Hyper-V networking question and UCS
-
יום שישי 04 מאי 2012 20:44
I'm new to Hyper-V, but have a project that will be using it. I am setting up a Hyper-V cluster with clustered shared volumes on Windows 2008 R2 Enterprise. The physical servers are Cisco B200M2's. I've successfully installed Windows 2008 R2 Enterprise and presented the appropriate storage. Both of the servers that will be in the cluster see the storage appropriately. I have the M81KR adapter in the machines and setup 4 nics for each Hyper-V server. 1 nic for management, 1 nic cluster shared volume, 1 nic for live migration, and 1 nic for the vm network. The nics for management, csv, and live migration are each on there own vlan. In UCS I've set them up so they can failover to the other fabric is there is a fabric interconnect failure. I configured each network interface with a static ip address. The management network interfaces are the sole interfaces that I've configured with an gateway IP address. The other interfaces just have an IP address configured. The problem I'm seeing is once I configured the csv and live migration networks with IP's I lost connectivity to the Hyper-V boxes completely. I can't connect these two Hyper-V boxes with cross-over cables for live migration and csv so how do I set these Hyper-V boxes up appropriately on UCS or for that matter anywhere where you aren't connecting the Hyper-V boxes with cross-ver cables. Thanks for any assistance.
RG
כל התגובות
-
יום שישי 04 מאי 2012 21:15
Hello RG,
How do you configure csv and live migration networks with IP's?
Do you give gateway ip address to these networks?
I think for the problem; there should not be gateway address for CSV and live migration. If you correct that issue you can use without problem.
As following picture.
Sincerely,
Murat Demirkiran
If the post helps you and remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. www.scvmm2012.com (TR Language) Please note that I'm not a Microsoft Representative or SpokesPerson.
-
יום שישי 04 מאי 2012 21:20that's the thing. I configured them without gateways. The Operating actually complains if you attempt to add gateways. The only network interfaces on the machines with gateways are the management interfaces. They were working up until the moment I configured IP's on the live migration and csv interfaces. The minute I did that I lost all connectivity to them remotely. The odd thing is if I go into the console of the machines through UCS I can ping the IP's of each machine. The have access to the internet for downloading patches, etc. I can ping the live migration, csv, and management IP address of one node from the other and vice versa. The firewalls are disabled and there are no firewalls between the machines I'm trying to manage and my workstation.
RG
-
יום שני 07 מאי 2012 05:53מנחה דיוןHi,
In UCS I've set them up so they can failover to the other fabric is there is a fabric interconnect failure.
>> What did you mean by UCS? What did you mean by “set them up so they can failover to the other fabric is there is a fabric interconnect failure”? Did you want to have network adapter failover when one of them fails?
Based on my experience, there is no need to configure network adapter failover in a Windows Server cluster configuration, except MPIO. I guess this is the cause of the issue.
-
יום שני 07 מאי 2012 17:40
It's not the fabric failover. When I did the NetApp/Cisco Microsoft FastTrack 1.0 solution, I used fabric failover on all the interfaces. Where are your vlans defined and how do pull the configuration down?
-
יום שני 07 מאי 2012 18:35
Hi John,
Did you update your firmware of UCS?
http://www.cisco.com/en/US/docs/unified_computing/ucs/release/notes/OL_25363.html
Sincerely,
Murat Demirkiran
If the post helps you and remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. www.scvmm2012.com (TR Language) Please note that I'm not a Microsoft Representative or SpokesPerson.
-
יום שני 07 מאי 2012 18:42
You don't configure any reduncancy for the network in your clusters? I would disagree. We have numerous MCSC and all of them have network teaming and it works fine.
UCS is Cisco's Unfified Computing System - http://www.cisco.com/web/solutions/data_center/ucs_servers.html?POSITION=SEM&COUNTRY_SITE=us&CAMPAIGN=ucs_servers&CREATIVE=UCS+-+Brand_UCS&REFERRING_SITE=Google&KEYWORD=cisco+ucs_p|mkwid_sPXd3yxz5_12732271093_272267a16717
RG
-
יום שני 07 מאי 2012 18:47
I defined my VLANs on the LAN tab under LAN Cloud > VLANs. Then I created som vNIC templates (one for the management vNIC, one for the clustered shard volumes, one for Live Migration, and one for the virtual machine network). I've configured each vNIC template for a different VLAN and then I associated the vNIC templates to my service profile templates which I built two machines out of.
My present firmware level is 2.0.(1t).
RG
-
יום שני 07 מאי 2012 18:53
Hi RG,
I dont think you have a problem in OS Layer.Everything seems normal in OS Layer.
As you said , you had created MSCS in other UCS part.
There should be Private network configured. Can you compare with Hyperv side.
Sincerely,
Murat Demirkiran
If the post helps you and remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. www.scvmm2012.com (TR Language) Please note that I'm not a Microsoft Representative or SpokesPerson.
-
יום שלישי 08 מאי 2012 00:43
We had to update the firmware. If this is a NetApp/Cisco hardware platform, you should be able to get the Fast Track deployment guide from your NetApp rep.
J
-
יום שלישי 08 מאי 2012 01:57מנחה דיוןHi,
Please provide network topology of your network in details, a diagram should be better.
.png)
