WEBVTT 00:00:00.960 --> 00:00:06.000 All right now for our last discussion of the day and it's going to be on EtherChannel bundling. 00:00:06.000 --> 00:00:11.120 How many of you are doing EtherChannel bundling? How many of you are doing EtherChannel bundling? 00:00:12.160 --> 00:00:16.959 Maybe you don't know what EtherChannel is. How about port channeling? How about channel grouping? 00:00:16.959 --> 00:00:20.959 How about NIC teaming? I'm just going to throw out those words. Okay we're talking about the 00:00:20.959 --> 00:00:26.719 same thing. EtherChannel is Cisco's terminology they use. It's their marketing term for port 00:00:26.719 --> 00:00:32.959 channeling, NIC teaming, port aggregation, all of that stuff. All right so when we look at this 00:00:32.959 --> 00:00:36.880 particular scenario we might ask ourselves, hey you know we're having some bandwidth issues. 00:00:36.880 --> 00:00:41.040 We're having some bandwidth issues here and here and here and here and here. So how can we increase 00:00:41.040 --> 00:00:45.040 the bandwidth between our switches? Well we can go ahead and we can upgrade to better switches 00:00:45.040 --> 00:00:49.360 with higher port speeds. Well good luck with that if you don't have the budget. Maybe you can upgrade 00:00:49.360 --> 00:00:52.720 just the ports on the switch if it's a modular switch. Again good luck with that if you don't 00:00:52.720 --> 00:00:56.639 have the budget. Last but not least you can just simply add more links between your switches. 00:00:57.360 --> 00:01:03.759 So you know that's pretty cheap. I can, if I have free ports, if I have some free ports on my device 00:01:03.759 --> 00:01:08.720 whether they're fiber ports or whether they're copper ports, doesn't matter right. If I have 00:01:08.720 --> 00:01:13.440 some free ports that we can utilize between our switches let's go ahead and add more switches. 00:01:14.559 --> 00:01:19.440 Add more links. That's all we're doing here. Okay but here's the problem. Here's the problem. 00:01:20.320 --> 00:01:26.320 The problem is Spanishy protocol. Spanishy protocol is going to rear its ugly head and say nope sorry 00:01:26.320 --> 00:01:30.160 I'm going to block all loops in this particular case. So even though we have eight links between 00:01:30.160 --> 00:01:35.440 all of our switches in this scenario we can only take this path, this path, and this path through 00:01:35.440 --> 00:01:40.559 this layer two environment. So this is not going to work for us right because we just wasted a 00:01:40.559 --> 00:01:44.480 whole bunch of links. We wasted a whole bunch of ports and we're not getting the value we want out 00:01:44.480 --> 00:01:48.400 of it. We want to increase our bandwidth right. We want to increase our bandwidth. So what do we have 00:01:48.400 --> 00:01:54.000 to do is we have to actually bundle these ports together. Basically take an elastic band and put 00:01:54.000 --> 00:02:00.000 an elastic band around them. That's what we're going to do. Bundle them together. Elasticize 00:02:00.000 --> 00:02:03.760 them together. Now I mean it's not just you physically putting an elastic band. You don't 00:02:03.760 --> 00:02:07.440 even have to put an elastic band around and you won't put an elastic band around them but I'm 00:02:07.440 --> 00:02:12.479 trying to give you a mental image of what you're actually doing here. We're aggregating those links 00:02:12.479 --> 00:02:16.960 together. So ether channel, link aggregation, port channeling, channel grouping, whatever you want to 00:02:16.960 --> 00:02:22.639 call it it's fine. Cisco's naming is ether channel bundling. Okay cost-effective scalable way for you 00:02:22.639 --> 00:02:28.960 to what? Get more bandwidth. Get more bandwidth. Fault tolerant meaning that if a link fails in 00:02:28.960 --> 00:02:32.880 the in there that's okay. The rest of them still keep working. Another one fails that's okay. They 00:02:32.880 --> 00:02:37.440 keep working whoever is left over in that particular bundle. Now we still have a problem with 00:02:37.440 --> 00:02:42.960 Spanning Sheet Protocol right because we do have a loop. This is a loop. This is a loop. This is an 00:02:42.960 --> 00:02:48.320 entire loop. So Spanning Sheet Protocol is still going to come along and block entire bundles. Hey 00:02:48.320 --> 00:02:52.880 this is still better though than blocking all the links except for three that we saw a little bit 00:02:52.880 --> 00:02:57.440 earlier. So this is still a lot better at least we've got eight here, eight here, eight here. So 00:02:57.440 --> 00:03:05.839 a total of 24 out of all the links that we have. So this is really good. All right so things to 00:03:05.839 --> 00:03:11.279 remember we have two or more physical links logically treated as one. Now for Cisco iOS 00:03:11.279 --> 00:03:17.360 more maxes out at eight sort of. It depends on the protocol you're using. So if you're using the 00:03:17.360 --> 00:03:23.360 protocol called PEG-P Port Aggregation Protocol which is Cisco proprietary then yes the max is 00:03:23.360 --> 00:03:29.600 eight. Eight in a bundle, eight max that you can use. But if you're using something like LACP you can 00:03:29.600 --> 00:03:40.800 have 16 in the bundle. Okay 16 in bundle but you can only use eight at a time okay. So two or more 00:03:40.800 --> 00:03:44.960 physical links logically treated as one but if using LACP you can have 16 but you can only use 00:03:44.960 --> 00:03:50.960 eight at a time. Meaning the other eight would be failovers, backups. So if all of a sudden one of the 00:03:50.960 --> 00:03:55.199 eight that are being used fails then one of the other eight that are remaining that aren't being 00:03:55.199 --> 00:04:01.360 used can be used. So you can have 8, 9, 10, 11, 12, 13, 14, 15, 16 it's up to you with LACP but PEG-P 00:04:01.360 --> 00:04:09.440 maxes out at eight. We get more bandwidth. Up here there's one logical port the Spanning Sheet Protocol 00:04:09.440 --> 00:04:15.600 that's why we can block that entire bundle or why we can allow all that traffic on that link. 00:04:15.600 --> 00:04:23.600 We can do this on layer two interfaces or layer three interfaces. So we can do it on routed ports 00:04:23.600 --> 00:04:29.760 or we can do it on access ports. We can do it on trunk ports and there's a variety of ways to load 00:04:29.760 --> 00:04:38.239 share as well. One also thing to note I want you to know about ether channel for IOS is that you 00:04:38.239 --> 00:04:48.799 can only create your bundles between two devices. So meaning that I cannot do a bundle like this 00:04:50.559 --> 00:05:00.000 unless we're doing what stack-wise. So if we take advantage of stacking then I can do that 00:05:00.000 --> 00:05:04.880 because remember every switch in a stack is connected together 00:05:08.480 --> 00:05:14.239 with a backplane cable called the stacking cable and they're treated as one logical unit 00:05:14.880 --> 00:05:20.399 almost like as if they were a single chassis based switch when it's all said and done. 00:05:20.399 --> 00:05:25.519 So instead of having a backplane and then all your modules are plugging into it you have individual 00:05:25.519 --> 00:05:30.399 switches that you connect to cable with specialized cables. The cables are the same as the 00:05:30.399 --> 00:05:35.679 backplane that would be inside of a chassis based switch. So essentially it's a way to do a chassis 00:05:35.679 --> 00:05:41.839 based type of switch without doing a chassis based type of switch stacking. So in this case because 00:05:41.839 --> 00:05:48.399 they're now treated as one logical entity then I could do bundles from those different devices. 00:05:48.959 --> 00:05:53.839 Okay same thing if we're doing VSS virtual switching service up at the distribution layer 00:05:53.839 --> 00:05:59.040 then yes we could do this as well. All right so that's something to keep in mind when we're talking 00:05:59.040 --> 00:06:04.480 about the bundling. So for the exam though right most of your ether channel examples are not going 00:06:04.480 --> 00:06:08.880 to be doing any type of stacking whatsoever. So it's important for you to remember that if they 00:06:08.880 --> 00:06:15.040 try to make a bundle like this that is wrong that is the problem that is why it's not working. 00:06:15.920 --> 00:06:23.119 Keep that in mind. All right so the protocols PEG-P LACP versus none what does that mean? Well 00:06:23.119 --> 00:06:28.239 PEG-P port aggregation protocol that's a cisco proprietary protocol as we said supports up to 00:06:28.239 --> 00:06:36.640 eight links only and whereas LACP is an industry standard protocol 802.3 ad 00:06:37.760 --> 00:06:44.000 right we can have eight links in use but up to a total of 16 total in a bundle for failover 00:06:44.320 --> 00:06:51.679 for failover sakes. So for the exam any scenario that talks about two cisco switches so if this 00:06:51.679 --> 00:06:56.239 is cisco and this is cisco you can use either or you can use either or it's up to you. 00:06:57.519 --> 00:07:03.519 But if the example has a cisco switch and a non-cisco switch then you have no choice but to 00:07:03.519 --> 00:07:09.920 use LACP keep that in mind. Right so cisco to cisco you can use either or but if it's cisco to a 00:07:09.920 --> 00:07:15.119 non-cisco you got to be using LACP and more than likely multiple spanning tree protocol as well. 00:07:16.480 --> 00:07:19.920 None means you're connecting the devices together without using a protocol 00:07:21.679 --> 00:07:25.920 so this would be really good in any type of scenario where for example if you had your switch 00:07:26.640 --> 00:07:32.399 and you were connecting to something like a server and you were doing nick teaming on your server 00:07:34.320 --> 00:07:39.679 and the server if it didn't support LACP then you could do none on this side over here 00:07:39.679 --> 00:07:42.959 meaning you're going to set up your bundle without relying on any type of protocol. 00:07:43.839 --> 00:07:46.880 But if the service supports LACP then you might as well go ahead and use LACP. 00:07:48.480 --> 00:07:54.399 Another example would be you have a switch connected up to a wireless LAN controller 00:07:54.959 --> 00:07:59.440 typically your wireless controllers do not support PEGP or LACP so you're doing your 00:07:59.440 --> 00:08:05.279 configuration here so therefore you'd have to have none over here. So just examples if the other 00:08:05.279 --> 00:08:10.799 device doesn't support a protocol then you would do none on this side no protocol on this side and 00:08:10.799 --> 00:08:16.559 the keyword we'll see later is actually on that's the keyword we'll use whereas PEGP and LACP have 00:08:16.559 --> 00:08:25.440 other keywords. Yeah so everything I'm talking about here is iOS because this is exam is on iOS 00:08:25.440 --> 00:08:30.399 okay so if you're if you're familiar with using like Nexus products and you have your VPCs with 00:08:30.399 --> 00:08:35.359 Nexus products or you're using more links with those Nexus products or anything like that then 00:08:35.359 --> 00:08:40.799 that that is not what this discussion is iOS this entire exam is testing you on your knowledge of 00:08:40.799 --> 00:08:49.599 Cisco iOS and iOS XE and iOS XE. All right now if there are new products out there that you've 00:08:49.599 --> 00:08:54.159 received recently that the numbers are different for them now and they allow you to do more they 00:08:54.159 --> 00:08:58.320 allow you to do different things again when this exam was updated and created right they're not 00:08:58.320 --> 00:09:04.239 testing you on that they're still testing on what I'm teaching you now and that is what the 00:09:04.239 --> 00:09:09.760 fundamentals the total fundamentals of the majority of products that are out there running Cisco iOS 00:09:09.760 --> 00:09:20.239 and iOS XE. All right so either channel we're going to focus on our links between DSW-1 and 00:09:20.239 --> 00:09:26.799 DSW-2 today. I'll show you some configurations some examples right but the first thing I always 00:09:26.799 --> 00:09:30.960 want to tell people about this and this is so important when you're doing an EtherChannel bundle 00:09:30.960 --> 00:09:41.440 and you are bundling these interfaces together that device expects every single one of those 00:09:41.440 --> 00:09:52.239 interfaces to be the exact same. All those four ports in this scenario have to be the same speed 00:09:52.239 --> 00:09:58.799 the same duplex if they're access ports they all have to be in the same VLAN okay if they're trunk 00:09:58.799 --> 00:10:03.599 ports they all have to have the same native VLAN if they're trunk ports they all have to have the 00:10:03.599 --> 00:10:12.479 exact same VLAN allowed list and if they're not then you're gonna have errors so you're gonna 00:10:12.479 --> 00:10:16.719 have to solve those problems so on the exam whenever you're troubleshooting EtherChannel 00:10:16.719 --> 00:10:22.799 bundles look for these things look for any of these things in the scenario in the situation 00:10:22.799 --> 00:10:27.039 because if any of them are out of alignment and they're not the same on all those physical ports 00:10:28.000 --> 00:10:34.799 that's your problem solve that problem meaning oh multiple choice question what does it say 00:10:34.799 --> 00:10:41.760 it says option C says to adjust the VLAN allowed list to include duh duh duh duh and duh great 00:10:41.760 --> 00:10:45.520 that's my answer because based on what we looked at that was the thing that was wrong 00:10:48.559 --> 00:10:53.119 again because it's iOS as we talked about here's a note just what we talked about two switches per 00:10:53.119 --> 00:10:59.599 bundle that's it so this switch going to this switch I could not do three switches like a VPC 00:10:59.599 --> 00:11:07.440 or anything like that or if we were doing stack wise where I could have more switches in the stack 00:11:07.440 --> 00:11:13.919 going to different switches that's not covered here that's not part of this exam this exam is 00:11:13.919 --> 00:11:21.599 hey iOS fundamentals two switches therefore all these ports must be going to one other device 00:11:21.599 --> 00:11:27.359 can't be two other devices in this scenario so again if all of a sudden you see a bundle 00:11:27.359 --> 00:11:30.960 like that on the exam right that's not right in this case 00:11:37.200 --> 00:11:42.719 so anytime you have issues it's pretty much usually related to misconfigurations 00:11:44.239 --> 00:11:48.400 misconfigurations so what has to match versus what doesn't match that's what you're looking 00:11:48.400 --> 00:11:52.799 for on the certification exam if it's if it's a troubleshooting scenario for ether channel 00:11:53.760 --> 00:11:59.919 all right peg p keywords are desirable in auto so if the scenario is about 00:12:02.000 --> 00:12:07.280 cisco devices connecting the cisco devices then you might have the opportunity to use peg p on 00:12:07.280 --> 00:12:13.599 the exam if they throw the term desirable out there that's peg p auto that's peg p 00:12:15.119 --> 00:12:19.599 right so desirable initiates the ether channel session auto waits for neighboring switch sounds 00:12:19.599 --> 00:12:26.799 a lot like dtp yes same idea so that means desirable works with desirable desirable works 00:12:26.799 --> 00:12:33.520 with auto but auto auto doesn't work together so again looking at it looking at a scenario on the 00:12:33.520 --> 00:12:38.719 exam if one side is desirable the other side is desirable that's not the problem one side's auto 00:12:38.719 --> 00:12:42.799 the other side's desirable that's not the problem both sides are auto there's your problem 00:12:43.520 --> 00:12:47.919 i'll show you how to set that up in the command momentarily lacp your keywords are active and 00:12:47.919 --> 00:12:53.200 passive so again if it's an lacp scenario so it could be two cisco devices or a cisco device in a 00:12:53.200 --> 00:12:59.359 non-cisco device lacp what are you going to do well if passive passive are on both sides not 00:12:59.359 --> 00:13:04.239 going to work but if it's passive and active it's going to work if it's active and active it's going 00:13:04.239 --> 00:13:08.080 to work why because active says initiate the ether channel session passive says wait 00:13:08.640 --> 00:13:16.000 so very straightforward it's about memorizing this and always remember that no protocol so 00:13:16.000 --> 00:13:20.239 if you don't use a protocol on one side and use the keyword of on which we'll talk about here 00:13:20.239 --> 00:13:26.640 momentarily on then the other side would have to be on it can't be lacp it can't be peggy so 00:13:26.640 --> 00:13:30.000 this side is on you can't have active here you can't have passive here you can't have 00:13:30.000 --> 00:13:33.440 desirable here you can't have auto here why because you can't have active here you can't 00:13:33.440 --> 00:13:36.799 you can't have active here you can't have passive here you can't have desirable here 00:13:36.799 --> 00:13:41.280 you can't have auto here why this is on the other side has to be on in that particular case 00:13:42.880 --> 00:13:45.280 all right time to do some config 00:13:46.159 --> 00:13:49.119 so 00:13:58.799 --> 00:14:04.799 perfect just reading my notes in regards to uh things we've talked about so let's go to dsw1 00:14:04.799 --> 00:14:09.359 right now and the first thing i want to do is i want to factory default reset my interfaces 00:14:09.359 --> 00:14:14.479 because we've done a lot on these four interfaces and we could definitely have some misconfigurations 00:14:14.479 --> 00:14:20.960 right now so it's always better to get them to exactly a state where you know they're all the 00:14:20.960 --> 00:14:27.440 same and in that way there you can build your configuration on top of that so default interface 00:14:28.000 --> 00:14:38.799 range gigabit ethernet 1 0 17 through 20 good and then we'll do the same thing over on dsw2 00:14:44.559 --> 00:14:50.640 all right and then over here on dsw1 i'm going to go interface range 17 through 20 00:14:51.520 --> 00:14:54.320 and i'm going to shut down all four of those interfaces why because 00:14:54.960 --> 00:14:57.679 either channel bundling is very very very 00:15:01.039 --> 00:15:09.919 finicky very dependent on order of operations therefore if you are leaving your interfaces up 00:15:09.919 --> 00:15:15.119 while you're doing it and you are making changes it's going to cause a lot of changes a lot of 00:15:15.679 --> 00:15:20.400 changes a lot of flapping a lot of of adjustments it's going to cause spanning tree protocol to 00:15:20.400 --> 00:15:24.239 go crazy and do a lot of topology change notifications and a lot of different things 00:15:24.880 --> 00:15:29.599 so the idea here is to you know bring it down the interfaces that you're going to be adjusting make 00:15:29.599 --> 00:15:34.640 all the changes and then bring them back up so that way there you're only doing a convergence 00:15:34.640 --> 00:15:37.760 twice once on the way down once on the way back up instead of like four or five times as you're 00:15:37.760 --> 00:15:45.679 doing your configuration also the differences between configuring one side versus another side 00:15:45.679 --> 00:15:52.640 could ultimately affect to your your um your negotiation and as a result of that just based 00:15:52.640 --> 00:15:56.960 on timing you could end up in a situation where you're gonna have to bounce all your interfaces 00:15:56.960 --> 00:16:01.599 anyways to get it to work which you don't want to have to do so it's better to in my opinion 00:16:01.599 --> 00:16:07.200 shut it down do your config bring them back up and then let the magic happen and then everything 00:16:07.200 --> 00:16:14.400 should be fine for you so what are we going to do here we are going to begin we are so far in 00:16:14.400 --> 00:16:18.799 gigabit one slash zero slash 17 through 20 we're going to start using the command of channel group 00:16:20.000 --> 00:16:29.359 and this is where you specify the number the number that will indicate which interfaces are 00:16:29.359 --> 00:16:36.320 part of which group so for example if i do channel group uh jeffrey says five jeffrey said five okay 00:16:36.320 --> 00:16:40.960 so if i create channel group five since we're configuring all four interfaces at the same time 00:16:40.960 --> 00:16:44.719 this is going to be in channel group five channel group five channel group five channel which means 00:16:44.719 --> 00:16:48.960 they're all part of the same group so what does that mean it means i can create another group 00:16:48.960 --> 00:16:53.919 here so we did another bundle here i could do group seven group seven i could do group six group 00:16:53.919 --> 00:17:00.880 six right so i have the number that you specify here indicates which interfaces go together in 00:17:00.880 --> 00:17:07.599 the bundle so gigabit 17 is part of group 5 18 is part of group 5 19 is part of group 5 20 is part 00:17:07.599 --> 00:17:11.119 of group 5 because i'm configuring them all at the same time with the interface range command 00:17:11.119 --> 00:17:15.359 if i did them all individually then i whatever command i'm about to type here i have to copy 00:17:15.359 --> 00:17:24.160 and paste to all those interfaces now it says mode so we'll type in mode and then next here's 00:17:24.160 --> 00:17:29.359 those keywords active auto desirable honor passive so the on means no protocol so pay attention to 00:17:29.359 --> 00:17:34.079 this on the exam right if they're asking you to do a config or troubleshoot or verify or anything 00:17:34.079 --> 00:17:38.319 of a config whether it's multiple choice drag and drop simulation doesn't matter fact is pay 00:17:38.319 --> 00:17:42.640 attention to the details of the scenario because if they turn around say they want you to use an 00:17:42.640 --> 00:17:46.400 industry standard protocol that means active or passive are your options they say they want you 00:17:46.400 --> 00:17:50.560 to the cisco proprietary protocol the desirable and auto are your options they say they don't 00:17:50.560 --> 00:17:55.199 want you to use a protocol then you on is your only option right but then also pay a little 00:17:55.199 --> 00:18:00.959 bit more attention to what if they say on dsw1 you must configure it with an industry standard 00:18:00.959 --> 00:18:06.400 protocol but you do not want it initiating the connection then you type in passive 00:18:08.640 --> 00:18:11.920 right what about a scenario where they turn around and they say something like you want 00:18:11.920 --> 00:18:17.520 to use a cisco proprietary protocol but you don't want it initiating the connection right you see 00:18:17.520 --> 00:18:21.599 just those little key words tell you exactly what you must be using in that particular case 00:18:21.839 --> 00:18:26.239 right so now that i've used auto on this side i have no choice but to use desirable on the other 00:18:26.239 --> 00:18:33.280 side notice what happened as soon as i typed this in notice how it created something called a port 00:18:33.280 --> 00:18:37.359 channel okay what is a port channel a port channel is now the virtual interface 00:18:37.359 --> 00:18:43.680 that is going to be used for this bundle so 17 18 19 and 20 are bundled together 00:18:43.680 --> 00:18:50.400 and they are now represented as port channel number five port channel number five 00:18:51.680 --> 00:18:58.079 so if i type in do show ip interface brief and we go all the way to the bottom 00:18:59.359 --> 00:19:04.560 you'll see that port channel 5 is now in my list of virtual interfaces that i have on my device 00:19:04.560 --> 00:19:08.239 right now it's down down because all the physical interfaces are down right now 00:19:08.239 --> 00:19:15.040 that's why it's down right but now port channel 5 is what type of interface it's a layer 2 interface 00:19:15.040 --> 00:19:20.640 why is it a layer 2 interface because my interfaces were layer 2 by default so since 00:19:20.640 --> 00:19:25.839 they're layer 2 interfaces the port channel is automatically a layer 2 port channel so what does 00:19:25.839 --> 00:19:35.359 that mean it means i can configure what oh boo i can configure vlans i can configure trunks i can 00:19:35.359 --> 00:19:42.239 configure whatever i need on that interface from a layer 2 perspective i can't configure an ip 00:19:42.239 --> 00:19:49.040 address i can't configure dynamic routing protocol why it's layer 2 so layer 2 port channels are 00:19:49.040 --> 00:19:55.839 going to allow me to do layer 2 things so let's go ahead right now and type in interface port 00:19:55.839 --> 00:20:05.920 channel 5 as an example and we'll type in switch port trunk encapsulation dot one q 00:20:05.920 --> 00:20:14.959 switch port mode trunk switch port no negotiate switch port trunk allowed 00:20:14.959 --> 00:20:24.640 vlans 10 20 50 200 switch port trunk native vlan 99 00:20:29.760 --> 00:20:34.000 oh this device doesn't like that cool all right but that's just a prime example of showing you 00:20:34.000 --> 00:20:37.199 that do show run interface 00:20:39.839 --> 00:20:44.640 port channel 5 now 5 right there's my configuration 00:20:49.520 --> 00:20:53.520 now what about the physical interfaces what i want you to realize is that the physical interfaces 00:20:53.520 --> 00:21:00.000 inherit the port channel configuration so we type in the same command but instead of port channel 5 00:21:00.000 --> 00:21:07.760 let's say gigabit ethernet 1.0 17 look at that it inherited the configuration that i typed 00:21:08.319 --> 00:21:12.719 so anything you type on the port channel goes to the physical interface so that's an important 00:21:12.719 --> 00:21:16.319 thing to realize too for the exam is that you don't want to start messing around with the 00:21:16.319 --> 00:21:20.319 physical interfaces once you have a port channel just mess around with the port channel anything 00:21:20.319 --> 00:21:28.880 you do to the port channel will affect the physical interfaces you know often i get a 00:21:28.880 --> 00:21:33.839 question well what if i wanted to change the speed of one of the interfaces you're not allowed 00:21:33.839 --> 00:21:38.880 because if you change the speed of one interface the other interfaces no longer match right 00:21:40.079 --> 00:21:45.040 so you have to change the speed of all the interfaces yes so you might as well just 00:21:45.040 --> 00:21:51.280 change the speed on the port channel which affects all the interfaces so you might as 00:21:51.280 --> 00:21:57.680 well just configure the port channel going forward so that way there you affect all the interfaces 00:21:57.680 --> 00:22:02.000 all right so there you go there's a prime example of how we would configure this 00:22:02.000 --> 00:22:10.479 so now let's go to dsw2 and on dsw2 interface range gigabit ethernet 1.0 17 through 20 okay 00:22:10.479 --> 00:22:14.560 and now what are we going to type in we're going to type in channel group now the channel group 00:22:14.560 --> 00:22:20.479 number doesn't have to match meaning that this number does not have to match the other side 00:22:20.479 --> 00:22:26.560 so i could use something like your 10 10 10 10 if i wanted to i don't have to do the same number 00:22:26.560 --> 00:22:32.560 but i like to do the same number why because then i know when i am troubleshooting that anything 00:22:32.560 --> 00:22:39.920 that is port channel 5 here is port channel 5 on the other side so i know that 5 goes to 5 6 goes 00:22:39.920 --> 00:22:47.040 to 6 7 goes to 7 8 that's how i like to do it but you don't have to do it that way now our mode here 00:22:47.040 --> 00:22:53.119 in this particular case is going to be what is going to be it has to be desirable why because 00:22:53.119 --> 00:22:56.479 the other side was auto so another thing to pay attention to on the exam is if you're doing any 00:22:56.479 --> 00:23:00.000 type of configuration scenario and one side's already configured and you're trying to do the 00:23:00.000 --> 00:23:04.640 other side well look at the other side and find out what did they do what did they do in their 00:23:04.640 --> 00:23:09.359 particular case well they used auto over here on those physical interfaces okay well then i have 00:23:09.359 --> 00:23:17.599 no choice but to use desirable on the other side rejected why why are you rejected what did i do 00:23:17.599 --> 00:23:24.079 wrong 00:23:26.560 --> 00:23:30.479 oh geez anybody see my mistake 00:23:34.880 --> 00:23:38.079 oops oops 00:23:41.599 --> 00:23:45.359 that's bad i hope that didn't mess everything up for today 00:23:45.359 --> 00:23:47.359 oh i gotta check my work 00:23:49.359 --> 00:23:55.199 ouch gotta double check before you hit enter raymond let's hope that didn't mess up the rest 00:23:55.199 --> 00:23:59.439 of the topology that's not because i'll be fixing stuff before tomorrow's class then 00:24:00.319 --> 00:24:01.839 all right so let's do that now 00:24:02.079 --> 00:24:04.319 um 00:24:10.719 --> 00:24:15.439 okay so what's the best thing to do what's the best thing to do what's the best thing to do 00:24:16.319 --> 00:24:20.880 okay let's go back to this range and let's do this 00:24:24.719 --> 00:24:27.599 okay now that i've done that 00:24:28.079 --> 00:24:32.079 let's go back to this and try 00:24:34.959 --> 00:24:42.160 all right got it all right so we're good there so all i do is negate it from all of those interfaces 00:24:42.160 --> 00:24:45.760 and then went back and just did the interfaces i was supposed to i hope i got away with it we'll 00:24:45.760 --> 00:24:52.560 see what happens this might be the worst demo ever for ether channel by raymond all right so now 00:24:52.959 --> 00:24:59.839 for ether channel by raymond all right so now the port channel the port channel 00:25:00.800 --> 00:25:03.359 okay so it was already created as you can see up here 00:25:05.359 --> 00:25:10.719 so it didn't recreate it so it already exists so we can type an interface port channel five 00:25:11.599 --> 00:25:16.400 and now we need the same config that we have over here so let's just go to this port channel 00:25:17.040 --> 00:25:22.079 and we'll copy and paste it my fingers are so crossed we are not toast on this demo 00:25:22.880 --> 00:25:27.760 what happens you just got to default it and start all over all right so it looks like we're okay 00:25:28.560 --> 00:25:31.040 so now that i have it configured the same way on both sides 00:25:33.040 --> 00:25:43.280 do show run interface port channel five that looks good do show run interface gigabit 1 0 17 00:25:43.920 --> 00:25:57.680 good 18 good 19 good 20 good okay what else was included in there seven seven's fine eight's fine 00:25:58.800 --> 00:26:04.000 good all right so excellent i think we're okay all right so now let's just bring up all the 00:26:04.000 --> 00:26:09.439 physical interface so interface range gigabit ethernet 1 size 0 size 17 through 20 let's 00:26:09.439 --> 00:26:15.280 confirm that's the right range yep no shut there we go fingers crossed basically what i'm looking 00:26:15.280 --> 00:26:21.199 for is port channel 5 up up when it's all said and done is port channel 5 up up when it's all 00:26:21.199 --> 00:26:29.439 said and done so we got port channel 5 up we've got port channel 5 up yes we got them both all 00:26:29.439 --> 00:26:38.959 right good so we should be a okay so show ip interface brief let's scroll all the way down 00:26:38.959 --> 00:26:48.719 to the bottom and we'll see port channel 5 is up up there it is okay let's take a peek now at 00:26:48.719 --> 00:26:56.880 something like show interface port channel 5 and you'll see the bandwidth here look at that it is 00:26:56.880 --> 00:27:03.359 what it is four gigabits per second theoretically all right it's not a perfect four gigabits per 00:27:03.359 --> 00:27:09.119 second because it's still for one gig link so you're only getting the performance of one gig 00:27:09.119 --> 00:27:16.000 links but you're getting that performance across four links at that time and you'll see that it 00:27:16.000 --> 00:27:21.760 shows you now which members are part of that ether channel bundle all right what else do we want to 00:27:21.760 --> 00:27:27.119 take a peek at oh let's take a look at show interfaces trunk now so we can see what this 00:27:27.119 --> 00:27:33.839 output might look at look like now and we'll see here that the port channel is showing up instead 00:27:33.839 --> 00:27:40.880 of the physical interfaces and our vlan allowed list as we created is showing up there right this 00:27:40.880 --> 00:27:47.439 is beautiful all right what about spanning tree protocol then show spanning tree mst1 00:27:48.319 --> 00:27:55.280 look it's port channel 5 that is showing up and look at the priority it gave it 544 00:27:55.359 --> 00:28:01.119 that's awesome 544 but look look at the cost remember when we were talking about long versus 00:28:01.119 --> 00:28:13.359 short i said a gig is a thousand i'm sorry gig is 20 000 so uh two gig would be 10 000 and four gig 00:28:13.359 --> 00:28:18.719 would be 5 000 and eight gig would be 2500 well there you go there's 5 000 for four gigs 00:28:19.599 --> 00:28:21.119 for spanning tree protocol cost 00:28:28.160 --> 00:28:35.680 all right so if you notice here then we're on dsw1 that entire port port channel 5 is now in 00:28:35.680 --> 00:28:41.680 the forwarding state so all four of those links are in the forwarding state we go to dsw2 do show 00:28:41.680 --> 00:28:50.479 span for mst1 we should see that it's the root port the entire port channel yep so the entire 00:28:50.479 --> 00:28:56.239 port channel so as before remember we were blocking three out of those four ports now 00:28:56.239 --> 00:29:04.239 all of them are the root port in this example all right and let's take a look at show ether 00:29:05.199 --> 00:29:10.479 summary so when we look at show ether channel summary we'll see here that we have 00:29:13.520 --> 00:29:19.439 the group number five so any interface that is part of group five is going to show up here and 00:29:19.439 --> 00:29:25.520 there they are 17 18 19 and 20 so if we add another interface to the group and we type in 00:29:25.520 --> 00:29:31.599 channel group five on that interface and we'd have to use the same mode so mode auto then it would 00:29:31.599 --> 00:29:35.040 show up in this list because we typed it as channel group five and if we had other 00:29:35.760 --> 00:29:39.040 channel groups here if we had other port channels they would show up in this output as well 00:29:39.040 --> 00:29:43.680 so for example we're on dsw2 right now if we had a port channel here and a port channel here we'd 00:29:43.680 --> 00:29:48.719 have three port channels and they would all show up here with the interfaces in them we look here 00:29:48.719 --> 00:29:57.839 at the port channel it says p05 or p yeah p05 uh su so su stands for what capital s means layer two 00:29:57.839 --> 00:30:04.000 which we've talked about already as we had this discussion and in use so this is good su means it 00:30:04.000 --> 00:30:11.280 is perfect it's working it's operational as layer two peg p if we were using lacp it's a lacp if 00:30:11.280 --> 00:30:16.800 it's an on not on it wouldn't say on typically it shows up as a dash so if you saw a dash especially 00:30:16.800 --> 00:30:25.760 on the exam the dash here would mean that you used the keyword of on and then these p's are good p's 00:30:25.760 --> 00:30:31.920 mean bundled in the port channel you want to see p's now here's the deal if you see anything else 00:30:31.920 --> 00:30:39.920 in this output on the certification exam so if you see d over here or you see eyes over here or you 00:30:39.920 --> 00:30:46.880 see s's over here or you see f's or m's or da da da da what i want you to realize is that there 00:30:46.880 --> 00:30:51.280 is a problem and you have to figure out what that problem is what does that mean you got to compare 00:30:51.280 --> 00:30:56.560 and contrast your configuration it is a configuration problem right remember we said speed 00:30:56.560 --> 00:31:00.640 master match duplex has to match for the access ports they all have to be in the same vlan verify 00:31:00.640 --> 00:31:05.520 that stuff look at that stuff what else we said if it's a trunk port the native vlan has to match 00:31:05.520 --> 00:31:09.839 we said the vlan allowed list has to match so if any of those things are not what they're supposed 00:31:09.839 --> 00:31:14.640 to be that is why you have a problem so you may be troubleshooting ether channel on the exam but 00:31:14.640 --> 00:31:18.959 in the end what are you really troubleshooting you're troubleshooting vlan allowed list you're 00:31:18.959 --> 00:31:25.119 troubleshooting um uh speed you're troubleshooting duplex you're troubleshooting you know things that 00:31:25.119 --> 00:31:29.680 are not directly related to ether channel they're related to other technologies but they have an 00:31:29.680 --> 00:31:35.920 effect on ether channel so look at compare and contrast the configuration and that is 00:31:35.920 --> 00:31:39.040 ultimately going to solve your problem whatever it happens to be and answer your question 00:31:42.000 --> 00:31:44.959 now if you see h h is not bad h is actually good 00:31:45.520 --> 00:31:51.760 so if you see an h it typically means that you well not typically it means you are running lacp 00:31:51.760 --> 00:31:56.800 first and foremost so notice we're running peg p so we would never see an h with peg p we would 00:31:56.800 --> 00:32:02.719 never see an h if there was dashes there but we would definitely see an h if we're using lacp 00:32:02.719 --> 00:32:09.280 and there are more than eight interfaces so the ninth interface would have an h the 10th interface 00:32:09.280 --> 00:32:14.479 would have an h the 11th interface would have an h remember you can have up to 16 interfaces 00:32:14.479 --> 00:32:19.839 but only eight are being used so the eight that are being used will have p's and the rest of them 00:32:19.839 --> 00:32:26.319 will have h's as hot standbys and then if one of the p's fail then one of the h's get promoted to 00:32:26.319 --> 00:32:34.640 a p and then that's the one that's being used so that's a benefit of lacp all right all right so 00:32:34.640 --> 00:32:39.280 there's our live configuration i know we are at the end of today's session but there's still a 00:32:39.280 --> 00:32:42.319 little bit more we have to cover with either channel and then we'll be done for the day 00:32:42.319 --> 00:32:47.439 right so my slides show you the configs basically as we've done them here folks and then this output 00:32:47.439 --> 00:32:53.040 here as well so nothing new but there if it's lacp it shows up as that but if it's not lacp it's on 00:32:53.040 --> 00:32:58.079 it would show up like that instead so just keep that in mind for the certification exam right 00:32:58.079 --> 00:33:02.479 we've already seen this output as well so it's about the bandwidth we never looked at this output 00:33:02.479 --> 00:33:06.959 but here's just another way show ether channel port channel it'll show all your different groups so in 00:33:06.959 --> 00:33:13.119 this case i created group 12 in this example so for port number 12 what do you have well you need 00:33:13.119 --> 00:33:18.719 logical slot and port numbers not exam relevant but you know number of ports that are in it what 00:33:18.719 --> 00:33:25.040 protocol did you use are you active and so on and so forth which ports are part of the bundle so you 00:33:25.040 --> 00:33:29.839 could get asked a simple question about something like this maybe maybe not probably not but just 00:33:29.839 --> 00:33:34.239 something to show you that you know being able to read these things uh based on the fundamentals 00:33:34.239 --> 00:33:38.479 that we've learned is the most important skill to have something like this i'll show you here 00:33:39.280 --> 00:33:45.199 be careful with the output they show you because if you look at this command it's talking about 00:33:45.199 --> 00:33:50.880 the neighbor so what you need to do is you need to look here this is dsw1 we're on dsw1 right now 00:33:51.920 --> 00:33:58.800 we're on dsw1 that's who we're on okay so if the neighbor neighbor means this device here 00:33:58.800 --> 00:34:05.040 so all of this is referring to this device down here even though we're on dsw1 pay attention to 00:34:05.040 --> 00:34:10.399 that sometimes they do these types of questions just to see are you paying attention it has 00:34:10.399 --> 00:34:15.600 nothing to do with your knowledge and everything to do with more can you read a question so in the 00:34:15.600 --> 00:34:23.600 end they might ask you um which of the following ports are participating in lacp whatever i mean 00:34:23.600 --> 00:34:30.800 that's pretty straightforward there's the ports on dsw2 okay dsw2 right or maybe you're looking at 00:34:30.800 --> 00:34:37.199 the answer options and they're asking you about um specific things you see here but some of the 00:34:37.199 --> 00:34:42.959 answer options are referring to dsw1 dsw1 dsw1 when in actuality if you pick that option even 00:34:42.959 --> 00:34:48.879 though the number is right they use dsw1 that means it's wrong you got to look for the one that's on 00:34:48.879 --> 00:34:56.320 dsw2 so just be paying attention to that on this exam right these flags here this row right here 00:34:56.320 --> 00:35:02.959 you see the a a means active so what does that mean it means dsw2 was configured with the keyword 00:35:02.959 --> 00:35:10.159 of active that's what that means so what does that mean dsw1 was configured with well we don't know 00:35:10.159 --> 00:35:16.159 by looking at this output but it may be active or it could be passive right so maybe the question 00:35:16.159 --> 00:35:21.919 is testing you on something like that right what what could dsw1 had been configured with 00:35:22.560 --> 00:35:27.919 so when you look at this output here like oh okay that's the active flag right the other option would 00:35:27.919 --> 00:35:33.360 be passive because it's lacp so if this is active then it could be this or that so then you're 00:35:33.360 --> 00:35:38.719 looking for an answer option that says active or passive but if these were all p's as an example so 00:35:38.800 --> 00:35:46.159 erase these and put p p p p right then you know that oh this is the neighbor of dsw1 00:35:46.159 --> 00:35:51.360 so this would have been passive which means dsw1 without a doubt has to be active in this case 00:35:54.239 --> 00:35:59.600 so priorities um and device id so that would be your base ethernet mac address that's what that 00:35:59.600 --> 00:36:06.159 would be and your priority this would be for um prioritizing interfaces so you have the ability 00:36:06.159 --> 00:36:11.600 to prioritize interfaces right meaning maybe have an interface be used more than another interface 00:36:11.600 --> 00:36:17.679 or when you have more than eight interfaces especially with lacp how do you ensure that 00:36:17.679 --> 00:36:22.159 certain interfaces are being used under normal conditions and if and only if there's a failure 00:36:22.159 --> 00:36:26.800 do you want the other ones to be used then you can mess around with the priority you can mess 00:36:26.800 --> 00:36:30.879 around with the priority meaning that if you give lower priorities to certain interfaces they're 00:36:30.879 --> 00:36:36.239 going to be used uh before the other interfaces are used and those other ones will be h interfaces 00:36:36.239 --> 00:36:42.879 as opposed to p interfaces in that other output that we saw right that's that's pretty much it 00:36:42.879 --> 00:36:47.360 for this example output just to you know teach you how to pay attention and and make sure that 00:36:47.360 --> 00:36:57.760 you can read certain things then we go to dsw2 and run the lacp to c okay so that command also 00:36:57.760 --> 00:37:04.320 works with pegp so that command does work with pegp um so i'm not going to switch everything to 00:37:04.320 --> 00:37:09.120 lacp on this device i really can't do that right now without having to kill everything 00:37:09.120 --> 00:37:15.520 bring it back but what we could do is um show pegp neighbor as an example 00:37:17.760 --> 00:37:21.360 all right so that's the different this one's slightly different the output here in this 00:37:21.360 --> 00:37:27.040 particular case but if you look and the flags are down here so a for auto mode remember we 00:37:27.040 --> 00:37:34.000 configured uh dsw1 for auto mode so there's the a for auto mode but if we had configured it 00:37:34.000 --> 00:37:52.000 um for desirable mode so if we look at um let's go to dsw1 show pegp 00:37:57.280 --> 00:38:06.719 one so show pegp internal shows me my internal information right so i was what i was configured 00:38:06.719 --> 00:38:12.639 with auto and if we do show pegp neighbor 00:38:16.080 --> 00:38:19.439 right you'll see in this case there's no flag so no flag 00:38:19.439 --> 00:38:23.919 means desirable so that's the way they do it for pegp slightly different so not the exact same 00:38:23.919 --> 00:38:26.320 output all right 00:38:33.679 --> 00:38:37.280 hope answer your question kind of to give you an idea what's going on there 00:38:38.479 --> 00:38:41.520 without me having to switch the entire environment over to lacp okay good 00:38:42.479 --> 00:38:46.719 all right so uh layer three ether channel bundle i won't do a configuration here in the live 00:38:46.719 --> 00:38:52.399 environment because we are well over the class time but basically what we're doing is we'll be 00:38:52.399 --> 00:38:58.719 using routed ports instead so with our routed ports we convert these to routed ports first 00:38:59.679 --> 00:39:04.239 and then we set up our ether channel bundle if you try to convert everything to routed ports 00:39:04.239 --> 00:39:10.000 after the fact it actually causes significant issues that you may not be able to recover from 00:39:10.000 --> 00:39:13.600 i've tried this in a live class before because somebody asked me to try it just i'll just just 00:39:13.600 --> 00:39:17.919 convert them to routed ports see what happens okay and it messed everything up it messed 00:39:17.919 --> 00:39:22.159 everything up and the only way to recover i spent about 30 minutes after class trying to troubleshoot 00:39:22.719 --> 00:39:28.399 we couldn't get it to work properly because the port channel did not want to become a routed port 00:39:28.399 --> 00:39:32.879 the physical interfaces were routed ports but the port channel would not become a routed port even 00:39:32.879 --> 00:39:37.600 if i typed it in on the port channel as well and it started creating conflicts and it would not 00:39:37.600 --> 00:39:42.080 start working right so basically wiped the whole thing out and configured it from scratch and 00:39:42.080 --> 00:39:47.520 everything worked properly at that point in time so something to keep in mind but i have had students 00:39:47.520 --> 00:39:52.560 say oh no it's worked for me in the past great perfect i loved hearing that but it didn't work 00:39:52.560 --> 00:39:56.479 for me in that live environment that day by changing the routed port so typically what 00:39:56.479 --> 00:40:00.639 you're going to do is you're going to default your interfaces convert them to routed ports 00:40:01.199 --> 00:40:06.959 and and how do you do that so default the interfaces could then go to the interfaces 00:40:06.959 --> 00:40:12.000 convert them to routed ports and then type in your channel group command and then create your 00:40:12.000 --> 00:40:18.000 port channel from there right and now because these were layer three interfaces this is a layer 00:40:18.000 --> 00:40:23.840 three port channel automatically so your port channel inherits the characteristics of the 00:40:23.840 --> 00:40:30.000 physical interfaces when it's created so a layer two interface creates a layer two port channel 00:40:30.000 --> 00:40:36.000 layer three interfaces create layer three port channels so therefore another thing to think 00:40:36.000 --> 00:40:39.760 about in the exam is if you have to bundle interfaces together and two of them are layer 00:40:39.760 --> 00:40:43.520 two and two of them are layer three you can't do that because it's not going to work they all have 00:40:43.520 --> 00:40:46.320 to be layer two or they all have to be layer three that's another thing to keep in mind 00:40:47.360 --> 00:40:51.919 right and now that it's layer three when i go to the port channel i can't create trunks i can't 00:40:51.919 --> 00:40:59.040 create vlans i can't do anything like that it is what ip addressing and i could also type in ip osp 00:40:59.040 --> 00:41:05.919 f1 area zero as an example and enable ospf on that port channel so now that would actually change 00:41:05.919 --> 00:41:12.080 my layer three environment to this so l3 would be this and now l2 you see it's on this side of 00:41:12.080 --> 00:41:18.239 the line now so i'm running ospf everywhere here or eigrp everywhere here now so yeah i'm routing 00:41:18.239 --> 00:41:23.600 over the bundled links that's really awesome so this is really great when you think about it on 00:41:23.600 --> 00:41:29.360 in a bigger scale environment so let's go back to our campus architecture example when you think 00:41:29.360 --> 00:41:36.000 about this campus architecture example and we erase everything we have here that is not required 00:41:36.000 --> 00:41:41.600 anymore and we think about the fact that the new environment this would be layer two this would be 00:41:41.600 --> 00:41:47.679 layer two and everything else is layer three well think about all of these being ether channel 00:41:47.679 --> 00:41:56.239 bundles as routed ports ether channel bundles as routed ports all of them everywhere everywhere 00:41:56.800 --> 00:42:03.520 and so now we've got layer three routed ports as ether channel bundles and therefore we're running 00:42:03.520 --> 00:42:12.080 ospf or eigrp on them so not only can we do load balancing within the ether channel bundle but we 00:42:12.080 --> 00:42:16.800 can also load balance across different ether channel bundles because of our dynamic routing 00:42:16.800 --> 00:42:21.360 protocols so for example if my dynamic routing protocol is deciding to load balance on two 00:42:21.360 --> 00:42:25.600 different paths then i could also load balance inside the ether channel bundle 00:42:27.040 --> 00:42:32.239 that is so cool i love it great stuff absolutely great stuff 00:42:34.399 --> 00:42:38.320 all right so the only difference here in this output that you're going to see is this value 00:42:38.320 --> 00:42:44.479 here everything else is still going to be the same so r for routed layer three so the exam if 00:42:44.479 --> 00:42:48.000 they're talking about a routed scenario and you see an s here well that's your problem it wasn't 00:42:48.000 --> 00:42:52.399 created properly or if they're talking about a layer two scenario and you see an r there well 00:42:52.399 --> 00:42:56.800 then again it wasn't created properly so that would be something you would have to solve okay 00:42:56.800 --> 00:43:00.959 maybe delete everything erase everything recreate it so that way they're the port channels created 00:43:00.959 --> 00:43:04.320 as layer two for the layer two interfaces or it's created as layer three for the layer three 00:43:04.320 --> 00:43:11.040 interfaces and so on and so forth all right and then what else do we have here we have our bandwidth 00:43:11.040 --> 00:43:19.760 for example as well right so load balancing for ether channel bundling is actually very finicky 00:43:19.760 --> 00:43:25.439 and it's not what we want it to be it's not as easy like our dynamic routing protocols load balance 00:43:25.439 --> 00:43:30.320 exceptionally well and that's why in newer modern day networks we really want to do routing 00:43:30.320 --> 00:43:37.120 everywhere and do our ether channel bundling as well with layer three links but still within the 00:43:37.120 --> 00:43:43.040 layer ether channel bundle it's it's not amazing like it doesn't excite me as much as it excites me 00:43:43.040 --> 00:43:48.959 when my dynamic routing protocols are doing load balancing and the reason why is because it's it's 00:43:48.959 --> 00:43:57.120 based on certain values like mac addresses or ip addresses or port numbers and it could be in a 00:43:57.120 --> 00:44:02.000 combination of different and if you pick the wrong one on your device load balancing is not actually 00:44:02.000 --> 00:44:08.399 going to work properly for you so there is a default on your platform and you can verify the 00:44:08.399 --> 00:44:13.199 default on your platform by typing in show ether channel load balance so right now india said you 00:44:13.199 --> 00:44:18.479 want if i type in show ether channel load balance you're going to see here that i'm load balancing 00:44:18.479 --> 00:44:25.760 based on source mac address that's it source mac address so what happens is the device when it 00:44:25.760 --> 00:44:31.040 receives a frame looks at the source mac address and then it uses that source mac address to decide 00:44:31.120 --> 00:44:36.800 which link to go across that's what it does but you'll see we have different options that we can 00:44:36.800 --> 00:44:41.360 choose from and depending on the platform will determine what options you can use now i haven't 00:44:41.360 --> 00:44:46.399 updated this for the new 9000 series of catalyst switches which are awesome by the way um but you 00:44:46.399 --> 00:44:51.199 know this is what we got okay and so when you look at these options here and you're not gonna have to 00:44:51.199 --> 00:44:54.959 worry about the models for the certification exam right but when you look at these options here 00:44:55.760 --> 00:44:59.919 right destination ip destination mac first of all does it ever make sense to 00:44:59.919 --> 00:45:04.560 load balance in your layer 2 environment based on destination mac address well let me answer that 00:45:04.560 --> 00:45:10.000 for you the answer is no why is that because most of the traffic comes from a pc and goes where 00:45:10.000 --> 00:45:17.600 in the layer 2 network almost 100 of your traffic goes from the pc to who in a layer 2 network 00:45:17.600 --> 00:45:23.439 the default gateway which is going to mean that everybody has the same destination mac address 00:45:23.439 --> 00:45:26.879 so that would be like the worst option you could use in your layer 2 network 00:45:27.520 --> 00:45:32.080 because the destination mac address is always the same so you would not want to do that in 00:45:32.080 --> 00:45:36.239 the layer 2 environment now elsewhere it might work fine but not in that layer 2 environment 00:45:37.600 --> 00:45:42.560 so what you choose is going to be the one that gives you the best results based on your traffic 00:45:42.560 --> 00:45:47.360 types based on your ip addressing based on your mac addressing based on port numbers as an example 00:45:49.600 --> 00:45:55.679 that's right meaning that what you choose you're going to have to test and how do you test it you 00:45:55.679 --> 00:46:00.560 choose it you implement it and then you monitor traffic flow on each link in the ether channel 00:46:00.560 --> 00:46:03.439 bundle and i'll show you the command that you can do that here i'll show you right now 00:46:03.439 --> 00:46:06.159 right the command you can do that with is right here 00:46:09.120 --> 00:46:13.439 clear counters and then do show interface the interface name the interface number and then 00:46:13.439 --> 00:46:20.000 include packets output yeah definitely i'm going to get to that that is like primarily like 00:46:20.879 --> 00:46:23.760 the best option you can use in reality when it's all said and done 00:46:24.879 --> 00:46:30.399 for most people that's the one that works best so when you do that for each interface then you'll 00:46:30.399 --> 00:46:36.800 be able to see hey are we getting a roughly equalized load balance from our users traffic 00:46:37.360 --> 00:46:40.959 now it's never going to be perfect because the load balancing with either channel bundling 00:46:41.919 --> 00:46:44.239 is based on what it's going to be based on 00:46:49.679 --> 00:46:54.479 using the ip address of the mac address but not the whole thing not the whole thing and it's also 00:46:54.479 --> 00:46:57.679 based on the number of links you have so let's say you have two links in your ether channel bundle 00:46:57.679 --> 00:47:06.000 that's it just two links and let's say we're load balancing based on destination ip so if we're load 00:47:06.000 --> 00:47:12.239 balancing based on destination ip what's the destination ip right here it's 40 so let me break 00:47:12.239 --> 00:47:19.199 that out we've got 0 0 1 0 1 0 0 0 2 links means we only look at the very last bit that's it just 00:47:19.199 --> 00:47:23.199 look at the very last bit and the reason why is because this link gets a value of 0 this link 00:47:23.199 --> 00:47:28.080 gets a value of 1 so whatever the very last bit is in the ip address is going to determine which 00:47:28.080 --> 00:47:33.600 link you use so if the destination ip is this then it's going across this link what if the 00:47:33.600 --> 00:47:39.199 destination ip was 42 well if it's 42 it also ends in a zero so it's going to use this link 00:47:39.199 --> 00:47:47.199 what if it's 44 so all again 44 would end in a zero as a binary so again it would use that same 00:47:47.199 --> 00:47:55.040 link so in this case because we only have two links all even destination ip addresses traffic 00:47:55.040 --> 00:48:00.959 would go on this link all odd destination ip addresses would go on that link so for some reason 00:48:00.959 --> 00:48:04.000 every single destination is even in your organization then you're not going to be 00:48:04.000 --> 00:48:09.919 doing any load balancing right so that's why you got to test that's why you got to test 00:48:11.439 --> 00:48:16.239 right what if you got four links though what do you got four links like i have now on the screen 00:48:16.239 --> 00:48:22.320 well now the numbering changes because it's 0 0 0 1 1 0 and 1 1 so now what does that mean you 00:48:22.320 --> 00:48:29.600 have to look at the last two digits now the last two digits so it ends in 0 0 it would go across 00:48:29.600 --> 00:48:33.919 this link if it ended in 0 1 it would go across this link if it ends in 1 0 it goes across this 00:48:33.919 --> 00:48:39.840 link if it ends in 1 1 it goes across that link what about if you have eight links well then if 00:48:39.840 --> 00:48:50.000 you have eight links it's 0 0 0 1 1 0 1 1 uh so 0 0 0 0 0 1 1 0 1 0 1 1 1 0 0 1 0 1 1 1 0 1 1 there 00:48:50.000 --> 00:48:56.479 you go so again if the last three values end in zero then it's using the first link if it's 0 0 1 00:48:56.479 --> 00:49:01.040 etc etc etc etc you get the picture now here's the thing about either channel bundling as well 00:49:01.840 --> 00:49:08.000 because it's based on this we're dealing with powers of two so what would happen for example 00:49:08.000 --> 00:49:16.239 if you only had three links well if you only had three links how many values would we have to use 00:49:16.239 --> 00:49:21.919 we still have to use two values so with three links we would still have to use two values 00:49:22.719 --> 00:49:28.959 so what does that mean you'd have 0 0 0 1 1 0 1 1 okay where where's 1 1 where's 1 1 we still 00:49:28.959 --> 00:49:35.919 need four values because it's 2 to the power of 4 or 2 to the power of 2 i mean right 2 to the power 00:49:35.919 --> 00:49:44.479 of 2 it's 4 we got four values but we only have three links well then that means one of these 00:49:44.479 --> 00:49:50.959 links for example is going to get both so now we've got 50 of the traffic going here 25 percent 00:49:51.679 --> 00:49:55.439 going here 25 percent going here so the moral of the story is this with either channel bundling you 00:49:55.439 --> 00:50:04.959 should use two four or eight links two four or eight links powers of two if you don't then you're 00:50:04.959 --> 00:50:09.360 never going to get a proper load balancing you can do it but you're never going to get a proper 00:50:09.360 --> 00:50:15.280 load balancing because other links have to take on that extra traffic so for example if you went 00:50:15.280 --> 00:50:21.600 ahead and you went with five links just five links that means three out of those five have to 00:50:21.600 --> 00:50:29.840 take on double duty for traffic because of the number of what the number of binary values that 00:50:29.840 --> 00:50:38.560 have to be referenced in that particular case right so that is how it works so it takes it 00:50:38.560 --> 00:50:44.879 takes time you have to decide do you want to use source ip do you want to use destination ip do you 00:50:44.879 --> 00:50:53.760 want to use source mac or destination mac right but and then monitor the traffic for a couple of 00:50:53.760 --> 00:50:57.199 days and find out what you're getting and if you're getting an even distribution across then 00:50:57.199 --> 00:51:01.520 you know it's pretty good but that doesn't mean it's still going to be good in a couple of months 00:51:01.520 --> 00:51:04.800 from now if you make changes elsewhere in your organization and ip addresses change elsewhere 00:51:04.800 --> 00:51:09.919 in your organization that can change everything so you have to constantly watch it however 00:51:10.560 --> 00:51:16.959 jeffrey says we use source and destination ip we use a combination of the two we don't just use 00:51:16.959 --> 00:51:23.600 like source ip or destination ip or source mac or destination mac we use source and destination 00:51:23.600 --> 00:51:27.520 combined together and that's where an exor operation comes into play so i'm going to use 00:51:27.520 --> 00:51:34.159 a mac address as an example but more than likely the um more than likely something like the source 00:51:34.159 --> 00:51:38.159 and destination ip would be your better option in the real world meaning if i'd try that one first 00:51:38.159 --> 00:51:43.520 folks and see if you get a nice low distribution with source and destination ip and if you do just 00:51:43.520 --> 00:51:47.520 stick with it don't waste more of your brain cycles trying to figure out there's another 00:51:47.520 --> 00:51:51.520 better option if you're getting good results of source and destination ip go with it if you can 00:51:51.520 --> 00:51:56.399 do source and destination port number on higher end platforms use that one because you have more 00:51:56.399 --> 00:52:01.199 variety in port numbers than you would in ip addresses because if you think about port numbers 00:52:01.199 --> 00:52:10.639 i could have a user going to the same server 10 times using three different source port numbers 00:52:10.639 --> 00:52:16.800 so if i do a source and destination port number combination that means that user's traffic could 00:52:16.800 --> 00:52:23.199 be distributed across those links really well as a single user so if using just source of destination 00:52:23.199 --> 00:52:30.159 ip typically one a user will more than likely use the same link over and over again more than not 00:52:30.159 --> 00:52:34.399 but if you do port numbers that's going to change drastically all right but in the end here what do 00:52:34.399 --> 00:52:40.320 we have in the end for example with the xor operation so if we're doing 33 and 44 because 00:52:40.320 --> 00:52:48.560 we're going to do source and destination in this case so 0 0 1 1 0 0 1 1 for 33 and then 44 is 00:52:48.560 --> 00:52:54.879 going to be 0 1 0 0 0 1 0 0 so with the x or you can see my little chart in the bottom right hand 00:52:54.879 --> 00:53:05.280 corner 0 0 equals 0 0 1 equals 1 1 0 equals 1 1 1 0 equals 0 so 1 and 0 equals what 1 and 0 equals 1 00:53:06.239 --> 00:53:15.040 1 and 0 equals 1 0 and 1 equals 1 so we think about our port 0 0 0 0 0 1 0 1 0 1 0 00:53:16.000 --> 00:53:25.439 0 1 1 1 0 0 1 0 1 1 1 0 1 1 1 then now source and destination in this case is 1 1 1 we're 00:53:25.439 --> 00:53:30.479 going to take the last link on the thing so typically with the xor operation you're going 00:53:30.479 --> 00:53:37.199 to get better results and so i like to uh try source and destination ip first if it works great 00:53:37.199 --> 00:53:40.879 if you're still not getting anything uh good out of that then start playing around some of the other 00:53:40.879 --> 00:53:46.719 ones and so definitely clear counters but before you clear counters always double check if you can 00:53:46.719 --> 00:53:49.760 actually clear counters before you do so the last thing you want to do is clear counters and if 00:53:49.760 --> 00:53:53.679 there is some sort of monitoring system that's relying on your counters and you go and clear 00:53:53.679 --> 00:53:58.800 them that could be a bad thing so always double check that first and then do a show interface 00:53:58.800 --> 00:54:03.679 include packets output and that'll let you double check so if i go over to dsw1 as an example 00:54:04.479 --> 00:54:11.760 show interface gigabit 1 0 17 pipe to include packets 00:54:13.600 --> 00:54:20.719 output as an example that's the that's what you would see as your option so now if we go to 18 00:54:22.719 --> 00:54:26.639 and we go to 19 and we go to 20 00:54:27.199 --> 00:54:34.080 17 right now i haven't cleared my counters in ages and we just set up load balancing so 00:54:34.080 --> 00:54:37.520 this is kind of misleading isn't it so that's why something like clear counters 00:54:40.959 --> 00:54:44.399 yes i want to clear all my interface counters and now if we go back and look through everything 00:54:45.199 --> 00:54:48.959 17 you'll see we're back down low so now monitor this for a couple of days 00:54:49.520 --> 00:54:54.399 come back and review to see where we're at in regards to all these different interfaces 00:54:54.399 --> 00:54:58.479 and then you'll be able to get a better picture of what is going on in regards to ether channel 00:54:58.479 --> 00:55:04.399 bundling and there you go folks we are officially done today's session and thanks for hanging out 00:55:04.399 --> 00:55:08.560 with me a little bit longer than we had intended i really do appreciate that but hopefully you have 00:55:09.439 --> 00:55:14.399 gleaned a lot of valuable information from the discussions that we have been able to have today