@computersavvy @gnwiii I did the memtests. I’m confused. But here are results:
*Case 1: [OS BOOTED BEFORE] Shutdown, electricity off, boot => errors
*Case 2: [NO OS BOOTED BEFORE] electricity on, reboot (via ESC [Memtest]) => no errors
*Case 3: [OS BOOTED BEFORE] Shutdown, electricity on, boot => errors
*Case 4: [NO OS BOOTED BEFORE] electricity on, reboot (via OS restart) => no errors
I also made photos because I can’t save a log file. If you need these, just tell me.
Should I try to regenerate the initramfs (dracut --regenerate-all --force)?
Should I try to start from an earlier kernel?
This seems to be a new motherboard, and supposedly new RAM, so one of the first steps I would do after having memtest86 show me errors would be to try reseating the RAM into the sockets on the motherboard, as well as verifying that the RAM is inserted into the proper slots as indicated by the documentation for the mobo. It may be a bad contact on the ram or it may be a mismatch in which slots are in use. Reseating the ram after checking the proper config should be able to rule both those out.
With memory errors reported by memtest86 and errors during boot the RAM should probably be the first approach.
Ok, just for me:
The errors will be sent in a second message (I hope I can upload all the photos in 1 message, if not, I’ll do more than 1 message).
I have an ASUS Prime B550M with G-Skill 32GB Ram, and I think you need to install your ram in the pair slots, so 1 and 3, or 2 and 4 or all 4 or a single one.
I agree with the comment by @jakfrost
This is why I suggested checking the documentation for that motherboard, and I have seen NO motherboards that support RAM installed the way those images show.
For my B550 mobo it suggests first using slots 2 & 4 then 1 & 3 as alternate for 2 DIMMs.
Those images show 16G installed with 2 DIMMS (8G each) in slots 2 & 3.
@computersavvy (pls tell me if that pinging is annoying)
I saw a mistake (red circle, lightning) there and I corrected it like the correct one (red circle, green hook):
I didn’t remove & place the RAM again (is that the meaning of “to reseat”?) because the RAM was already in the right place…BUT if reseat means to remove and place the RAM again (in this case), I’ll do that of course (I’m not sure that’s why I was asking). Here’s the placement for the RAM:
Sorry if I’m confusing in any way!
I suggested checking the docs, and you posted the image from the docs.
Clearly that shows that for 2 DIMMs they should be in slots A2 & B2. From the numbering in memtest it would appear that you have the DIMMs in A2 & B1. I believe that A2 & B2 are both gray slots while A1 & B1 are black. I have checked the online images at asus to confirm. Your image seems to show both in the gray slots, but the seating issue would be critical.
Reseating means to reinsert them into the proper slots especially when you have been asked to check the documentation and verify the proper installation. (If they were inserted properly to begin with then reseating would mean remove and reinstall in the same location)
The second DIMM, (Upper in the first image) was clearly not fully seated as can be seen with your upper red circle. Failure to fully seat a DIMM is a common cause of errors since it results in poor contact.
If the DIMMs are seated in slots A2 & B2 as suggested and now properly and fully seated then repeat the memtest. If errors are fixed then good. If not then we may need to do more digging.
empty
Ok, thanks for the explanation/clarification. I’m going to remove & reinstall the DIMMs into the slots again then if the thunder stopped (it just started).
empty
Ok, I’m going to do that once I removed & reinstalled the DIMMs + if the thunder is over.
Cya!
I’m back. The errors still consist. I hope the mobo isn’t damaged because of this weird behavior (slots & errors) 
.
The errors:
Try removing one of the DIMMs and use only one in slot B2 then test again. It may be a damaged or bad dimm, but usually it would only be one.
If one tests bad then swap and try the other. If both test bad in that slot then try the other slot (A2) one at a time.
Ok, just for me:
I have to turn off the computer again, thunder is back 
.
testing this way will identify if the issue is DIMM or MOBO.
New board and DIMMs should be covered under warranty if replacement is needed.
Hey Jeff, I’m done with testing. Here are my results:
DIMM 1, Slot B2, no initramfs:
DIMM 2, Slot B2, no initramfs:
DIMM 1, Slot A2, no initramfs:
DIMM 2, Slot A2, doesn’t boot at all:
At first attempt the mobo turned off & on again - it was like in a loop, repeating that until I turned off the computer. Second attempt the DRAM control LED on the mobo lighted up, nothing else.
It looks like the DIMM 2 is a bad one. Another thing which I wonder about is that slot B2 still gets recognized as Slot 3 even if it’s the 4th slot.
The memtest86+ docs note the lack of consistent names and numbers assinged by vendors to slots, so recommend removing and rotating modules to identify the one causing failures.
Ok, Dimm 1 passes in both slots.
Dimm 2 fails so should be replaced.
You could use DIMM 1 in slot B2 (single dimm arrangement) while getting the replacement for DIMM 2.
Note that memtest does not always order things in the order that the mfgr uses for designating slots.
I have to send both back because they were both within one package. But that’s totally fine.
What stands “mfgr” for?
Hmmm did I get you right? Do you suspect the mobo is damaged? I identified which DIMM isn’t working. The DIMM 2 (I simply named it like that so that I don’t confuse DIMM 1 with DIMM 2) is the bad one.
“mfgr” is short for “manufacturer” and is a commonly used abbreviation.
Many vendors are willing to do a cross-ship in situations like yours, so you may be able to continue with one DIMM until the replacement is received, then send back the old pair. Worth trying.
