PCIEPORT AER: Correct Error Received

[Kotuku52]

Hi There,

Currently having an issue with a fresh install of Proxmox.
System: Acer Veriton X2680G - Intel Core i5 11400 / 64GB / Latest BIOS
SSD's: 2x500GB Samsung Pro 980 (1x on a SilverStone ECM21-E) - setup on ZFS
(side note issue - the Silverstone SSD sometimes drops off, what can I do to diagnose this?)

Now I actually have 2x of these setups and both are displaying the same thing.

The log is getting spammed with Correct Error Received messages. I have tried to do pcie_aspm=off etc on the grub file.
I also found this thread with a similar issue but didn't work: https://forum.proxmox.com/threads/w...e-error-detected-samsung-980-m-2-nvme.108162/
(Bit of a dumb question, the system is booting up with systemD - do these grub tweaks actually work in this instance?)



Thanks for any help.

 

[leesteken]

(Bit of a dumb question, the system is booting up with systemD - do these grub tweaks actually work in this instance?)

If your system is booting with systemd-boot then you need to put kernel parameters in /etc/kernel/cmdline (single line, no support for comments). You can always check which parameters are currently passed with cat /proc/cmdline.

EDIT: You could also try pci=noaer as a work-around, but you may actually have PCIe issues and I don't know how to diagnose and fix that.

 

[Kotuku52]

If your system is booting with systemd-boot then you need to put kernel parameters in /etc/kernel/cmdline (single line, no support for comments). You can always check which parameters are currently passed with cat /proc/cmdline.

Thanks for this - thought I was going mad lol.
This is what is there by default for /etc/kernel/cmdline
root=ZFS=rpool/ROOT/pve-1 boot=zfs

For reference am I adding pcie_aspm=off (and any other boot arguments) after the boot=zfs or a new line under it?

Thanks.

 

[leesteken]

For reference am I adding pcie_aspm=off (and any other boot arguments) after the boot=zfs or a new line under it?

I may have been a bit unclear: only one single line is supported. Put everything after one another separated by space(s) on a single line. Do not put a # in front of the first (or any other) line and put stuff on the second line. Really just use one single line, which is the first and only line in the file.

 

[Kotuku52]

I may have been a bit unclear: only one single line is supported. Put everything after one another separated by space(s) on a single line. Do not put a # in front of the first (or any other) line and put stuff on the second line. Really just use one single line, which is the first and only line in the file.

That makes sense.

So I put in the Boot Arguments - single line - but both times the cat /proc/cmdline doesn't show the arguments? - is this normal?
I'm using proxmox-boot-tool-refresh each time after changing the /etc/kernel/cmdline - this the right procedure?

Thanks so much so far.

 

[leesteken]

So I put in the Boot Arguments - single line - but both times the cat /proc/cmdline doesn't show the arguments? - is this normal?
I'm using proxmox-boot-tool-refresh each time after changing the /etc/kernel/cmdline - this the right procedure?

If you are using systemd-boot (UEFI & ZFS rpool) then proxmox-boot-tool refresh should be enough. It is strange that cat /proc/cmdline does not show your changes after a reboot.
Can you please attach the /etc/kernel/cmdline file and show the output of proxmox-boot-tool refresh and cat /proc/cmdline (in CODE or ICODE tags instead of screenshots)?

 

[Kotuku52]

initrd=\EFI\proxmox\5.13.19-2-pve\initrd.img-5.13.19-2-pve root=ZFS=rpool/ROOT/pve-1 boot=zfs

This was the output of the refresh tool:

Running hook script 'proxmox-auto-removal'..
Running hook script 'zz-proxmox-boot'..
Re-executing '/etc/kernel/postinst.d/zz-proxmox-boot' in new private mount namespace..

But this appeared in the shell but did not show up in the output:
Copying and configuring kernels on /dev/disk/by-uuid/1F73-152A
Copying kernel and creating boot-entry for 5.13.19-2-pve
WARN: /dev/disk/by-uuid/1F73-4CC1 does not exist - clean '/etc/kernel/proxmox-boot-uuids'! - skipping

Does this help? I had to add an extension to cmdline to get it to upload

Thanks

 

[leesteken]

It's not clear to me which is the /etc/kernel/cmdline and which is the cat /proc/cmdline in your message as the Code and the attachment are identical.

Copying and configuring kernels on /dev/disk/by-uuid/1F73-152A is not followed by any actual lines that copy the kernel(s), so I'm not sure if that (existing) ESP partition is actually refreshed.
WARN: /dev/disk/by-uuid/1F73-4CC1 does not exist - clean '/etc/kernel/proxmox-boot-uuids'! - skipping makes me think that you have a second ESP partition but it is not configured correctly (or changed later) and therefore not updated. But you appear to boot from an ESP partition which does not have UUID 1F73-152A and also not UUID 1F73-4CC1 and which is therefore not refreshed with your changes to /etc/kernel/cmdline.

Please have a look at the documentation on how to configure proxmox-boot-tool. What is the output of ls -l /dev/disk/by-uuid/?