Hi there; I would like to have some feedback on this to see if there are any glaring items that stand out that should be changed in our Network setup.
I can't speak for the Network Ethernet Switching side of things as that's out of my AOR & AOE(expertise). We have been having major Network issues affecting our Hosts & VMs, and I just want to make sure the VMWare network config is sufficient.
Here's a breakdown of our VMWare vSphere 4.1 environment:
We have two HP DL580 G5 ESXi Hosts (8 CPU licence-split 4 & 4) in separate telecomm buildings, each with 128GB RAM, four HBA's going to a EVA8000 SAN over a dual fabric FC network, and six GBNics, split between two Avaya (Nortel) 8600 Core switches.
Two NICs on each host are dedicated to the VMKernel vSwitch (Management & vMotion), and four NICs are for the general network traffic.(4 uplinks from each host into the vDist switch)
The vDist switch has no options configured. There is no IP Pool, not using NetIOC, no traffic shaping and each VM has one connection and a static IP assigned to it. Although we have many VLANs (for clients) , nothing in the VMWare infrastructure is outside the default VLAN.
Two HBA's go to each Fabric FC Switch. The main VMFS/VMDK Datastore is on the SAN and the user data SAN LUNs are configured as RDM's.
I have attached a screen shot of our dvportgroup traffic, and some of it doesn’t make sense to me. i.e. - Why do some of the ports indicate traffic , when there is no “connectee”?
If you'd like me to post more of the config, please be specific.
Any feedback would be appreciated.
Thanks.
Gregg
Message was edited by: GWOD1 Forgot to add that even though the two Hosts are Clustered, they are on different "legs" of the network. Although the VM's data disks are not Fault Tolerent disks (being RDM's), I thought that if one leg went down, the hosts would lose contact with each other, and the "surviving" host would call the VM from the Datastore(SAN), and there would only be a momentary lag, not a complete loss of access.
Message was edited by: GWOD1 One more thing: I have HA, with Host Monitoring. I had DRS on but just now turned it off after re-reading the description: (attached as DRS.jpg)