.
Click to expand...

Optiplex 3070 Micro (RT8168) has the same issue, network works for a short time, then all connections are gone

Going to activate the 8168-DKMS to see if that solves the issue tonight
 
Last edited:
jdancer said:
My next step was to re-create the monitors manually by disabling the service and removing /var/lib/ceph/mon/<hostname> directory.

Then ran 'pveceph mon create'. After awhile it timed-out. Running 'journalctl on the failed monitor service shows the following:

Jun 25 13:29:03 pve-test-7-to-8 systemd[1]: Started ceph-mon@pve-test-7-to-8.service - Ceph cluster monitor daemon.
Jun 25 13:29:04 pve-test-7-to-8 ceph-mon[8161]: *** Caught signal (Illegal instruction) **
Jun 25 13:29:04 pve-test-7-to-8 ceph-mon[8161]: in thread 7fe8c0b1da00 thread_name:ceph-mon
Jun 25 13:29:04 pve-test-7-to-8 ceph-mon[8161]: ceph version 17.2.6 (810db68029296377607028a6c6da1ec06f5a2b27) quincy (stable)
Jun 25 13:29:04 pve-test-7-to-8 ceph-mon[8161]: 1: /lib/x86_64-linux-gnu/libc.so.6(+0x3bf90) [0x7fe8c11bdf90]
...
Jun 25 13:29:55 pve-test-7-to-8 ceph-mon[9402]: NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.
Jun 25 13:29:55 pve-test-7-to-8 systemd[1]: ceph-mon@pve-test-7-to-8.service: Main process exited, code=killed, status=4/ILL
Jun 25 13:29:55 pve-test-7-to-8 systemd[1]: ceph-mon@pve-test-7-to-8.service: Failed with result 'signal'.
Jun 25 13:30:05 pve-test-7-to-8 systemd[1]: ceph-mon@pve-test-7-to-8.service: Scheduled restart job, restart counter is at 6.
Jun 25 13:30:05 pve-test-7-to-8 systemd[1]: Stopped ceph-mon@pve-test-7-to-8.service - Ceph cluster monitor daemon.
Jun 25 13:30:05 pve-test-7-to-8 systemd[1]: ceph-mon@pve-test-7-to-8.service: Start request repeated too quickly.
Jun 25 13:30:05 pve-test-7-to-8 systemd[1]: ceph-mon@pve-test-7-to-8.service: Failed with result 'signal'.
Jun 25 13:30:05 pve-test-7-to-8 systemd[1]: Failed to start ceph-mon@pve-test-7-to-8.service - Ceph cluster monitor daemon.

Seems to point to a corrupt binary, compile, or something else. No idea.

Going to do a clean install of Proxmox 8 and see if I get the same error when manually creating the monitors.

I'm testing the migration script on a test cluster which does have quorum and nodes can ping each other.

This is why you always test before pushing it to production.
Click to expand...

Did a clean install of Proxmox 8 and using the no-sub Quincy repository.

Still got the 'Caught signal (illegal instruction)' message.

It's pointing a bad re-compile or the Ceph monitor binaries no longer are supported on AMD Opteron 2427 CPUs.

Ceph was working fine under Proxmox 7.
 
Last edited:
  • Like
Reactions: kifeo
elcoco said:
Optiplex 3070 Micro (RT8168) has the same issue, network works for a short time, then all connections are gone

Going to activate the 8168-DKMS to see if that solves the issue tonight
Click to expand...
8168-DKMS Driver works like a charm, rock-stable again
 
jdancer said:
It's pointing a bad re-compile or the Ceph monitor binaries no longer are supported on AMD Opteron 2427 CPUs.
Click to expand...
The quincy release is quite popular, and we have no other reports of this, same w.r.t. Proxmox VE 8 (compared to the amount of systems reporting to our package repo infra).

OTOH, your CPU is over 14 years old and quite probably neither compiler projects, library developers nor Ceph or us test on that old HW, so glitches in there can go unnoticed. I'd recommend installing the newest BIOS/Firmware and CPU microcode available for that platform, with a bit of luck this makes the CPU compatible enough.
 
mjaxon said:
Any insight on the procedures to do this? I'm facing the same problem and it's unclear in my research what to do.
Click to expand...
Sure, no problem

  • Proxmox Shell
  • sudo nano /etc/apt/sources.list
  • Add non-free at the end of the first bookworm main contrib line
  • Control + X, save exit
  • sudo apt update
  • sudo apt install dkms
  • sudo apt install r8168-dkms
  • reboot
  • ethtool -i %interfacename% (e.g. ethtool -i enp1s0) to check the loaded driver, should show r8168
 
  • Like
Reactions: nonomega, pschneider1968, alsicorp and 1 other person
For me, it was a lot simpler to just migrate vms to another host, upgrade one host at a time moving vms back once done. Simple, clean, no old files left behind.
 
Wrong guest memory report by balloon with fresh PVE 8.x and fresh Windows 2019 as guest

- Fresh PVE 8
- Fresh Windows Server 2019 as guest
- No software but latest Virtio drivers (0.1.229) + balloon service + qemu Guest Agent

1687934846137.png
 
Hi,
vaschthestampede said:
I point out that using proxmox 8 I can't mount SMB/CIFS shares.
Click to expand...
yes, unfortunately the smbclient changed how it handles an empty domain parameter in Bookworm, so you'll run into an issue when trying to add a share with empty domain in the UI.

As a workaround, use the CLI or in the UI, add the storage without the enabled checkbox and then edit and enable it afterwards.

It will be fixed with future versions of libpve-storage-perl >= 8.0.2: https://git.proxmox.com/?p=pve-storage.git;a=commit;h=b27da68320ba004e9a0f490707b86ef01af2c669
 
  • Like
Reactions: vaschthestampede, pschneider1968 and DerDanilo

Current Release Family: NVIDIA vGPU Software 15​

vGPU SoftwareLinux vGPU ManagerWindows vGPU ManagerLinux DriverWindows DriverRelease Date
15.3525.125.03529.06525.125.06529.11June 2023

NVIDIA-Linux-x86_64-525.125.03-vgpu-kvm.run
Has anyone tried installing this driver on PVE 8 ?

Everything works fine on 7.4-15
 
Last edited:
You must log in or register to reply here.
Top Bottom