R710 2nd CPU Install - E1410

Ah yeah ... memory MUST be evenly configured for each processor.

Hi Flash.

Actually I just cracked it.

Original system single Xean E5504 @ 2.00GHZ Family 06 Model 1A stepping 5

32 GB RAM 4 x 8GB 800MHZ ECC DDR3 1.5 Advanced ECC mode.

system booted fine in this config with memory in 4 consecutive slots starting closest to CPU 1 on left side of MB looking from the front.

Second CPU was purchased.  Exact same specs. From Dell.  Adding it to CPU slot 2 , added 6th fan, Moved memory to slots 2 and 3 in Bank A and Bank B.  System now boots.

i believe I had the RAM in the wrong slots and result was FANs were going crazy and no POST.  LED on from posted "system Software Event" Processor Sensor IERR was asserted.  Moving the CPU's around showed them both to be good.

Thanks for looking anyway...

purchases,

I do see that you are using E5504 processors and I am using E5540's (unless that was a typo). So you have both processors in the server, then you have memory in slots 2 and 3 in both banks but you are skipping slot 1? Were you able to put anymore in? I have 48GB RAM to put in this server. lol.

I called Dell for phone support on this to see if out of warranty support could help but the guy gave me some run around AFTER 20 minutes of asking me questions knowing that at the beginning he saw it was out of warranty and knew I was not the original owner, he said there is nothing he can do because free phone support isn't offered on servers after 2008 or something, complete bs Dell.... So much for being top customer service.

Thanks
Jessy

Im going to put more in (96GB total) but all I have handy is 32GB.  I used the matrix from the hardware owners manual here:-

http://www.dell.com/support/Manuals/us/en/04/Product/poweredge-r610

To answer your question though, yes I am skipping 1 and using 2 and 3 on Banks A and B and Advanced ECC mode in the BIOS, per the matrix.

Just for the issue of upgrading Firmware for Systems that run under ESXi (i also have some running on ESXi):

Get the OMSAlive from the System section of the Downloads for your Model. Burn it, boot it, so you should have a centos Linux running.

Take a second Machine, download the FW for your model as Linux binary (.bin) and put it on a USB stick.

use the CentOS to update the FW. (Terminal -> [dir where your USB stick is] -> ./[firmwarefile] -> follow the instructions on-screen

another possibility is to use the SUU (which is really much to download)

___________

How to find out the Stepping? put the according CPU to socket 1, go into bios and open your eyes. do the same with the other CPU and you will see if there's a difference.

hope this helps.

You won't find OMSALive under downloads for your system.  The latest version is located here:

http://linux.dell.com/files/openmanage-contributions/omsa-71-live/OMSA71-CentOS6-x86_64-LiveDVD.iso

Did you use some type of head magnifiers? These processors do not have pins like the older ones - if these have pins they are tiny, almost impossible to tell if they are bent.  But I am having all kinds of problems with a CPU/RAM upgrade on a PE R710, so I am more than willing to check and make sure no pins are bent.  The processors are flat though, and do not have anything resembling pins (at least what I consider to be pins).  I have 2 of these R710 servers that are completely identical and both have the latest BIOS.  One of them worked perfectly when I upgraded the CPU and RAM.  I removed the CPU and replaced it and added a second CPU and also added 96 GB of RAM using the ECC mode (2,3) (5,6) and (8,9) populated.  Each DIMM is the same size and speed and even the same brand.  So again, I did the exact same upgrade on both servers - removed old CPU, added 2 identical new CPUs, and upgraded the RAM by pulling all the RAM out and adding back 12 modules of 8GB each.  One server worked perfectly but the other one is just not cooperating.  I tried clearing the NVRAM - and at least now it is recognizing all the RAM but one of the CPUs is getting an error... the E1410 error.  I think what I am going to do is take it down to only one CPU on the problem server and see if that works.  But if I find anything else out I will update the thread.  If you can explain to me what the pins look like, since I do not seem to know, I would be extremely grateful.  Many thanks!

Sam

Can confirm the issue was caused by bent pins on CPU1 for me as well. The reason this is fine is because the pins were bent were on the Quick Path Interconnect (QPI) side Intel uses to link 2 CPU's together. If you don't have a second CPU installed, then there's about 150 PIN's on CPU1 that aren't active, which is why this issue doesn't crop up until you try and install a second processor.

Look for the south side of the socket to look for the bent pins, I was able to spot mine due to the irregular look of the pins (like trees in a forest, there's a uniformity to the pin layout).

I just experienced the exact same thing with an R510 that I was trying to add a 2nd CPU to.  However, when I added the 2nd CPU, not pre-boot screen or error of any kind.  But thanks to this post, I looked at pins under the 1st CPU and found a few bent.  Literally must have come from Dell that way because I have never had CPU 1 out before.  

Fixed pins and walah!!

Thank you for this.  I had been pulling my hair out on this one.