Does not boot with new kernels after upgrade from 1.1

[mmenaz]

Hi, I've upgraded from 1.1 to 1.2 following the online instructions. I suppose the upgrade brought me to 1.3 directly, since at reboot the kernel list in grub menu was:
2.6.24-7-pve
2.6.24-5-pve
2.6.24-2-pve (old one)
Unfortunately, with most recent two kernels boot starts briefly and then "freezes" with the screen printing something like this:
------
Initrd /initrd.img-2.6.24-7-pve
[linux-initrd @0x372d.....]

at the bottom (jumping from lines above)
Kernel alive
Kernel direct mapping tables up to 230000000 @ 8000-12000
------
I've tried a lot of bios settings, also I updated to latest bios, but nothing solved.
Now Proxmox only boots with "old", 1.1 kernel (2.6.24-2-pve).
Any clue?
MB Asus M3A78, AMI ACPI BIOS rev. 0703 (now upgraded to 1606)
CPU AMD Athlon 64 X2 Dual Core Processor 4200+
Memory 8GB

I had also a problem with a Asus M3A78-EM, Phenom II X4 920 Processor, 8GB RAM, where I just tried 1.2 and 1.3, that refused to boot (never tried 1.1), but don't remember the last message printed. I just had to change MB with the one of my PC that worked for sure (asus M3A, processor AMD Phenom9550, 8GB RAM) .

Wondering if I'd better leave Asus (I will anyway, since they are so against GNU/Linux) or there is some bios /kernel setup I can try.
If the former, which MB manufacturer do you suggest? Better AMD or Intel (for compatibility).

Last question... can I keep running this "mixed" proxmox installation?
proxmox:~# uname -r
2.6.24-2-pve
proxmox:~# pveversion -v
pve-manager: 1.3-1 (pve-manager/1.3/4023)
qemu-server: 1.0-14
pve-kernel: 2.6.24-8
pve-kvm: 86-3
pve-firmware: 1
vncterm: 0.9-2
vzctl: 3.0.23-1pve3
vzdump: 1.1-2
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1

And kvm module loaded is kvm-83

Thanks a lot

 

[tom]

Hi, I've upgraded from 1.1 to 1.2 following the online instructions. I suppose the upgrade brought me to 1.3 directly, since at reboot the kernel list in grub menu was:
2.6.24-7-pve
2.6.24-5-pve
2.6.24-2-pve (old one)
Unfortunately, with most recent two kernels boot starts briefly and then "freezes" with the screen printing something like this:
------
Initrd /initrd.img-2.6.24-7-pve
[linux-initrd @0x372d.....]

at the bottom (jumping from lines above)
Kernel alive
Kernel direct mapping tables up to 230000000 @ 8000-12000
------
I've tried a lot of bios settings, also I updated to latest bios, but nothing solved.
Now Proxmox only boots with "old", 1.1 kernel (2.6.24-2-pve).
Any clue?
MB Asus M3A78, AMI ACPI BIOS rev. 0703 (now upgraded to 1606)
CPU AMD Athlon 64 X2 Dual Core Processor 4200+
Memory 8GB

I had also a problem with a Asus M3A78-EM, Phenom II X4 920 Processor, 8GB RAM, where I just tried 1.2 and 1.3, that refused to boot (never tried 1.1), but don't remember the last message printed. I just had to change MB with the one of my PC that worked for sure (asus M3A, processor AMD Phenom9550, 8GB RAM) .

Wondering if I'd better leave Asus (I will anyway, since they are so against GNU/Linux) or there is some bios /kernel setup I can try.
If the former, which MB manufacturer do you suggest? Better AMD or Intel (for compatibility).

Last question... can I keep running this "mixed" proxmox installation?
proxmox:~# uname -r
2.6.24-2-pve
proxmox:~# pveversion -v
pve-manager: 1.3-1 (pve-manager/1.3/4023)
qemu-server: 1.0-14
pve-kernel: 2.6.24-8
pve-kvm: 86-3
pve-firmware: 1
vncterm: 0.9-2
vzctl: 3.0.23-1pve3
vzdump: 1.1-2
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1

And kvm module loaded is kvm-83

Thanks a lot

We heard several times about issues with similar hardware.

Currently we have no Opterons/Phenoms in our test lab so we cannot test in detail.

Normal approach to a working solution:
Hardware sellers or customers send their hardware to us and we validate/test. Sometimes its enough for a few weeks, for some parts it makes sense to integrate it the test cycles. so feel free to send us your server for testing or think about sponsoring the needed parts for our test labs.

 

[mmenaz]

We heard several times about issues with similar hardware.

Currently we have no Opterons/Phenoms in our test lab so we cannot test in detail.

My "main" system is a AMD Athlon 64 X2, so probably is related with general AMD cpu or chipset.

Normal approach to a working solution:
Hardware sellers or customers send their hardware to us and we validate/test. Sometimes its enough for a few weeks, for some parts it makes sense to integrate it the test cycles. so feel free to send us your server for testing or think about sponsoring the needed parts for our test labs.

I'm just a lonely developer that has provided a cheap PC to the customer of my old legacy program. Delivery cost to you and back could probably overcome the cost of the whole hardware itself (Italy).
As soon as I receive some money from him, I've planned to donate some to you and other FOSS technology I'm using, but will not be enough for your needs.
What frightenes me is that now I hardly dare to suggest using proxmox if there is such an high risk that it will not work with the hardware my next client (hope to have one) could buy. If you tell me that Intel is a safer choice, I will suggest Intel next time.
Hope you can find a discontinued AMD X2 or buy a very cheap one with the few euros I will be able to donate, in any case. I think that testing also on AMD hardware is a requirement for Proxmox in the future.

 

[tom]

Just to make clear - using cheap hardware is a general no go for a virtualization project. we will not try to support outdated and not suited desktop hardware.

I would recommend to start with at least Intel Xeon (x3xxx) processors on a entry level server motherboard, e.g. Intel S3200SH or similar - also AMD on a similar level will work.

Best: Intel Xeon 55xx series or the latest AMD´s.

 

[mmenaz]

Well, my customer wanted to buy HP hardware, but I discouraged him because is Proxmox would have not be able to run, I would have to refund him a lot of money. Suggesting assembled hardware, you can still choose high quality components but replace "offending" ones with low cost (i.e. MB around 100 euros).
Having had a very risky experience with a Fujitsu Xeon (don't remember the model) that had kvm refusing to start and proxmox 1.1, that was fixed, fortunatly the day of the installation, with 1.2, I'm a bit puzzled. I don't dare to upgrade him to 1.3, of course.
Some hardware works only with old version, some other with the last one is scaring, isn't it?
Is not a blame to you, of course, just to tell you about my situation that, I think, is the one of others also, and look together what can be done about.
I.e. I could collaborate, if you have "beta versions" with debug or other troubleshooting options enabled, to help you in testing (I'm not a guru, need detailed instructions about how to help and dump useful info if booting freezes).

 

[tom]

Well, my customer wanted to buy HP hardware, but I discouraged him because is Proxmox would have not be able to run, I would have to refund him a lot of money. Suggesting assembled hardware, you can still choose high quality components but replace "offending" ones with low cost (i.e. MB around 100 euros).
Having had a very risky experience with a Fujitsu Xeon (don't remember the model) that had kvm refusing to start and proxmox 1.1, that was fixed, fortunatly the day of the installation, with 1.2, I'm a bit puzzled. I don't dare to upgrade him to 1.3, of course.
Some hardware works only with old version, some other with the last one is scaring, isn't it?
Is not a blame to you, of course, just to tell you about my situation that, I think, is the one of others also, and look together what can be done about.
I.e. I could collaborate, if you have "beta versions" with debug or other troubleshooting options enabled, to help you in testing (I'm not a guru, need detailed instructions about how to help and dump useful info if booting freezes).

why don´t you buy just a server which is known to work? then there is no risk.