Cluster 2008 R2 with iSCSI
Hardware and OS:
2 server (Node1 and Node2) with Windows 2008 R2 Enterprise,
3 network adapters on each server configured as follows:
INT LAN (10.10.10.1/24, 10.10.10.2/24)
Heartbeat (1.1.1.1/30, 1.1.1.2/30)
iSCSI (192.168.0.1/24, 192.168.0.2/24)
1 SAN iSCSI MSA2312i with IP 192.168.0.3/24
Purpose:
Create a cluster with iSCSI SAN
I haven't other servers, then the two nodes are also domain controllers.
I installed 2008; I created the domain; I have enabled servers with the iSCSI initiator and connected them to the SAN; I have formatted the two Lun (Lun0 = Quorum; Lun1 = Data); I added on the servers, the Failover Cluster feature; I ran The Cluster Validation Wizard, with zero failure and warning.
At this point I created the cluster with ip 10.10.10.3/24, assigned as the quorum Lun0; configure the network cards :
INT (LAN client and cluster traffic)
Heartbeat (traffic cluster)
iSCSI (no traffic for cluster and client)
Now I want to test failover, so I disable the INT LAN card on Node1 .
From a client PC I noticed that the ping has stopped and then is restarted.
In the management of the cluster, the quorum has moved on Node2.
Until here everything ok. (right?)
Now I have re-enable INT LAN card on Node1.
I disabled the LAN card INT on Node2.
From a client PC I noticed that the ping has stopped and is not restarted.
In the management of the cluster, the quorum is still ONLINE on Node2, but the quorum is unreachable.
If I re-enable the INT LAN on Node2, the cluster is still not working.
The only way to bring life the cluster, I have stop and start the service on cluster Node2.
How can I fix?
Thank you.
p.s: sorry for my english ;)
2 server (Node1 and Node2) with Windows 2008 R2 Enterprise,
3 network adapters on each server configured as follows:
INT LAN (10.10.10.1/24, 10.10.10.2/24)
Heartbeat (1.1.1.1/30, 1.1.1.2/30)
iSCSI (192.168.0.1/24, 192.168.0.2/24)
1 SAN iSCSI MSA2312i with IP 192.168.0.3/24
Purpose:
Create a cluster with iSCSI SAN
I haven't other servers, then the two nodes are also domain controllers.
I installed 2008; I created the domain; I have enabled servers with the iSCSI initiator and connected them to the SAN; I have formatted the two Lun (Lun0 = Quorum; Lun1 = Data); I added on the servers, the Failover Cluster feature; I ran The Cluster Validation Wizard, with zero failure and warning.
At this point I created the cluster with ip 10.10.10.3/24, assigned as the quorum Lun0; configure the network cards :
INT (LAN client and cluster traffic)
Heartbeat (traffic cluster)
iSCSI (no traffic for cluster and client)
Now I want to test failover, so I disable the INT LAN card on Node1 .
From a client PC I noticed that the ping has stopped and then is restarted.
In the management of the cluster, the quorum has moved on Node2.
Until here everything ok. (right?)
Now I have re-enable INT LAN card on Node1.
I disabled the LAN card INT on Node2.
From a client PC I noticed that the ping has stopped and is not restarted.
In the management of the cluster, the quorum is still ONLINE on Node2, but the quorum is unreachable.
If I re-enable the INT LAN on Node2, the cluster is still not working.
The only way to bring life the cluster, I have stop and start the service on cluster Node2.
How can I fix?
Thank you.
p.s: sorry for my english ;)
ASKER
I noticed that the network cable from the ex-Node1 in Failover Cluster Manger -> Networks -> Cluster Network 1 (my LAN INT) both cards are Fail.
Why?
Why?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
I did the test in a virtual environment, with 2 member servers and 1 DC, same result, so even if the cluster is created with 2 dc there is no difference.
I added a resource to the cluster with Lun1 and ip 10.10.10.4/24.
I dropped time Fail to 1 hour (the default is 6 hours).
I did the same tests.
I have disabled INT LAN on Node1 and the cluster is passed on Node2.
I have re-enabled the INT LAN on Node1, I waited 1 hour and I disabled the INT LAN on Node2. The cluster is passed on Node1.
At this point I said, it works ok finally.... mah ...... I got the bad idea to do another test.
Instead of disabling the card INT LAN servers, because hardly anybody disable the network card but it is more likely that is missing the link to the network, I tried unplugging the network cables in LAN INT network card of the servers.
On Node1 I unplugged the cable from the rj45 network adapter INT LAN and unfortunately the cluster is not passed on Node2.
Why does this happen? For the cluster is not the same thing to have the network card disabled or missing network card links ?