Creating a Link Aggregation Group with vSphere 5.5 and LACP

[Music] vmware vsphere 5.5 comes with the ability to create some really cool link aggregation groups on the distributed switch using a multitude of hashing algorithms that rely on LACP so we're able to form dynamic port channels between your host and whatever physical switch that it's plumbed into let's walk through that process it's pretty simple but there's some various steps that you have to learn and pretty much all of this is new in 5.5 if you've made a port channel or an ether channel using 5.1 or earlier you'll notice that this process is drastically different so to begin I'm going to assume that you don't have your distributed switch if you do you can skip this part so we're actually I've got a demo folder inside of my networking container here and I'm just gonna create a new distributed switch from scratch and we'll call it I'm not going to be very creative we'll call it VDS - LACP to denote that this is the distributed switch that I'm forming LACP pour channels on you can call it whatever you like the name really doesn't matter now this is important if you want to do the new enhanced LACP support which is a feature provided with the distributor switch 5.5 every host that'll be attaching to the switch has to be running ESXi 5.5 or later and your vCenter obviously needs to be 5.5 or later as well or else you won't even see this option so go ahead and select 5.5 and click Next now we can choose the amount of uplinks that will be available for the switch just like you normally would nothing changes here I have two available and I'm gonna make a default port group but I'm not gonna call it D port group because that's kind of silly right let's call it something a little more descriptive we'll call it the production VMs maybe that's where you put your production VMs in fact I'll say on B LAN 105 there we go it's a semi descriptive port group name that gives me a little bit of a clue as to what's going on again call it whatever you want that makes sense for your environment click Next you'll review to make sure that it's not going to explode on your face and as long as it looks like this you can click finish and you're good to go at this point the wizard is going to do is wizardry and create a distributed switch for you and we'll go ahead and expand this and see if it's done that now it's still working okay now that the switch has been created let's select it and actually build a link aggregation group so I'm gonna go to the manage tab for this switch and then you'd want to click settings which happens to be already clicked and then LACP which is the link aggregation control protocol and it's what we're going to use to build the link aggregation group so that we can form a port channel between our hosts that are running ESXi and the physical switches that they're plumbed into using their network adapters so even though we don't have any hosts on this switch yet I want to go ahead and create the lag so that it's available when we go to add a host to the switch we're gonna click on the green little plus sign here that lets you create new link aggregation groups or lags click plus there and we're going to enter a little bit of a wizard here so the name I'm gonna call it the Spongebob oh one link aggregation group it doesn't really matter what you call it and in a work environment Spongebob may be a questionable name for it we're gonna make this port channel consume two different up links or two different ports and two is the minimum obviously you can't have a group of links unless you have two or more I'm gonna change the mode from passive to active and all that really means is that we're going to actively seek out the other end of our interfaces so we're going to advertise the fact that we want to form a port channel with the other side of the equation the physical switch that were connected to if we chose passive then we would be relying on the physical switch that were plugged into to do those advertisements I typically like to have both sides active just as a fail-safe to make sure that either the virtual switch or the physical switch is going to advertise the desire to form a port Channel then we have the load balancing mode now if you remember in previous versions of vSphere this would have to be route by IP hash and that basically meant it would look at the source and destination IP address and hash those values in 5.5 I'll just kind of scroll slowly you have lots of different options you can basically use TCP and UDP which is layer for data you can use the IP address which is layer 3 right here you can use the MAC address which is layer 2 or as you can see here we can also use the VLAN tag which is part of the Ethernet 802 dot1q so you have tons of options about how you want to actually do your load balancing it's gonna really depend on what you're putting on the port channel to make sense for you there's no one right answer I'm just gonna stick with the source and destination IP address because it's what we used to use and whatever it's just the one I'm going to choose it looks like okay and in a moment we will now have a new lag called spongebob zero one it uses two ports so we can have two network adapters tied into this lag the mode is active so it's gonna actively reach out to its peer members on the physical switch and ask it to form a port channel and the VLAN settings will be inherited by the port group that uses the lag we're not defining it on the on the lag itself we're gonna let the port group that's using the lag define what VLANs will be available to that port group so now we need to actually add a host to this switch I can just go up to the actions menu here and click add or manage hosts it's already on add host which is good and I only have one host to add so this always kind of confuses me at first I see I want to add a host and list is empty that just means that you have to go to the little button here says new hosts and choose a host in this case I only have a single host called ESX 2 so I'm gonna select that host click OK and now we see the list has a new host with the new in brackets or in parentheses rather and it says ESX 2 is connected and ready to go so we'll click Next now I don't have any VM kernel adapters to migrate to this switch so I'm going to uncheck that I do however want to manage my physical adapters and assign those network adapters to the up links for my link aggregation group so I'm gonna leave this checkbox selected click Next and now I want to add my two different network adapters in this case I have 4 available but the one that I want to use our VM neg 1 and VM Nick - I've already determined that those are the two network adapters that are plugged into my upstream switch so I'm gonna click on one and then assign the uplink and you'll notice there's the usual uplink 1 and uplink two that you're probably used to seeing and now we have specific ports for our link aggregation group port 0 and port 1 so it gives you the name which is the spongebob oh 1 and at the end of it it puts a 0 and ones in the note port 0 and port 2 1 so the two ports so I'm going to put uplink 1 on the first port and I'm gonna sign up link 2 to the second port and now we can reconfirm that and see that the - spongebob link aggregation group ports are now assigned to VM nic 1 & DM nick - I can click Next there's no impact - I scuzzy that's fine click Next again and finish now we're adding the host and during that add process it's going to map the physical network adapters that are going up to my upstream switch to this link aggregation group and you could have multiple link aggregation groups you could have a pair or multiple ports going to one set of switches for production let's say another one for maybe a DMZ zone or a web facing zone whatever it makes sense for you you can have many different lags you're not limited to one like you used to be now that that's been added we need to do some work on the physical switch to actually build the other side of the link aggregation group before I do that however I'm going to go to my host and I'm gonna look up what ports it's using using lldp so here we can see a list of the physical adapters that are on ESX 2 and if I go down to Nick 1 I can click on it and I can click lldp and I'll let me know we don't actually have lldp turned on yet give me just one second we'll go back to the network and let's actually turn that on so I'm going to edit this particular switch and go to advanced and we'll change it from cisco discovery protocol or CDP to link layer Discovery Program and I'm gonna set it to both advertise and listen like okay on that let it apply there we go now we can go back to the host and take a look at its information so I'm gonna scroll back down look at Nick one go to lldp and there we get information that this is attached to modulo 1 or rather switch one module 0 port 6 so now I know that I need to modify port 6 and this one's on port 12 so port 6 and port 12 are the two ports on the physical switch that need to be put into a link aggregation group so I'll switch over to my HPV 1910 switch and let's go ahead and do that work to form the lag and this will depend very highly on what kind of switch you're using this is the switch that's in my lab so it's what we're gonna use so I'll click on link aggregation and all I did was I logged in to the web platform clicked on network and then click on link aggregation and I already have one lag going to one of my psychology storage arrays so we're not going to touch that one and that's aggregation group number one so let's create a new one called link aggregation group number two so we'll go through this little wizard it defaults to the number two because that's the next available number and we don't want static because that would be just a regular ether channel with no advertisements going back and forth we want to use dynamic because that will use the link aggregation control protocol so that tells you a little hint right there saying hey LACP is enabled if you choose this option and that's what we want so now I'm going to choose port six and port 12 let's scroll down just a hair here and you can see aggregation interface ID number two is going to use member port six and 12 and it's dynamic so we're using LACP I can click apply let's go through the process of actually building that interface and it's pretty much done there's not a lot to it we can see now it's kind of hard to see it's a very dark gray number two meaning this interface and this other interface interface six 12 are now part of the bag or the bonded aggregation of number two ID number two so now we can go back to the summary and we can take a look at this actual aggregation group number two here so right now it's telling me off the bat that this is not configured properly it's not happy about it there we go it just takes a little bit of time for the LACP data units to pass back and forth and actually form the link aggregation group I think it's just the fact that the HP switch while quite robust is not really meant for I guess that kind of enterprise instant-on via the web platform it probably takes a little bit time for that information to be sent to the web GUI but if you just click on the button a few times you'll see that it changes from selected ports to standby ports 0 and that you want to see the reason for being unselected is just a pair of dashes meaning that they're selected they're operational and they're up so there's a few pieces of housecleaning left we need to actually set the VLAN tagging on our new lag or bag as it's called inside of this HP switch so if I click on VLAN here and go to port detail will see that this port here this is port 6 but has the two with the underline meaning that this is lag number two and when I click on it actually drops down here and says this is the Gigabit Ethernet port 1 0 6 we'll see that it's currently an access port now prior to making the change of adding the link aggregation group to the HP switch it was a trunk port and it was passing 3 different VLANs however now it's an access port if you look at this port as well it's also an access port and the actual aggregation group right here it's right down here it's hard to find sometimes click on that we see that it is also an access port as well now there's no need to change the individual ports by making changes to the link aggregation group itself it changes the underlying ports in this switch but that may not be the behavior on your physical switch so what I'm going to do over here is go to modify port and we're going to select the link aggregation group number 2 the bag number 2 I'm going to change the link type to a trunk which is right here will allow us to pass multiple VLANs and I'll apply that change will click close there so now that that change has been initiated I can now tell it which VLANs to tag on the trunk so I'm gonna select the bag number two again tell it what VLANs to tag and I'm going to do to 51 through 253 because those are the VLANs that I need in my environment I'll click apply and there we go now if I go back to the port detail we can verify that these changes have been made the way we want click on port the tail then I'll click on the BA Gigi we can see that the native VLAN or the untagged VLAN is going to be VLAN 1 the 3 VLANs that we're tagging and passing along are 251 through 253 and that it's a trunk port and we can verify this with our other ports by clicking on port 6 and port 12 and as you can see by making that change to the link aggregation group it's also gone ahead and done us the courtesy of changing the underlying interfaces to match so that's it for the switch it's now operational all of the ports are passing the VLANs that we want they're set to a trunk you may have a need to make them an access port in a lab but for most realistic environments they're going to need to be a trunk we'll do one last check on our link aggregation group and we can see that everything is still selected nothing is stand by and everything's operational we now need to switch back to the VMware environment one more time and I've got this port group here the production VMs VLAN 105 I just made the VLAN number up let's let's actually make that a realistic number so let's make that VLAN number 251 so there we go okay so now the name of the port group I'm gonna select it makes more sense we want to pass VLAN 251 for these production virtual machines using our lag which is currently not set up so I've got the port group selected I'm gonna click manage then settings and then we click this edit right here if you're feeling lazy you can just go to the Actions menu and edit directly from there but I like showing different ways to get to it now we need to make some changes to the VLAN ID so I'll click on VLAN we're gonna change the none to VLAN 251 and we're gonna go to the teaming and failover and this is something brand new with vSphere 5.5 we can actually see the port group or the link aggregation group rather as an object it's an uplink in and of itself and it has kind of the universal symbol of two lines with a circle around it meaning that it's a link aggregation group so we're gonna drag that with the arrow we're gonna click on the arrow move it all the way up to the top and this little red square comes up and says hey you can't have a mixture here that's not supported you're gonna need to move the other up links to unused and that's fine because if you remember there's nothing actually bound to uplink one we've bound the physical network adapters to the Spongebob pork channel so I'm gonna drop down uplink to two unused and we're gonna drop down this other uplink to unused as well and it's gonna say hey I don't care what load balancing protocol you use here what algorithm use here we're gonna use whatever the algorithm that you chose in the lag for this port group so you don't even have to change this to the route by IP hash they used to use the warning is telling you that it's completely worthless and it doesn't really matter so we've got this configured to use the spongebob link aggregation group click OK we've also set the VLAN and at this point you're done you've actually got a port group here and with like summary and it's going to be here we go teaming and failover it's going to be using the active uplink called spongebob o1 it's not using those physical up links anymore it's not using those independent up links it's using the lag called spongebob o1 and we can see that it's using VLAN ID 251 it's right there so at this point you could put virtual machines on this porc group and actually pass traffic across that particular network of VLAN 251 don't miss out on my future videos become a youtube subscriber today do you crave more content on home labs technical certifications deep dives product reviews and geeky shenanigans wall network is also available in blog format at wall network.com

Channel: Wahl Network

Views: 98,011

Keywords: link aggregation, LAG, LACP, vsphere, esxi, vmware, networking, wahl network, esxi 5.5

Id: VRZJL5QDBCU

Length: 18min 14sec (1094 seconds)

Published: Fri Feb 21 2014