Multichassis Link Aggregation Groups Learning Byte

hello my name is Jodie Holmes and I'm a courseware developer with the Juniper Networks education services group in this learning byte we are going to be discussing multi chassis link aggregation groups or MC lags we will give you a brief overview of what an MC lag is and then show you how to configure and monitor an MC lag setup if you're not certain what a link aggregation group is we have to lag learning bytes available on the Juniper website one in English and one in Chinese so first off let's go ahead and get started with an MC lag overview multi chassis link aggregation enables a device to form a logical lag interface with two or more other devices the MC lag devices use Interjet control protocol or ICCP to exchange the control information between themselves the diagram on the slide shows a basic MC lag setup the ce-1 router is an MC lag client device that has two physical links configured as a standard lag group that is configuration on c e one is the same as with a standard lag bundle in fact the ce-1 device is not even aware of the MC lag happening between p1 and p2 p1 and p2 devices use ICCP to coordinate with each other to ensure that data traffic is forwarded properly this type of basic MC lag setup provides only node level redundancy on the provider side this slide shows a couple of examples of how MC lag can be implemented the upper example adds two more links to our previous example for some link level redundancy however as before there is still only node level redundancy on the provider side the lower example shows a much more thorough setup in this example we have introduced a second customer device ce2 along with dual links between each of the devices with this setup there is no single point of failure and as before the provider site setup is transparent to the customer side that is c1 and c2 are not aware they're lags are connected to two different devices there are two types of states that you can configure MC lag to operate in active standby and active active each state type has its own set of benefits and drawbacks the original MC lag mode is active standby and is supported on both dense port concentrator or DPC and modulator port concentrator NPC line cards active standby mode allows only one provider edge node to be active at a time using LACP the active PE device signals to the customer edge device that it's links are available to forward traffic as you might guess a drawback to this method is that only half of the links in the CES lag are used at any given time however this method is usually easier to troubleshoot than active active because traffic is not hashed across all links and no shared Mac learning needs to take place between the PE devices the active active mode is supported on MPC line cards only in an active active setup all links between us EE and P devices are active and available for forwarding traffic because of this traffic has the potential need to go between the PD of PE devices to egress so active active mode requires an inter-caste link to be configured between the two PE devices this link is used to share traffic between the PE devices this slide demonstrates the traffic flow that can be expected from both active standby and active active modes in the active standby model you can see that all traffic flows across the active link first and then continues on its way from the PE one device in the active active model all links are active and traffic can be forwarded across any of them many factors come into play when deciding which model to use and where and a full discussion about this is outside the scope of this learning byte alright let's get into configuring and monitoring MC lags okay so you can see our sample topology on the right we have three routers see e one PE one and PE two the mode we're going to configure first is the active standby mode and we'll start with the ce-1 router now if you recall the c1 router has no idea that he is connected to multiple chassis x' and therefore his configuration is just that of a standard lag and to do that just first of all you create the aggregated Ethernet interface set chassis aggregated devices Ethernet device count 1 and then you go ahead and add member links to that ae0 interface you just created so in this case we're going to set 1 1 0 gig either options 802 3 ad and specify our newly created a e 0 interface and we'll do the same for the gig e 1 1 1 member link and all we're going to do now is set up the a 0 interface with LACP active and configured as a simple trunk allowing VLAN 500 so set interfaces a 0 aggregated either options LACP active and then set interfaces a 0 unity a family bridge interface mode trunk VLAN ID list 500 and if we show that interface you can see the simple configuration we've done there go ahead and commit this and we'll see what our newly created a e0 interface is doing show interfaces terse 0 and it's currently admin up a link down which is to be expected the other side has not been configured yet ok we'll move on to the PE 1 router now the initial part of the setting up this on the PE 1 router is exactly the same so we're going to go ahead and create the aggregate a Ethernet interface first and add the member link to that in this case it's Giggy 1 1 0 now after you're creating that the first thing you want to do is create the ICCP portion and the first thing you want to do there is set up the service ID now the service ID is used by ICC p2 identify each instant and we'll match between PE 1 and P 2 that is set under this switch option service ID hierarchy we're going to specify 1 and now we can actually create the ICC P portion of this that's done in the Edit protocols ICCP hierarchy and the first thing you do is set your local address in this case we are 10.10 10.1 and you set up the peer 10.10 10.2 we give it a redundancy group ID list we'll specify one this this will come into play later on when we actually create the multi chassis portion of this we'll give it a lightness detection minimum interval value of 300 and a multiplier of 3 if we do a show there we can see the configuration we've just done we'll go ahead and commit this now to save some time I've already set up the PE 2 router on this so we should just be able to do a run show ICCP and you can see that we're established and up and up so that shows that ICCP is working properly another command that you can use to verify this is the run show BFD session detail command and you'll want to look for the client ICCP realm line in there which again is just another way to verify that ICCP is working properly ok now that we've verified ICCP is working properly with be2 we can concentrate on the AE 0 portion of this config so we'll go ahead and navigate there top at it interfaces AE 0 and we currently have nothing there first thing we're going to do is configure the LACP portion of this set aggregated ether options LACP active to other items under here LACP system ID and zero zero colon zero zero colon zero zero now this system ID must match between p1 and p2 and as well the admin key in this case we're going to use the value of 1 now the next bit of config for this will be under the set aggregated ether Options mcae portion we're going to give it an MCA e ID of 1 again this matches between PE 1 and PE to the next portion is the redundancy group this must match the configuration you set for ICCP next up is the chassis ID in this case we're going to set PE 1 is 0 P 2 is set with a chassis ID of 1 next will be the actual mode in this case since we're setting up back to standby mode we're going to set it as active standby obviously this matches between the two routers and the final thing is going to be the status control mode in this case we're going to make p1 as active and p2 is set as standby and finally we're going to set up this interface as a simple bridge with VLAN 500 on it so set unit 0 family bridge interface mode trunk VLAN ID list of 500 so if we take a look at our config you can see the LACP portions the mcae portions and the unit 0 we just configured we go ahead and commit that I'm going to enable it on PE 2 as well again this was just to save a bit of time go ahead and switch back to PE 1 do a run show interfaces terse a 0 we're now showing up up we cannot just do a run show interfaces mcae to look at that output you can see that our local status is active and our peer status is in standby just like we configured one more note take a look at the peer IP MCP state of na this will come into play once we switch this over to active active mode and finally you can do a run show LACP interfaces and you'll notice that in this case the PE one router has been chosen as the active node you can tell that from the muck state of collecting distributing if we switch over to p2 and we run this command run show la CP interfaces you'll notice that our muck State is waiting again this shows that p1 is the active router and just to confirm even further if we jump back over to the ce-1 router and we do this we'll see actually both states going so you can see that collecting distributing is the active node and the other interface is showing as attached now this information is sent from the PE routers to the seee device and again it has no idea that it's connected to two different routers so we've just configured active standby mode and now we'll move on to modifying this configuration for active active mode converting this to an active active setup is simple first we'll go back to the PE one router and we're going to change from mode active standby to mode active active and then the final thing we need to do is set up the Interjet link between the PE 1 and PE 2 routers you do that with a set multi chassis protection give the peer address followed by the outgoing interface and obviously on PE to the peer address is going to be 10.10 to attend dot 1 I've already set that up to save us a bit of time commit this and if we go back to the ce-1 router and we do a run show LACP interfaces you'll see that both its links now are in collecting distributing muck state which means both are available for forwarding well that concludes our learning byte on multi chassis link aggregation groups thank you very much have a nice day visit the Juniper education Services website to learn more about courses view our full range of classroom online and e-learning courses learning paths industry segments and technology specific training paths Juniper Networks certification program the ultimate demonstration of your competence and the training community from forums to social media join the discussion

Channel: JuniperNetworks

Views: 30,755

Rating: 4.8666668 out of 5

Keywords: Juniper, Juniper Networks, multichassis, LAG, active-standby, active-active, aggregation, link, learning byte

Id: 1byapa89L9w

Length: 14min 15sec (855 seconds)

Published: Tue Aug 13 2013