Yum Update To Centos 7.4 breaks Ryzen 1800x NVME workstation

[sidestreetpost]

Hello all.

So I ran "yum update" today and that seems to have broken my Centos 7.3 installation.

I now get an error message:

"Failed to open \EFI\CENTOS\grubx64.efi - Volume Corrupt
Failed to load image \EFI\CENTOS\grubx64.efi: Volume Corrupt
start_image() returned Volume Corrupt"

I tried reinstalling Centos 7.3 and do a yum update but still the same error

I've tried googling to see if I can just yum update to specific 7.3 packages after a clean Centos 7.3 install but that does not seem to exist because the repos are "latest versions" or nothing at all.

I tried booting into Centos 7.3 usb image rescue mode and ran grub2-install /dev/sda but it had modinfo.sh missing errors.

I also tried "grub2-mkconfig -o /boot/efi/EFI/redhat/grub.cfg" but the same error as above regarding "Volume Corrupt" appears again.

Totally regretting running "yum upate" right now so casually.

What am I missing? I'm a bit confused.

[desertcat]

OK this might be caused by a few problems which you need to check.

1) UEFI and EFI are half of a boot process that you must check, the other half is BIOS. You boot one, and you are toast if the machine is in the other. My buddy and I have IDENTICAL ASUS motherboard based machines we scratched build. When we initially built the machines I set mine to BIOS and he to the default which was UEFI. While I could get through the install process he could not even get the machine to the install starting point. As soon as he switched to BIOS and NOT UEFI it was smooth sailing. So your starting point is to check your BIOS first and seeing if it is using a UEFI enabled device or a device based in BIOS (ie NOT UEFI). You will see the same device listed twice: once as UEFI and once as NON-UEFI. It could be that the load order changed and it trying to load the wrong version of the device. If not sure trying booting from both the UEFI and NON-UEFI version of the device and see if it makes any difference.

2) This is similar to a problem I am now facing with the "Trash-can Monster" ( OLD HP machine I "rescued" from the dumpster) which keeps giving me the following message: "NVM Checksum Is Not Valid". This has lead to a LOT of hair pulling.You could simply have a too new piece of equipment and are simply lacking drivers necessary in CentOS 7.4 Try running the following command as root in a knosole: lspci. Second download (from a different machine of course) a copy of Knoppix 7.x.x. This is a Linux Utility Disk. See if your machine will boot from this. Likewise download both a "Live" version of "Linux Mint" as well as CentOS 7.4 "Live" and see if you can or can not both live versions. I quickly discovered that both Knoppix and Linux Mint found both my Wired Ethernet card and my Wireless WiFi cards in the "Trash-can Monster", neither found my On-board Ethernet; OTOH my CentOS 7.3 "Live" version FAILED to detect any Ethernet controllers. After some more horsing around I downloaded a driver for the 3com Ethernet card, and it WORKED!! Sadly the same could not be said for the Wireless WiFi card -- downloaded a new driver for it and it still is not detected; I also found a driver for the On-board controller, and it too still does not work.

3) I remember reading reading something in the release notes about NVMe in the release notes thinking it might solve my problem.

https://wiki.centos.org/Manuals/Release ... ntOS7.1708

Hope these ideas point you in some directions to try.

[sidestreetpost]

Thank you so much desertcat

In the end we had to rebuild the workstation from scratch, but thankfully the 7.4 ISO worked fine as a clean install.

We of course had to go ext4 instead of xfs because for some reason ext4 is the only format that will work with the NVMe.

Thanks again.