SCVMM Deploying Clusters
In this video we demonstrate how quick and easy it is to create a cluster in SCVMM.
Prerequisites
Identify Cluster Name & IP Address To Use
- If using DHCP address will automatically be assigned
- Can source from SCVMM IP Pool
Plan LUN usage
- Witness
- CSVs
Steps to deploy cluster with SCVMM
1. Verify networks are in place
- Logical Switch deployed with SCVMM, see our previous post!
2. Verify storage is in place
- Presented to hosts with SMI-S via SCVMM, See our previous post!
3. Create Cluster Wizard
- Cluster Name (this is the name you will use for managing the cluster)
- -Select a runas account that has administrator privileges on the hosts and the ability to create computer account in AD
4. Select nodes you would like to be included in the cluster
5. Set cluster IP address from IP Pool, manually define an IP, or leverag DHCP
6. Configure storage LUN uses, format, label and create CSVs
7. Configure virtual switches if appropriate
- This was already complete, See our previous post!
Pingback: Deploying and Using SCVMM – Part 0/6 | Hyper-V Rockstar
I have created the vnics as you mentioned teaming all my phsyical interfaces into a logical switch using VMM. When running the cluster creation through VMM, it doesn’t show the IP address page – the only way i seem to be able to get this to show is if i have a physical adaptor free (removed from the logical switch).
It looks like you didn’t have to do that according to your videos – can you confirm what is happening here, and why it doesn’t seem to allow the IP address page to display if I have all my physical adaptors allocated into the team. If it is the case that it cannot have this configuration set up – is there any point in configuring the logical switches before hand?
Thanks
A
Hi A,
I’ve seen this happen rarely (once or twice). When I have seen it, I have resolved by removing nodes from the list that I planned to use until the IP Address Page does show. I would recommend going through this wizard and add only a single host first, and see if the page shows up. If it does, I would add a second and look again. Use this process iteratively until you determine which host is causing the problem. Once you have determined that, review that host carefully with the other hosts to determine what may be different. If you want to move forward quickly, build the cluster with the hosts that seem to work fine, and then use Failover Cluster Manager to add another host later. When you refresh SCVMM, the other host will also show as in the cluster.
Thanks for your swift reply sir!
I have only 2 hosts to add to the cluster (this is us testing VMM out before purchasing).
Neither host allows for the IP address page to be added.
Each server has 4x 1gb nics, which are teamed in a logical switch as per your configurations in these guides. I have 4 vnics hooked into seperate networks all with IP pools configured. They are all receiving IP addresses and default gateways from these networks that are configured to receive those settings from VMM’s IP pools. They are not using DHCP.
As far as I can see, both hosts are configured identically – the only exception being the name of the adaptor as a captical M for management, rather than a lower case m one one host compared to the other.
Any other ideas?
Thanks again
Hi Ax,
In your reply you mentioned that all of your networks have default gateways. Can you further elaborate? Typically you should only have a default gateway on a single vNIC, the one used for the management of the host. The other vNICs should have only an IP address and Subnet Mask. The cluster network, live migration network, and even networks used for iSCSI storage should not have gateways.
Also, have you applied SCVMM Update Rollups?
Hi Eric,
Thanks again for getting back!
Interesting you say that, because the microsoft article here:
https://support.microsoft.com/en-us/kb/2679673
mentiones that each network should have a default gateway?
I have installed all the rollups.
Many thanks for your help and your time sir!
Kind regards
A
Hi Ax,
I don’t think the article to which you supplied a link is particularly clear on which vNICs require a gateway. When a cluster is created, all of the networks on each node are enumerated to the extent they are on different subnets. When a network is found with a gateway, the cluster assigns it as a role 3 network. When there isn’t a gateway, it is assigned as a role 1 network, and will be used for cluster communication, CSV communication, and Live Migration. We recommend to leave gateways off of all adapters except the management adapter. I don’t think any of this really answers your question.
Another method to move past this issue is to configure all the networks from SCVMM using a logical switch as you’ve done, and then use failover cluster manager on one of the hosts to create the cluster. Once this is completed, you can refresh one of the hosts in SCVMM and it will recognize the cluster. This may be the best way in your scenario.
Further to your comment, I have adjusted the networks as you mentioned (with the only vNIC having a default gateway being the management network) and it still will not show the IP address page on either node.
Thanks
Aaron
As an aside, I have already done what was mentioned here:
https://support.microsoft.com/en-us/kb/2679673
Hi this video series is absolutely the best I have ever seen for training. You explain things better than even the paid services. Just wondering when we can expect videos 5 and 6?
Thank you so much
Part 5-6 and 6-6 missinf
Where are video 5 and 6?