Ok going to keep this on the simple side. Running esx 5.5 and vCenter 5.5. The ports on the switch are configured with channel-group # mode active, no I am not using a vDistributed Switch. What would happen if I changed them to the recommended config of channel-group # mode on? Yes this is a production environment and it has been this way for a long time.....surprised the consultant I brought into the setup our environment missed this. Everything else is like what is described in VMware KB: Sample configuration of EtherChannel / Link Aggregation Control Protocol (LACP) with ESXi.... I just am unsure what to expect when I make this change.
You won't lose connectivity, but better to do in a maintenance window anyways. Even better, when you have a maintenance window, remove the etherchannel for your storage traffic completely. It's likely not needed and only adds complexity, especially if you're using iSCSI. Check out Chris Wahl's numerous posts on the subject.
Edit: I should say that I'm surprised an LACP bundle is working without LACP configured on the ESXi host. That seems odd.
Please consider marking this answer "correct" or "helpful" if you found it useful.
VMware, Cisco Data Center, and NetApp dude
Message was edited by: Mike Brown
IMHO, the results could be catastrophic. It's not immediately clear if it's just the SAN ports or also the ESXi interfaces that are misconfigured. If this involves your ESXi host interfaces, and your change causes the ESXi management interfaces blip, this could cause fake HA events (uncheck the 'Enable Host Monitoring' HA setting during maintenance) or worse, the VM storage could blip causing deeper wounds. I would recommend gathering a CDP Report using your favorite PowerCLI script (example) and sharing that with your network team so the interfaces being scrutinized are well understood.
Before doing any activity, ensure that all ports listed in your CDP have clear port descriptions and the upstream port channel health is known and understood. I do agree that in your case, simply getting rid of the channeling would simplify your setup but making any change of this sort could have dire consequences at this point. Since this is production, consider engaging your network switch vendor and/or VMware Support to be safe.
Note: Remember that mode active error checks before creating the channel participants whereas mode on will not.
MikeB, while vCenter shows the ports on the host show as up the port-channel shows as down, each individual port is up, on the switch. Going to read those etherchannel posts now. May be this isn't as big of a deal as I am using iSCSI.
Other Mike, Well I figured I'd put the host into Maintenance Mode before I started to reconfigure the switch ports. I figured better safe that sorry.
You can configure Ether channel on SAN for following reason
1) To Increase the throughput.
2) for Load balancing
3) for High Availability
You need to configure teaming by creating bonds if you have HP SAN storage and there are set of commands need to run at Switch level.
there are multiple types of teaming option available in HP Storage.
After spending some time reading vSphere 5.5 Documentation Center I think I see where my issue is. Take a peak at this screen shot of my vSwitch that connects to the SAN(Dell Equallogic by the way.
Now from my understanding is that I would need 3 more VMKernal Ports, all with IPs in the same vlan, change the teaming policy so that each as a physical adapter as active and the rest as unused. Then bind the new VMKernal ports to the software ISCSI adapter on the storage page. Question now is about the other two VMFT and vMotion, can I leave them as is? Maybe bind each of them to just one physical NIC?
I still can't believe the consultant missed this.........