Dell PowerEdge M620 issues

[orioon]

Hello,

we are having problems with Dell PowerEdge M620 machines.
http://www.dell.com/us/business/p/poweredge-m620/fs
These are blade servers which are installed in a special chassis:
http://www.dell.com/us/dfb/p/poweredge-m1000e/pd

Already updated Bios to the lastest version, as well as the iDRAC and other components which might not seem important.

One thing to note is that we have 2 different hardware versions of this machine.
The older ones have two E5-2620 CPUs while the newer ones have two E5-2620v2 CPUs installed.

I am seeing the following errors:
1) It immediately shows 4 ECC errors on the new servers (with E5-2620v2), no matter if I use single or parallel CPU mode. I also tried to change the address range (like 3g start, 4g end), but that didn't help either. Tried booting from USB and Virtual Media. The older model does not have the issue.
Posted a result screenshot, got no log file for this one right now.

2) Mouse support is broken when using Dell iDRAC (latest version), KVM works fine though.
Is there any way I can select items from the list on the left (like CPU Selection) without a mouse? Left arrow and TAB don't seem to work sadly.
I think keyboard shortcuts for this would be really great or am I just missing it?

3) When trying to view detailed SPD info no modules are display at all (Happens on both models)

4) CPU temperature is not shown, it stays at 0C (Happens on both models)

I hope you can help me figuring out these issues

I mailed you the log from an older blade without the ECC issue and the results file from a newer one without this issue.
(Yes I know, the other log file would be much more interesting but obviously Memtest86 will overwrite the previous log file if you boot it in an other machine.... It would be much more better if naming rules were identical to the results)

Kind regards,
Benjamin

[keith]

Thanks for the logs. We'll have a look at it and get back to you.

1) It immediately shows 4 ECC errors on the new servers (with E5-2620v2), no matter if I use single or parallel CPU mode. I also tried to change the address range (like 3g start, 4g end), but that didn't help either. Tried booting from USB and Virtual Media. The older model does not have the issue.
Posted a result screenshot, got no log file for this one right now.

Looks like it may be a MemTest86 bug. We will have a look at the log file you sent us to try to debug it.

2) Mouse support is broken when using Dell iDRAC (latest version), KVM works fine though.
Is there any way I can select items from the list on the left (like CPU Selection) without a mouse? Left arrow and TAB don't seem to work sadly.
I think keyboard shortcuts for this would be really great or am I just missing it?

Yes, there are keyboard shortcuts to access the different screens on the left sidebar. The character in parenthesis is the keyboard shortcut you need to press to switch to the corresponding screen. For example, you would press 'c' for (C)PU Selection.

3) When trying to view detailed SPD info no modules are display at all (Happens on both models)

4) CPU temperature is not shown, it stays at 0C (Happens on both models)

Both issues are chipset specific, so we need to look at the details in the log file.

(Yes I know, the other log file would be much more interesting but obviously Memtest86 will overwrite the previous log file if you boot it in an other machine.... It would be much more better if naming rules were identical to the results)

MemTest86 shouldn't overwrite any existing log file. New logs are appended to the existing file.

[orioon]

MemTest86 shouldn't overwrite any existing log file. New logs are appended to the existing file.

Thats good to hear, in this case the log should contain a short run on both servers at least
Thanks for looking into this, much appreciated.
Edit: Just found they start at line 7830, that's perfect then!
Logs may end a bit weird as I used to pull the flash drive out during runs.

I also noticed that the run time seems to be a lot different between both models, the older model seemed to skip areas of the tests.
Don't know if this is a bug or something else, or what I could do to get more details on this.
As the sever also seemed to crash during the test I should get an other one of the old models to rule out hardware issues.

Could you maybe explain in which manner the Pro version offers customizable reports?
That seems very useful along with the other features like ECC injection.
Also I would like to know in which manner the Pro version can be kept up to date from a long term point of view.
is there something like a time limit for updates or is it maybe linked to major releases?

Thanks in advance!

[keith]

I also noticed that the run time seems to be a lot different between both models, the older model seemed to skip areas of the tests.
Don't know if this is a bug or something else, or what I could do to get more details on this.
As the sever also seemed to crash during the test I should get an other one of the old models to rule out hardware issues.

If you can run the same tests with the same configuration on both machines, we can compare the execution times from the log files.

Could you maybe explain in which manner the Pro version offers customizable reports?

In the pro version, there are 3 template files (header, footer, stylesheet) that you can edit to customize the appearance of the report (for example, if you need to include a company letterhead or need to use a standardized stylesheet for your organization). These elements are hardcoded in the Free version.

Also I would like to know in which manner the Pro version can be kept up to date from a long term point of view.
is there something like a time limit for updates or is it maybe linked to major releases?

For the moment, upgrades to major releases are free for users who purchased the Pro version.

[orioon]

If you can run the same tests with the same configuration on both machines, we can compare the execution times from the log files.

I sent you the files when I was back at work. I had to abort the slower machine at the end of pass 3 though.

For the moment, upgrades to major releases are free for users who purchased the Pro version.

Just noticed that the feature page provides the information.
"Free upgrades for 12 months after purchase."
That looks fair, so we have a concurrent user license with 12 months of support and lifetime usage if I am not mistaken.

Really wondering if the freezes did occur after the run finished...
Will give this a try again.