Unsolved
This post is more than 5 years old
1 Rookie
•
106 Posts
0
1202
September 7th, 2018 03:00
M640 IPMI correctable memory error
I have inserted 6 brand new M640 blades into a M1000e chassi.
They all seem to work fine, but they all report "correctable memory error" when checking their health via IPMI.
They all have iDRAC version 3.21.21.21 (and cant find any upgrade for this, so its the latest i guess). BIOS v1.4.5 on all, and CPLD (whatever that is) v1.0.0 on all.
Using FreeIPMI tools i get a warning for ID 51 which is memory bank A = "correctabl memory error".
All 6 blades say the exact same thing. None of the 10 older blades in this chassi have any problems (they are M630s).
And checking the health by logging in on the iDRAC web UI, everything seems fine, like this:
All 6 new blades have same 2x32Gb memory.
What should i do? :)
Daniel My
10 Elder
•
6.2K Posts
0
September 7th, 2018 09:00
Hello
You can review the hardware logs and run diagnostics on the memory. Single bit or correctable memory errors can occur without being reported. They only produce an error when the error rate exceeds thresholds. The system maintains a log of memory errors. The IPMI command you are using may be reporting that there are correctable memory errors reported against the modules. If you are not seeing the errors reported by our management and monitoring tools then it is not exceeding thresholds.
http://www.dell.com/support/
Thanks
Alexander-36725
1 Rookie
•
106 Posts
0
September 11th, 2018 00:00
Ok @DELL-Daniel My, thanks for the explanation!
One thing that i would like to fix though is that in these 6 new M640 blades, i dont see the same IPMI-IDs presented.
From 4 of the servers i see this: https://hastebin.com/azopazoxag.rb
From the remaning 2 i see all this: https://hastebin.com/gexidataki.rb
All those 6 servers have iDRAC9 Enterprise (v3.21.21.21).
Can i upgrade/update something to make the servers "identical" when it comes to IPMI? All servers have the same specs (same CPUs, memory, disks, controllers, etc etc).
wbr / Alex
Daniel My
10 Elder
•
6.2K Posts
0
September 12th, 2018 09:00
I don't know what IPMI IDs you are referring. Those pictures are not pulling up for me. The iDRAC acts as the BMC, so updating the iDRAC to the same firmware revision should have them all respond the same. The system BIOS may also affect the enumeration. If the IDs are are not somehow static then there may not be a way to get them all enumerated the same way.