Hi community,
TL;DR: Running apt dist-upgrade and rebooting Proxmox VE on Dell Optiplex kills the mainboard.
Usually I would be really careful with such a statement. But I find this case so curious I just want to share my experience here, also because I cannot find a single clue of anybody on the net experiencing something similar. Not having an explanation of what is happening to the system drives me a bit crazy ;-)
I'm new to Proxmox VE and wanted to get a playground setup up and running on a rather old Dell Optiplex 7010 SFF (3rd gen Core i5, 32GiB RAM, 500GB Samsung 840 EVO SSD). To be honest: I'm not 100% sure if I had the first installation on the original 320GiB HDD instead.
Major BIOS settings: no legacy boot, just UEFI, PXE, SecureBoot disabled, TPM disabled
Dell Diagnostics (ePSA) runns without issues.
Booted Proxmox VE 8.4 ISO (9.0 just not yet available) via netboot.xyz and ran the graphical installation without any issues, using ext4 filesystem, no advanced config. Rebooted maybe a few times without any issues (playing around with Intel AMT/KVM, but no changes to PVE installation).
I disabled the enterprise and ceph repos and added the no-subscription repo and ran the dist-upgrade:
Now the fun part...Reboot shows no issues on shutdown, but after that I only get a black screen (not even POST :-o). I then fiddled around with clearing CMOS settings via jumper etc., but wasn't able to resurrect the PC.
After a bit of browsing my favorite Internet search engine I found out that it is rather common to have a defective BIOS chip on the Optiplex 7010... and indeed, after desoldering the BIOS chips and comparing their contents to the original BIOS image (version A29) it turned out that one of the chips contained just garbage. I flashed the original contents via a SPI flash programmer, verification runs ok, soldered the BIOS chips back on the mainboard. Outcome: PC is constantly turning on and off every few seconds.
Small detour: I'm a software engineer and electrical engineer and because I was really curious I investigated a bit deeper into the Vcore circuitry on the mainboard and was able to verify with rather good confidence (thanks to schematics and datasheets being available) that the Vcore power controller goes into a short circuit protection, resulting in the shutdown of the PSU subsequently power cycling the board.
So far I was of course disappointed, but as the hardware was rather old and maybe more than 2nd hand, I just wrote it off. I then bought another refurbished 7010 with lower specs but just used the mainboard from this system in my first setup. Installed Proxmox VE 8.4 without any special care about what exactly is happening in every step, because at this point I still had the assumption that the first HW just died by coincidence). And now on the second HW, after the obligatory apt dist-upgrade I had EXACTLY the same observations as with the first install: garbage in BIOS chip, short circuit on Vcore, etc.
For curiosity I invested another few bucks into a new mainboard and used the CPU and RAM from the second Optiplex now (Core i3, 4GiB RAM). So every component was now different from the other setups (also the PSU) - with the exception of the SSD.
Yeah and what should I say ... same bricking procedure as usual
This time I pulled some logs from the SSD (see attachments). I hope that anybody here can tell me what's going on. A solution would be a nice to have, but if there is no way out, I'll just give it another try with some different HW
Best regards,
Reinhard
TL;DR: Running apt dist-upgrade and rebooting Proxmox VE on Dell Optiplex kills the mainboard.
Usually I would be really careful with such a statement. But I find this case so curious I just want to share my experience here, also because I cannot find a single clue of anybody on the net experiencing something similar. Not having an explanation of what is happening to the system drives me a bit crazy ;-)
I'm new to Proxmox VE and wanted to get a playground setup up and running on a rather old Dell Optiplex 7010 SFF (3rd gen Core i5, 32GiB RAM, 500GB Samsung 840 EVO SSD). To be honest: I'm not 100% sure if I had the first installation on the original 320GiB HDD instead.
Major BIOS settings: no legacy boot, just UEFI, PXE, SecureBoot disabled, TPM disabled
Dell Diagnostics (ePSA) runns without issues.
Booted Proxmox VE 8.4 ISO (9.0 just not yet available) via netboot.xyz and ran the graphical installation without any issues, using ext4 filesystem, no advanced config. Rebooted maybe a few times without any issues (playing around with Intel AMT/KVM, but no changes to PVE installation).
I disabled the enterprise and ceph repos and added the no-subscription repo and ran the dist-upgrade:
code_language.shell:
# cat /etc/apt/sources.list.d/pve-no-subscription.list
deb http://download.proxmox.com/debian/pve bookworm pve-no-subscription
# apt update && apt dist-upgrade && apt autoremoveNow the fun part...Reboot shows no issues on shutdown, but after that I only get a black screen (not even POST :-o). I then fiddled around with clearing CMOS settings via jumper etc., but wasn't able to resurrect the PC.
After a bit of browsing my favorite Internet search engine I found out that it is rather common to have a defective BIOS chip on the Optiplex 7010... and indeed, after desoldering the BIOS chips and comparing their contents to the original BIOS image (version A29) it turned out that one of the chips contained just garbage. I flashed the original contents via a SPI flash programmer, verification runs ok, soldered the BIOS chips back on the mainboard. Outcome: PC is constantly turning on and off every few seconds.
Small detour: I'm a software engineer and electrical engineer and because I was really curious I investigated a bit deeper into the Vcore circuitry on the mainboard and was able to verify with rather good confidence (thanks to schematics and datasheets being available) that the Vcore power controller goes into a short circuit protection, resulting in the shutdown of the PSU subsequently power cycling the board.
So far I was of course disappointed, but as the hardware was rather old and maybe more than 2nd hand, I just wrote it off. I then bought another refurbished 7010 with lower specs but just used the mainboard from this system in my first setup. Installed Proxmox VE 8.4 without any special care about what exactly is happening in every step, because at this point I still had the assumption that the first HW just died by coincidence). And now on the second HW, after the obligatory apt dist-upgrade I had EXACTLY the same observations as with the first install: garbage in BIOS chip, short circuit on Vcore, etc.
For curiosity I invested another few bucks into a new mainboard and used the CPU and RAM from the second Optiplex now (Core i3, 4GiB RAM). So every component was now different from the other setups (also the PSU) - with the exception of the SSD.
Yeah and what should I say ... same bricking procedure as usual
This time I pulled some logs from the SSD (see attachments). I hope that anybody here can tell me what's going on. A solution would be a nice to have, but if there is no way out, I'll just give it another try with some different HW
Best regards,
Reinhard