Poweredge 2950 III Cpu chk and DRAC 5 problems

Question

Hi :)

I have a poweredge 2950 III at home, I use it for my Plex server, gaming server, and backup server. I have install windows server 2012 r2.

Bios version: 2.7.0

DRAC 5 version: 1.20

BMC version: 2.50

PERC 6/i integrated version: 6.3.3.0002

2x Xeon E5450 3GHz

32GB of ram.

Soo ! :) 2 problems ! First, my server crash sometimes when i think he get full used maybe ? I was renderring plex thumbnail and he crash..

I got two errors:

1_CPU Machine Chk: processor sensor, transition to non-recoverable

2_System Software event: run-time critical stop was asserted

I try to check on internet without real help.. i update bios bmc perc without any change.

My second problem is I can't update my DRAC.. I try to install the version 6.5 by sending the update by the internet page who tell me the page get too long to reply. I try to update with windows update file, he find my drac and say he can update it to 6.5 then he stop and say try to restart pc...

Hope you can help me ! Thanks you ! :)

DELL-Chris H · Answer

Quentin3225,

Did you recently update the BIOS to current? As the issue may be due to the BIOS being updated to current and the Drac is so far behind. It is best practice to walk them up together, as you can see issues if the gap is too wide between them. You may want to try walking it up by starting with version 1.40 or something and then try to go to the latest version afterwards. Also, is are you seeing anything else in the hardware log as well?

Let me know what you see.

savvy2 · Answer

check the CPU for overheating?

go to DRAC

see the page, on sensors and logs.

check for overheating

is your server room (cough cough ) air conditioned or room at 100f?

Booted from USB?

Booted from SAS , HDD, sata or raid,? (PERC)?

what is your boot media, ? if a HDD does it pass all SMART tests?

if using raid for the boot, we need to do special tests to see all smart parameters.

x2 do all firmware updates, in the correct order, per dells instructions.

quentin3225 · Answer

It was happening before I update everything. I try to find older DRAC5 version but i always get to the last 1.60 version... Do you know where I can find older drac5 version ?

It's not a heat problem, I got no error and my server is only 24*c and when I stop it and open nothing is hot expect ram but that's normal. No other error in drac. I boot from a hard drive who is in raid 1.

I do test and it appears that's when the CPU get full load (about 90%) the server crash. I have removed my 32gb of ram and only place 8gb. And guess what it works... I do the same test and it works this time no crash. So that's probably a ram problem ? I will try to increase ram and see how it goes.

savvy2 · Answer

by crash , BSOD crash, blue screen of death>?

you have bad RAM.

try them 1 by 1, in bank 0 , to see of all sticks are ok.

but use 1 CPU only to test all sticks 1by1 then with 2 CPU, and 1 stick each. and then add them.

are they all same brand and model or a mix.

to be sure , the only test if RAM is to run piles of code./data trough it,

load photo shop 10 times, or

most memory tests are very crude tests, and only find hard failures.

Bad RAM can be a hair puller, that is for sure.

i guess this is not ECC ram?

quentin3225 · Answer

Yeah, blue screen ^^

I will try one by one thanks. But how can i set only one CPU ?

They are all the same : https://www.itcreations.com/view_product.asp?product_id=22860

they don't look to be ECC. Are they good or should I change them ?

Thank you ! :)

theflash1932 · Answer

The 2950 can ONLY take Fully Buffered (FB) ECC RAM. The slots are keyed so you can't put the wrong RAM in. Non-ECC/non-registered/non-FB RAM will NOT fit in the slots.

CPU CHK can be many things: CPU, slot (motherboard), RAM, or add-in cards. Do you have any add-in cards besides the PERC?

CPU CHK was also VERY common on this machine and was addressed by BIOS update, sounds like you do, but make sure everything including the PERC and backplane firmware is up to date.

If you are troubleshooting the CPU's, CPU slot 1 MUST be occupied, so if the issue is reliably repeatable, then try both CPU's in slot 1, leaving slot 2 empty.

Make sure that memory specs match. Run Dell's 32-bit diagnostics on the RAM (clear the hardware logs first). If you can force an error on one, then swap them to see if the error follows the DIMM or the slot.

Do you have errors in your Windows Event Log?

quentin3225 · Answer

Hey :)

Sorry I was busy with work. So I first check for updates, all is done and I think I'm good. I try to test one ram at the time but it still crash. I try to use the Dell diagnostic but I probably doing it wrong... I try to boot with my USB key but when it try to boot on USB it say unplug everything. I try many things but still don't work, I'm probably the problem haha.

The only thing who is plug to my mother board is the perc, drac and a dvd reader. I saw something who speaks about it and I don't use it so I remove it. For the moment I have also removed my cpu 2.

I try multiple test with cpu a 100% and ram at 100% and no problem. I increase ram and try again and no problem for the moment. I will keep going and increase ram to 32gb try multiple stress test and if everything is good I remove the cpu 1 and try the other cpu who was in slot 2 and start again by increasing ram and check if I got any issues. Could the dvd be a problem ? I hope the problem come from the cpu and not from the motherboard .

quentin3225 · Answer

Back with some news :)

I was using only one cpu (cpu number 1 in slot one)

I try multiple test and put the cpu and ram at 100 % of use with 8, 16, 24 and 32gb. I don't have any crash so that's great !

I then do the same with the second cpu (cpu number 2 in slot one)

I do the exactly same and nothing happened no crash !

I try to put back the cpu 1 in slot 2 and it crash with the first stress test ... Could it be a perc issue or the motherboard..? How can i check that ?

Thanks you :)

PowerEdge Hardware General

Poweredge 2950 III Cpu chk and DRAC 5 problems

Was this post helpful?