Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr
-
@dafyre said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
@black3dynamite said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
I wonder if its the CPU model that you are assign to.
This would be my guess. It looks to be related to the Intel i10nm Ice Lake(?) CPUs.
Yeah, the few Google results I found mentioned a CPU, although I thought I saw skylake.
-
Vultr replied -
"Hello,
The software one your VPS is outside of our scope of support.
Thank you."
I replied -
"I understand. Just thought your guys might want to figure out why your instances can't handle fully-patched versions of Fedora." -
@wrx7m said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
@dafyre said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
@black3dynamite said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
I wonder if its the CPU model that you are assign to.
This would be my guess. It looks to be related to the Intel i10nm Ice Lake(?) CPUs.
Yeah, the few Google results I found mentioned a CPU, although I thought I saw skylake.
What Vultr "level" is the VM?
-
@wirestyle22 said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
@wrx7m said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
@wirestyle22 said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
@wrx7m I would try to create a new instance with their image of 29 and upgrade to 30 to see if it works. If that doesn't work then I'd try to re-create what you are running on a custom iso and see if you can upgrade that successfully.
I can try the upgrade of a stock instances to see what happens. If not, I can just start at 30 and install and configure it as needed. Then snapshot it and make the necessary customization. Just like I did to get these 4 servers.
In the future I would definitely avoid using their images. I have run into some insanity and I wasn't sure if it was a mistake I made or the image. Ended up being the image.
What insanity was that? If their image is bad, they have support for that. If you don't use it, you are left on your own. I would not advise avoiding their images.
-
@wrx7m said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
I replied -
"I understand. Just thought your guys might want to figure out why your instances can't handle fully-patched versions of Fedora."This isn't quite stated properly. The issue is that Fedora cannot properly handle the Vultr instance, not the other way around. Vultr is handling Fedora 30 just fine.
It is Fedora to whom this should be reported, it is in their court to fix, not Vultr's.
-
@wrx7m My guess is that either there is a configuration difference that would be easiest to ask Vultr to provide, or there is a package missing that Fedora does not detect as needed and install properly that we should be able to find by doing a package listing and comparing. I have many working Fedora 30 instances (upgraded from 29) on Vultr, so we have baseline systems that we know work and made the transition to compare against. It might take a bit of searching, but likely we can track down what Fedora is leaving out.
-
As a starting point, are these installed when you install via ISO?
# dnf list | grep edac edac-utils.i686 0.16-20.fc30 fedora edac-utils.x86_64 0.16-20.fc30 fedora edac-utils-devel.i686 0.16-20.fc30 fedora edac-utils-devel.x86_64 0.16-20.fc30 fedora -
@scottalanmiller said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
@wirestyle22 said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
@wrx7m said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
@wirestyle22 said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
@wrx7m I would try to create a new instance with their image of 29 and upgrade to 30 to see if it works. If that doesn't work then I'd try to re-create what you are running on a custom iso and see if you can upgrade that successfully.
I can try the upgrade of a stock instances to see what happens. If not, I can just start at 30 and install and configure it as needed. Then snapshot it and make the necessary customization. Just like I did to get these 4 servers.
In the future I would definitely avoid using their images. I have run into some insanity and I wasn't sure if it was a mistake I made or the image. Ended up being the image.
What insanity was that? If their image is bad, they have support for that. If you don't use it, you are left on your own. I would not advise avoiding their images.
I installed a CentOS image at one point and their image had selinux disabled by default. When I enabled it, the VM wouldn't boot. I put a ticket in and the reply was "we disabled selinux because our customers have problems with it enabled" and my reply was "if it is only disabled then why when I enable it does the VM break?". They said they would look into it. I checked back in 3 or 4 times and the reply I got was "we are working on the issue". I just never used their images again. They are customized in a way where I don't really know what they changed and if I run into an issue in the future I may not be able to fix it.
This is my reasoning anyway.
-
@scottalanmiller said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
@wrx7m said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
@dafyre said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
@black3dynamite said in Fedora 30 - EDAC skx: Can't Get tolm/tohm Error After Upgrade From 29 (Vultr):
I wonder if its the CPU model that you are assign to.
This would be my guess. It looks to be related to the Intel i10nm Ice Lake(?) CPUs.
Yeah, the few Google results I found mentioned a CPU, although I thought I saw skylake.
What Vultr "level" is the VM?
Compute 55GB SSD LA and Dallas.
-
@scottalanmiller said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
As a starting point, are these installed when you install via ISO?
# dnf list | grep edac edac-utils.i686 0.16-20.fc30 fedora edac-utils.x86_64 0.16-20.fc30 fedora edac-utils-devel.i686 0.16-20.fc30 fedora edac-utils-devel.x86_64 0.16-20.fc30 fedoraThe error is present using their own installer - Compute 55GB Server Type, Fedora 29 or 30.
However, I also tried an install using the custom iso option and net install. That one yielded the error during the first boot after install, as whatever patch causing the issue was applied during the install.
-
@scottalanmiller said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
dnf list | grep edac
From one of their own server type installers, as shown in my previous post.
-
@scottalanmiller said in Fedora 29 and 30 - EDAC skx: Can't Get tolm/tohm Error on Vultr:
As a starting point, are these installed when you install via ISO?
# dnf list | grep edac edac-utils.i686 0.16-20.fc30 fedora edac-utils.x86_64 0.16-20.fc30 fedora edac-utils-devel.i686 0.16-20.fc30 fedora edac-utils-devel.x86_64 0.16-20.fc30 fedora
From one that was upgraded to 30. -
Vultr support was interested after my previous response.
"We understand your concerns and appreciate you bringing issues to our attention, however, we are not seeing this issue with our Fedora 30 build when running updates. It is possible that something is different when upgrading the distribution manually and this is not something we would be able to troubleshoot as we are not seeing the issue.
Does the OS itself throw an error when attempting the update? If there is an error being throw by the OS we can try to assist in figuring out what it means and try to point you in the right direction, but as a self-managed service our in depth troubleshooting steps for software issues is extremely limited.
Thank you for your understanding."
I provided them with specifics on how to produce the error.
-
Vultr got back to me last week with an email -
"Hello,
We apologize for the delayed response to this ticket, while we investigated this matter further.
We are continuing to review our OS template and application images, and should be applying comprehensive changes across all available deploy images in the near future.
Your continued patience and understanding is appreciated in the meantime.
If you have any followup questions or concerns, please feel free to reach out."
