![]() ![]() ![]() Done for now - will update sometime in 2019. Most of it was done using Ansible, which saved tons of time and makes user additions and other changes reproducable for new nodes on the HPC cluster. Type cd msutil to navigate to the Msutil directory where nvrboot.efi is located. Once I got it to boot and installed the SGI-specific software, it was minimal effort to get everything running again. At the shell prompt, type the drive letter or file system number of the system partition, such as C: or FS n, where n is the file system number of the system partition. Click settings (big button on the left) -> boot -> uefi hdd priority -> make sure that the top one is Microsoft Boot Manager (if i recall correctly) That fixed it for me (MSI mobo too) I couldn't update windows 10 is what happened and the efi shell only happens when windows mode is on. Next time I have some dedicated downtime, I’ll give it a try and update this post with the results.įor now, I have a working machine running RHEL 7 and humming (actually screaming the fans never slow down) away in the basement. I reached out to Red Hat and they gave me something to try, but I haven’t had a chance, since it would require rebooting a production machine. I copied the /boot directory from another SGI UV machine running RHEL 7 and that seemed to work as a bandaid. However, this additional step makes rebooting more difficult and when you are rebooting a 120 core and 1.5TB RAM HPC, your heartrate is already higher than normal (if not only becuase of the heat coming off of the servers) and additional work is not ideal. Slowly, through various support articles and Stack Overflow threads, I pieced together a solution that worked. It turns out that I had to manually launch the grub boot loader, but I had no idea how. I was prompted with a: Shell>ĭespite lots of Duck Duck Go-ing (is that a verb yet?), I was stuck. It was the UEFI shell and the commands were extremely limited. It wasn’t the BASH shell that I expected, nor any varient that I had come to know (e.g. I am not one to shy away from shells, indeed I prefer them over using a Linux GUI for most of my server admin, but this shell gave me pause. I then accessed the boot menu on startup (F12), and booted from the USB stick using UEFI. I downloaded the shellfull.efi from Intels EFI Developers Kit, renamed it boot圆4.efi and placed it in efiboot on a FAT32-formatted USB stick. #BOOT FROM EFI SHELL INSTALL#What wasn’t a known issue is that the Boot Loader wasn’t properly configured during install and when the system reboot, it stopped at a UEFI shell. So, the answer is sorta in the original question, just not quite. The install went fairly smooth, save the unresponsive screen from the Live DVD (which is a known issue with SGI UV systems). I got in early that day to make use of the window and got to work. By this time I had moved all of our user’s home directories to an NFS share, moved almost all applications to environmental modules, and started using Ansible for configuration, meaning that I could finally do a clean install of the new OS with minimal configuration on the other side. Because of the version discrepency, we had to implement work arounds for using, say, the latest version of gcc. #BOOT FROM EFI SHELL UPGRADE#I had scheduled some maintenance time for an upgrade to our main HPC, which was still running RHEL 6. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |