DL 580 G5 with ESXi 5.0 show Memory warnings

[odenbach]

Hi,

we have upgraded to ESXi 5.0 on 10 blades and on 6 DL 580 G5 and have installed the HP ESXi offline bundles.  After the upgrade each DL 580 G5 box shows 32 memory warnings (status unknown). There were no errors with ESXi 4.1.

We have tried the original VMware ISO + HP offline bundles and also the HP provided ISO, both with the same result.

Has anyone experienced the same problem? What can I do?

Thanks for help,

Christopher

[Jags_21]

Hi,

I wonder if the problem is with the System Management Homepage (SMH) i.e which is there in HP offline bundle and HP ESXi 5.0 image. If you can reinstall any of the server with vmware image without offline bundle and check what the status it shows, before that please check if the server is upto date on the firmware's if not update the same.

Thanks,

[odenbach]

Hi,

with ESXi there is no System Management Homepage (SMH), so there is no other place I could check the status of the memory modules. Well, not quite: Interestingly, we use Icinga to monitor our hardware, and the used plugin check_esxi_hardware.py does not show any errors.

The firmware is up to date, I did this right before installing ESXi 5.0.

Christopher

[Michal Kapalka (mikap)]

Hi,

if you have the ILO configured i would start there to check HW status and of course the IML log,

if you have there some entries clear all of them.

If you have the HP management agent installed, ( if you are using ESXi HP branded it should be automatically installed )

Michal Kapalka (mikap)

[odenbach]

Hi,

I have already checked the IML and the HW state in iLO - everything is fine! We experience the problem on six machines, so it must be a software issue. The question is whether it is an HP problem or a VMware problem.

On ESXi there are no real "management agents" as there were on ESX. With ESX there was the system management homepage (SMH) and the management agents which could be contacted byb SNMP. On ESX there are only CIM providers, so there is no SMH and no SNMP. It does not make a difference if we use the VMware ISO and the HP offline bundles or directly the HP ISO.

A search revealed at least two other people experiencing this problem on DL 580 G5:

http://communities.vmware.com/message/1851259

Christopher

[Michal Kapalka (mikap)]

Hi, as you already mentioned, that this is on six machines, and if you have the latest firnware on the proliants, i would recommend to open a support call on vmware support ( if you h ave a valid SnS contract ). Michal Kapalka (mikap)

[Michal Kapalka (mikap)]

Hi,
 
i spoke with my friend, and he told me that proliants have some spacial counting on the memory,
 
so if ths counter is not cleared with the smartstart CD, som special procedure,
 
the error should ocupied which you are recieving from the ESXi servers,
 
if some technikan replaced the memmory he should reset the counter.
 
Check this link :

HP ProLiant 300/500/700 Servers - How to Troubleshoot Memory Related Issues?

Michal Kapalka (mikap)

[odenbach]

Hi,

I have checked the memory with the HP SmartStart DVD, as expected no errors. There were no errors with ESXi 4.1, so why should there after installing ESX 5? And the exactly the same errors on six different machines?

No, there must be something wrong with either the HP offline bundle which collects the data from the machine or with ESX 5 which then interprets this data.

I have opened up a case with VMware, we will see if they can find any reason.

Christopher

[odenbach]

Hi,

VMware sent me back to HP. They said, that VMware just displays the data they get from the HP cim provider. If there are wrong error messages displayed, these errors must come from the HP cim provider.

What can we do now? I do not have any HP carepack on these rather old boxes to open a case at HP. Is there any other way I can have this confirmed (and perhaps repaired) by an HP engineer?

Thanks,

Christopher

[allencrawford]

We have the same issue.  We just got started with ESXi 5.0 and sure enough, every HP DL580 G5 we've upgraded has this problem.  So far the only other model we've upgraded is the 580 G4, which is unsupported, but it doesn't toss any memory errors.  Not a good test.  We may grab a DL380 G5/G6 we have in our lab and try it real quick.

As for getting it fixed, I'm working with VMware and others and will likely open a ticket with HP as well.  I'll keep this thread posted.  But, if anyone from HP is on here, please feel free to help us.