avij wrote:No, I don't think this is caused by UEFI.
I'll preface this by saying that when I had this problem was at the release of CentOS 7.0. My buddy and I have EXACT Mirror Image Machines -- EXACT same hardware, hardware makers, etc. --as we built them at the same time... we simple took one order and doubled it. This makes trouble shooting any problem more-or-less easy, as we can replicate any problem. The heart and soul of the systems is based around a pair of ASUS M5A97 R2.0 Motherboards with an AMD FX-6300 hexacore CPU's and 16GB of G.Skill Sniper DDR3 1600 MHz 9-9-9-24 RAM sticks, and a ASUS nVidea GeForce GT 640 Graphics Card.
When CentOS 7.0 came out I got to be the "Official Test Guinea Pig" for the install. I always do a Custom Install partition by partition, he always does a Default LVM install. The first problem we encountered was even in my Custom Install, it *insisted* that there be a /boot/efi partition, which was the same with his. I'd did my install, booted up the machine, and after a few flickers, I ended up with with nothing but a black screen. He then installed it his way and ended up with the EXACT same results. It took several reinstalls before I set it up with a /boot partition -- not a /boot/efi -- partition and that solved my problem. My buddy was still relying on the DEFAULT LVM w/ /boot/efi partition and getting the same black screen. The next thing he did was a modified Custom Install where he modified his LVM install but changed /boot/efi => /boot. We also had to change the DEFAULT hardware device from UEFI => Legacy. With that we had no problems. With one minor exception, and that was the microcode problem, which occurred after the first Welcome To CentOS message we have had NO PROBLEMS with either machine (except for a Power supply that died 6 months in, and a dead 1TB Seagate HDD that died last December). The only thing that has changed is we have now added another 16GB of the G.Skill Sniper DDR3 1600 9-9-9-24 sticks of RAM bring us up to 32GB, and have added a Crucial 500 GB SSD, and now a pair of Seagate 2TB drives.
The last full install of CentOS 7 I did was CentOS 7.2, and since then updated by the roll over method. The ONLY thing that I figured out was that there is *something* about UEFI that caused us a LOT of grief that simply disappeared when we switched back to a simple Legacy BIOS. The OP's problem may NOT be due to UEFI, however it is still worth a shot IF he has the option of installing via Legacy BIOS to see if that solves the problem. But I can tell you the problems we experienced sound very similar to the OP's, though our machines were set up to be be used as a pair of light duty Workstations rather than servers. Right now the issue over UEFI -vs- Legacy BIOS is a big one as I am specing out eventual replacements based around the ASUS X370-PRO MB w and AMD Ryzen 1800X CPU. There is one potential hiccup: The BIOS is ONLYUEFI ....
