Need assistance in seeing duplicate MAC address in a Virtual Connect Domain

[chuckk281]

Robert was trying to diagnose a customer MAC address problem:

******************

Team

I have a customer that is seeing a problem with a specific MAC address duplicated seven times in the environment, we delete the profile and reboot to post and the real MAC shows up. We power off the blade, assign a profile and with no power applied to the blade the MAC address changes back to the duplicated address.

My immediate need is someone or some type of documentation that walks through the MAC assignment process end to end in a C-class blade environment with VC. This would include any level of MAC interaction in VC, OA and the Blade server itself.

Flow chart would be great if it exists.

Current overview of configuration:

• When a profile is assigned you can watch the real mac cage to duplicate mac, sometimes back and then eventually it sticks to the dupe mac address.
• We are not running VCEM
• Chassis are in 3 different data centers.
• Different VLans.
• Not connected to each other in any way.
• VC MAC is set to static not virtual.
• VCs in chassis are configured for one vlan out one port to a CISCO network. The configuration makes VC act like a bridge to the Cisco network.
• Replaced system board and problem returns.

Any thoughts or ideas would be great..

*******************

DIane had some help:

*****************

Check out this document on pages 63 and 64.

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02516247/c02516247.pdf

*******************

And then Jon also stepped into the conversation:

*******************

        On each and every enclosure that features this duplicate MAC address, refer to the table at http://bizsupport1.austin.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&taskId=110&prodSeriesId=3601866&prodTypeId=18964&objectID=c02494590&printver=true .

    Find out what address range the address came from within all applicable server profiles returning this MAC address.  Then compare it to the current HP predefined address range specified within the domain.  They should be the same.

    This should be true for only one enclosure/domain.  If multiple domains feature the same HP predefined address range setting (i.e. HP predefined range 1), then there is an opportunity for duplicate MAC addresses appearing on the network.  The HP predefined range setting must be unique for each domain.

    If the HP predefined address range is correct, then deleting and recreating the same server profile would not make much of a difference.  It would get the same address from the same pool.  However, on those domains in which the current HP predefined address range setting does not include the duplicate MAC address, it is not possible for the new server profile to get the same MAC address as it is obtaining an address from the current (different) pool of MAC addresses.

*************

Any other thoughts or suggestions for Robert?

[BeltwayBenny]

I've got the exact same issue as the original poster.

We have the following MAC address in 3 different datacenters connected only by Layer-3 so the MACs are not crossing datacenters.

Duplicate MAC is D8:D3:85:DF:E0:F0.  Each blade chassis is in its own domain and we are using the factory MAC option at Ethernet > MAC Addresses > Use the static, factory default MAC Addresses

Our MAC is not in this table:

http://h20566.www2.hp.com/portal/site/hpsc/template.PAGE/public/kb/docDisplay/?sp4ts.oid=3601866&spf_p.tpst=kbDocDisplay&spf_p.prp_kbDocDisplay=wsrp-navigationalState%3DdocId%253Demr_na-c02494590-1%257CdocLocale%253D%257CcalledBy%253D&javax.portlet.begCacheTok=com.vignette.cachetoken&javax.portlet.endCacheTok=com.vignette.cachetoken

This is causing some major issues for us, anybody have any ideas?

I've removed the server from VC profile, booted with no profile, reset the system via F9, etc, etc with no fix.

Ben