Raptor Computing Systems Community Forums (BETA)
Raptor Computing Systems Hardware => Blackbird => Topic started by: bw42 on July 15, 2020, 10:36:42 pm
-
The system says something is there, but it doesn't show the size.
I didn't notice any errors on boot.
I swapped the modules around to see if it worked with the one that wasn't detecting and it works fine in the A1 slot.
I've read some systems use half the RAM for ECC, so you can only use half of what is installed.
Does the Blackbird do something like that?
-
No. My Blackbird has its RAM stick in B1 and sees all of it. What type of RAM?
-
Hi,
My Blackbird came with 16GB (Samsung) on the back slot (relative to the user seating in front of the computer). I've bought another 16GB from Crucial (CT16G4RFD8266 16Gb Ddr4-2666 Rdimm, 16 Gb (1 X 16 Gb), Ddr4 Sdram, 2666 Mhz Ddr4-2666/Pc4-21300, 1.20 V, Ecc, Registered, 288-Pin, Dimm) installed it and it worked for a few days.
Then it stopped worked and I've tried all combination front/rear slot samsung/crucial and found that both sticks work on the front slot but not on te rear slot, giving me only 16GB. Neither work on the back slot alone and I suspect the back slot is not working or there is some firmware issue that was triggered when I added the additional stick on the back slot.
I opened a support ticket on Raptor (my machine is brand new and came with a defective power supply, which took a month to be replaced). It seems your issue is similar to the one I'm experiencing.
Have you opened a ticket with Raptor?
Regards
-
The RAM I bought is Micron 2x32GB DDR4-2400 ECC Registered.
The system picks it up and works with either module, as long as its in A1.
Just doesn't pick up the module in B1.
So that's disappointing.
-
Oh. I forgot to mention: I had asked support about the memory I was intending to buy (Crucial) and they replied that it should work with the memory stick that came with the machine (Samsung).
Have you checked this link: https://wiki.raptorcs.com/wiki/POWER9_Hardware_Compatibility_List/Memory ?
Regards
-
Yes, I've looked at the wiki page.
The memory works, both modules work, but only in the A1 slot.
It shows something is in the B1 slot, but it doesn't use it.
-
It sounds like the RAM was GUARDed out by the firmware. It probably detected some issue with it and disabled it.
You can try clearing the GUARD partition via the BMC when the host off:
pflash -P GUARD -c
This won't fix the reason why it was GUARDed out, but it should re-enable the slot.
-
w00t!
After running that command in the BMC it is picking up both RAM modules in the OS.
Is there a log in there somewhere that says why it was guarded out?
-
Hi, I'm new to Raptor. How do I issue this command? After powering up and before booting? On the boot menu?
Thanks
-
Hi, I found it on petitboot (shell). It worked for me also, but it remains unknown why that happened.
-
Thank you meklort!
-
guys this problem just happened to me too. I had 16GB with 2 banks of 8 each, lately I got a 16GB one and wanted to use it additionally but as soon as I put it in slot 1b, it stopped seeing me that slot and now it only detects slot A1, not seeing the slot anymore B1 ... do you think I also have to execute that command by entering the Blackbird BMC?
-
guys this problem just happened to me too. I had 16GB with 2 banks of 8 each, lately I got a 16GB one and wanted to use it additionally but as soon as I put it in slot 1b, it stopped seeing me that slot and now it only detects slot A1, not seeing the slot anymore B1 ... do you think I also have to execute that command by entering the Blackbird BMC?
It happened to me again last week and I ran that command again and it worked. I do not what caused the memory bank to be guarded and I hope it won't happen soon again.
-
ok thanks cchincz, I'll try to do that command at this point and I hope it works for me too ... Thanks
-
In the end I also went through the procedure of deleting the GUARD partition through the BMC and it seems that now slot B1 has started working again. Very strange, however, as a defect, in the end I did nothing but replace one RAM bank with another and obviously everything was off and misaligned. It is probably a bug in the BMC software. Note that I performed the procedure not from a Linux system but from a Mac Book Pro M1 ...