Windows Server TechCenter >
Windows Server Forums
>
Clustering
>
CSV - Acces from nodes to ClusterStorage\VolumeX
CSV - Acces from nodes to ClusterStorage\VolumeX
- Hi all,
Last week I deployed two Hyper-V Server 2008 R2 (core) in order to virtualize some servers at our organization.
Everything is going fine, but I can only acces to c:\clusterstorage\volume1 from the node which owns the CSV disk at that moment. If I try from the other node, I can just get to c:\clusterstorage, but I can't access to Volume1.
I know that csv is only intended for Hyper-V, but I think that when I deployed both servers I could acces to Volume1 from both nodes (although copy or whatever was going redirected through the node which owned the csv disk resource).
What I can do is to move the CSV from one node to another (this allows me to access ClusterStorage\Volume1 from the node that I hadn't access before and stopping having acces from the node I use to have access) and I even can live migrate virtual machines.
I'm also receiving Event 1069, FailoverClustering:
Cluster resource 'Cluster Disk 1' in clustered service or application 'Cluster Group' failed.
This events happens after rebooting Hyper-V server and before loging in.
Thanks in advance,
Ll
All Replies
- Maybe this posts helps:
http://social.technet.microsoft.com/Forums/en-US/windowsserver2008r2highavailability/thread/5b0f981e-3714-4280-97b1-73ffe31913f6
Otherwise I need more specific information about your configuration (NICS,Nodes..)
Bye
ThorstenWujek - I think I have a mess with my NICS.
Environment:
EMC Clariion SAN
Active Directory + DNS server (192.168.0.211)
Nodes:
Node1
Name: stahyperv01
OS: Microsoft Hyper-V server R2
FC fabric to connect to the SAN
5 Physycal NIC's
Joined to the domain
Node2
Name: stahyperv02
OS: Microsoft Hyper-V server R2
FC fabric to connect to the SAN
5 Physycal NIC's
Joined to the domain
SAN:
2 LUN's
Quorum: 1 GB
CSV: 725 GB
Presented to both nodes
Node1 NIC's:
LAN01: (Physical NIC) Hyper-V external network for VM to connect to the LAN. This NIC in the Hyper-V has the checkbox "Allow management operating system to share this network adapter" cleared in order to be used only with VM. When I ipconfig the server this NIC is not listed.
DMZ01: (Physical NIC) Hyper-V external network for VM to connect to the DMZ. This NIC in the Hyper-V has the checkbox "Allow management operating system to share this network adapter" cleared in order to be used only with VM. When I ipconfig the server this NIC is not listed.
LiveMigration: (Physical NIC). Used for live migrate the VM.
IP: 10.10.20.106/24
DNS: 192.168.0.211
HeartBeat: (Physical NIC). Used for cluster and CSV communicattions.
IP: 10.10.10.106/24
DNS: 192.168.0.211
Management: (Physical NIC). Used for manage
IP: 192.168.0.106/24
GW: 192.168.0.101
DNS: 192.168.0.211
Local Area Network* 958: Microsoft Failover Cluster Virtual Adapter
IP: 169.254.1.116/16
Node 2 NIC's:
LAN01: (Physical NIC) Hyper-V external network for VM to connect to the LAN. This NIC in the Hyper-V has the checkbox "Allow management operating system to share this network adapter" cleared in order to be used only with VM. When I ipconfig the server this NIC is not listed.
DMZ01: (Physical NIC) Hyper-V external network for VM to connect to the DMZ. This NIC in the Hyper-V has the checkbox "Allow management operating system to share this network adapter" cleared in order to be used only with VM. When I ipconfig the server this NIC is not listed.
LiveMigration: (Physical NIC). Used for live migrate the VM.
IP: 10.10.20.108/24
DNS: 192.168.0.211
HeartBeat: (Physical NIC). Used for cluster and CSV communicattions.
IP: 10.10.10.108/24
DNS: 192.168.0.211
Management: (Physical NIC). Used for manage
IP: 192.168.0.108/24
GW: 192.168.0.101
DNS: 192.168.0.211
Local Area Network* 249: Microsoft Failover Cluster Virtual Adapter
IP: 169.254.2.24/16
Cluster:
Validate Cluster before created: OK
Name: ClusterHyperV
Access Point to manage cluster (in the Create Cluster Wizard): I selected Management NIC (and assigned 192.168.0.107) and HeartBeat NIC (assigned 10.10.10.107).
DNS:
Three hosts registered as ClusterHyperV:
192.168.0.107
10.10.10.107
10.10.20.107
Three hosts registered as stahyperv01:
192.168.0.106
10.10.10.106
10.10.20.106
Two hosts registered as stahyperv02:
192.168.0.108
10.10.10.108
10.10.20.108
**This cause that sometimes DNS resolves the hosts as the ip 10.10.10.X or 10.10.20.X and makes impossible to connect hosts from the Lan (which is 192.168.0.X). I think I should delete those records from the DNS server, but I prefer to keep it as it were registered automatically.
Failover Cluster Manager:
Cluster Core Resources:
- Network: 10.10.10.107 and 192.168.0.107
- Disk Drives: Cluster Disk 1 (Volume Q) (Disk witness in Quorum)
- Cluster Disk 2 (CSV)
- HeartBeat: Allow Cluster networ communication on this network and Allow clients to connect through this network (I marked this checkbox in order to bring cluster network core resource online because it was offline and I couldn't bring it online).
- LiveMigration: Allow Cluster networ communication on this network .
- Management: Allow Cluster networ communication on this network and Allow clients to connect through this network .
The problem is that when I created the cluster I could access c:\clusterStorage\Volume1 from both nodes, and now I can just acces from the node which owns the CSV disk resource.
I'm wondering if I can destroy the cluster and create it again, but I have some VM in the CSV and I don't know if when I will add the storage again to the CSV all the VM will still be there or everything wil be lost.
I think this is a mess, so thanks a lot if you can help me, and thanks a lot if at least you've read till this last line.
Ll - Why is your cluster resource in the 10.10.10.x network when your lan is 192.168.0.x ?
can you please provide a Get-clusternetwork|ft ?
Bye
ThorstenWujek - The cluster resource is in the 10.10.10.x network because I thought it must be a different network from lan in order to operate the cluster in a correct way.
PS C:\Users\Administrador> import-module failoverclusters
PS C:\Users\Administrador> get-clusternetwork|ft
Name State
---- -----
HeartBeat Up
LiveMigration Up
Management Up
- Percu1
a good rule o fthumb is to have one CSV per node and each node own the volume, or even better more than one volume per node.
you have to remember if you need to tkae ht evolume1 offline for main then all of your virtual machines go offline, I would add more volumes ot the cluster min one bper node plus Quorum.
William Bressette, Network Architect, Horn IT Solutions, http://wpjsplace.spaces.live.com/blog- Proposed As Answer byWPJB Thursday, November 05, 2009 3:02 AM
- Marked As Answer byTim Quan - MSFTMSFT, ModeratorMonday, November 09, 2009 2:12 AM
- Unmarked As Answer byPercu1 Monday, November 09, 2009 7:49 AM
