I'm having some problems getting this running. I've got two IBM blades, both of which have the Cluster and iSCSI target roles installed. I have created two virtual disks on the first blade for iSCSI use.
If I then set the disks up to be allowed initiators of itself and the other node and also set the initiator up on both blades to access this target, I can pass the disk portion of the failover cluster check with no issues.
However, once I create the cluster and try to add disks inside the cluster, they cannot be seen. Is this not a valid configuration? The link below seems to suggest that you can.
I'm not having a whole lot of luck here. How can the VHD be brought online to get it seen by the cluster without also being an initiator for itself?
Otherwise, what is the best way to create a 2-node file server? It seems like the the target MUST also be an initiator. I just need to be able to have some failover for some file/database storage.
One additional question, do the VHDs need to be created on both servers?
Loopback: There are cases where you want to run the initiator and Target on the same machine; it is referred as “loopback”. In Windows Server 2012, it is a supported configuration. In loopback configuration, you can provide the local machine name to the initiator for discovery, and it will list all the Targets which the initiator can connect to. Once connected, the iSCSI virtual disk will be presented to the local machine as a new disk mounted. There will be performance impact to the IO, since it will travel through the iSCSI initiator and Target software stack when comparing to other local IOs. One use case of this configuration is to have initiators writing data to the iSCSI virtual disk, then mount those disks on the Target server (using loopback) to check the data in read mode.