Have been reading up on iSCSI networks , and get the point about binding,
Just want clarification about hosts that have say 2 seperate iSCSI networks amd 3 vmkernel ports
e.g
iSCSI Network A
vmk0 - 172.20.50.10
vmk1 - 172.20.51.10
iSCSI Network B
vmk2 - 172.20.26.20
I'm guessing I bind all the vmkernels to the iSCSI adapter vmhba32 ?
iSCSI network B has already been setup (badly) by another company and is in production , so we can't fix it (touchy customer).
> I'm guessing I bind all the vmkernels to the iSCSI adapter vmhba32 ?
Yup. That would be the right approach. You'll see a few additional log messages about "network unreachable" when the daemon tries to log into
the 172.20.50.x targets through vmk2 and the 172.20.26.x targets through vmk0 and vmk1. These are expected and not harmful.
Andy
You need to bind all the vmk nics to SW iSCSI adapter. You may refer to http://blogs.vmware.com/kbtv/2011/01/how-to-configure-iscsi-port-binding-on-vsphere-4.html and the video to get details about 'How to configure iSCSI Port Binding on vSphere 4'
~Sudhish
You say the "network unreachable" messages are expected but not harmful. When I setup a config like this I see HBA rescan times go up significantly. With two vmkernel interfaces the times go up to around 90 seconds. On a config with four vmkernel interfaces the time goes up to around 5-6 minutes. Starts to look broken.