Zach Gibbs, Content Developer, Juniper Networks

Seamless EVPN-VXLAN Stitching – Verification

Learning Bytes Data Center
Zach Gibbs Headshot
Screenshot is from Zach Gibbs demo, showing white code on top of a black screen, as the EVPN-VXLAN verification process is taking place. Words like “input packets” and “output packets” and “VXLAN endpoint type” are shown with data. The bottom paragraph is highlighted in blue.

Juniper Learning Byte: How to verify seamless EVPN-VXLAN stitching operations

Juniper’s Zach Gibbs walks you through the process of seamless EVPN-VXLAN stitching. This video is most appropriate for users with a high degree of knowledge and skill with data center technologies.

Show more

You’ll learn

  • An example of the topology involved, with two different data centers 

  • How to dive into the topology to make sure things are working correctly

  • Step-by-step how the verification process works 

Who is this for?

Network Professionals Business Leaders


Zach Gibbs Headshot
Zach Gibbs
Content Developer, Juniper Networks


0:00 [Music]

0:12 hello my name is zach gibbs and i'm a

0:15 content developer within education

0:18 services inside juniper networks and

0:21 today we will be going through the

0:23 seamless evpn vxlan stitching

0:26 verification learning byte

0:28 all right so here is our topology in

0:31 this topology we have two different data

0:33 centers dc1 and dc2

0:36 and then in dc1 we have two qfx devices

0:39 that's spine one and leaf one and then

0:42 we have the host one device and host one

0:44 is a part of vlad10 and uses vni5010

0:49 and you can see the address here is


0:52 it's in that subnet and uses dot one

0:55 and then in dc2 we have spine 2 and leaf

0:58 3 which are qfx devices and then we have

1:01 host 2 which is a part of vlan 10 and

1:03 uses vni 5010 so same vni and vlan as

1:07 host 1.

1:08 and then host 2 here

1:11 uses the

1:14 ip address and

1:16 we have

1:17 seamless evpn vxlan stitching configured

1:20 and working

1:22 and if you're interested in the

1:23 configuration i do have other learning

1:25 bytes that go over the configuration so

1:27 please check those out

1:29 and then with this

1:31 so

1:32 it is working and functioning host one

1:35 can communicate with host two

1:37 and so things are working but we want to

1:40 dive into the individual qfx devices and

1:43 make sure that things are working

1:44 correctly

1:46 and so one other thing to point out or

1:48 some other things to point out is the

1:50 loopback addresses of the qfx devices

1:52 spine1 uses 192 168 100.1

1:56 spine 2 uses dot 2. leaf 1 uses diet 11

2:00 and leaf 3 uses dot 13. you can see the

2:02 host addresses here as well we already

2:05 discussed those

2:06 and then we have the interconnect

2:08 vxlan tunnel parameters now this is

2:11 important here that uh you have the esi

2:14 from as you configured on spine1 you see

2:16 it here and spine2 you see it here now

2:19 that's a big those are some big numbers

2:20 to try to memorize but all you have to

2:22 realize is how i name these

2:25 is with the esi with spine1 it's 1 1 1 2

2:28 1 3 1 4 and so forth with spine 2 it's 2

2:32 1 2 2 2 3 2 4 and so forth so keep that

2:34 in mind that'll make it easy to

2:36 recognize which esi interconnect esi

2:39 that is is from spine1 or spine2 and

2:42 then you have the route distinguishers

2:43 here for spine1 and spine2 based off the

2:46 loopback ip addresses for those devices

2:49 and then the route target we see here

2:51 that's the interconnect route target is

2:54 uh target colon one colon one two three

2:56 and then you see the interconnect vni

2:59 that will be shared across the

3:01 interconnect

3:02 vxlan tunnel okay so with that let's go

3:04 ahead and jump to the cli and get this

3:06 started

3:08 so here is the cli for leaf one let's

3:11 look at the show ethernet switching

3:13 vxlan tunnel endpoint remote

3:16 command

3:17 and we see some output here

3:19 and let's go through it so we can see

3:21 that

3:22 this is the local address the the local

3:25 anchor point we can see that we're using

3:27 uh interface the loopback interface and

3:30 we see the loopback address that we're

3:32 using for that so that's the ip that's

3:33 associated with the local vxlan tunnel

3:36 anything that starts from there and then

3:38 you can see here that we do have one

3:39 vxlan tunnel coming in

3:41 it's from the spine1 lupec address we

3:44 know that that 182 168 100.1 is that

3:48 spine1 loopback address

3:50 and we see the vtep interface there

3:52 and then we see this mode this mode says

3:54 rnve

3:56 now that tells us it's a regular vxlan

3:58 tunnel and that's exactly what we want

3:59 to see this isn't a

4:02 interconnect vxlan tunnel and when we

4:04 look at spine1 we'll be able to see the

4:06 difference there it won't be rnve for

4:08 the tunnel that goes from spine 1 to

4:10 spine 2. there'll be a tunnel that's

4:11 coming from leaf 1 to spine 1 that will

4:14 show rnve under the rv temp mode field

4:18 so keep that in mind then we can see the

4:20 vni here that is being used and that 5 0

4:22 1 0 that's exactly what we want to see

4:25 and then notice the the v tip interface

4:27 we can look at that in more detail to

4:30 see what's happening here and so

4:32 let's go ahead and clear interface

4:34 statistics just say all

4:37 and then show interfaces vtep

4:40 and there's going to be the vtep

4:42 interface uh the the vtep interface that

4:45 is being used here is the vtep

4:48 32769

4:49 and you can see we have some input and

4:51 output packets let's go ahead and do

4:52 that command again and you can see that

4:54 it's incrementing and that's perfect

4:55 that's what we want to see the traffic

4:56 is currently flowing between host one

4:58 and host two and we can see that it's

5:00 actually going into that vxlan tunnel so

5:03 perfect that's what we want to see

5:05 and so let's go ahead and jump to spine

5:07 one and look at the same thing so let's

5:09 do the

5:10 show ethernet switching vxlan tunnel

5:12 endpoint remote command

5:15 and you can see here we have some more

5:16 information and that's good because we

5:18 have we should be having two vxlan

5:21 tunnels terminate on spine1 which we do

5:24 have so the first one you see here you

5:26 can see the ip address for the rv tab

5:29 ipfilled

5:31 shows that it's spine 2's loopback

5:34 address

5:35 and then we can see the vtep interface

5:37 and that's

5:38 3270 and then we have the when vtep

5:42 output here

5:44 under the rv temp mode fill now if you

5:47 see that that tells you right there that

5:49 it is an interconnect

5:51 vxlan tunnel and because the next output

5:53 you see here this is the loopback

5:55 address of leaf one and you also see in

5:58 here the vtep interface that is

6:00 vtep.32770

6:03 and also under the rvtep mode

6:06 field we have rnve so we know that's

6:09 just a plane vxlan tunnel here is a

6:12 spine two let's do the show ethernet

6:15 switching vxlan tunnel endpoint remote

6:18 command

6:19 and we can see here that we have

6:21 a tunnel coming from 192 168 100.1 which

6:25 is spine1

6:26 and then we can see the vtep interface

6:29 that is associated with it and we do

6:31 have wan vtep under the rv tip mode

6:35 field that's perfect that's what we want

6:37 to see that is the interconnect vxlan

6:39 tunnel and then we have

6:42 192 168 100.13 which is leaf three

6:46 and we can see the rnve

6:48 vtep mode or rv tip mode

6:51 and so that's a normal vxlan tunnel

6:53 that's exactly what we want to see so

6:55 let's jump to leaf three and just take a

6:56 quick look there show ethernet switching

6:58 vxlan tunnel endpoint remote

7:01 and perfect see that rnv e

7:04 and we can see it's coming from

7:06 spine2 and of course the vnio5010

7:10 and so things look really good there

7:11 it's exactly what we should see

7:14 okay so with that let's go ahead and

7:17 look at these uh vtep interfaces so

7:20 clear

7:21 in our clear statistic our interface

7:24 statistics all

7:25 and get rid of any sort of

7:28 information that and so we can start

7:30 fresh and so then let's do the show

7:32 interfaces vtep

7:34 and recall that it's vtep32769

7:37 and we see that here

7:39 here's the interface and we see input

7:42 and output packets going in and out of

7:44 that tunnel that's what exactly what do

7:46 we want to see

7:48 and we can run that command again and we

7:50 can see that output is incrementing and

7:53 then if we jump to spine two

7:55 throw interfaces v temp actually let's

7:57 clear the statistics first clear

8:00 interface statistics all

8:03 then show interfaces vtep and with this

8:05 it's

8:07 vtep32770 and vtep327

8:13 and we can see these interfaces we can

8:15 see we have input and output packets

8:19 on both those interfaces we can run that

8:21 command again and see that those packets

8:24 those counters are incrementing that's

8:26 perfect that's what we want to see and

8:29 we can look at spine1 to verify that as

8:38 well and we can see here that we do have

8:42 packets incrementing that looks great

8:44 for those vtep interfaces

8:47 and one other thing i do want to point

8:49 out that i kind of forgot before is the

8:50 vx lan endpoint address you can see that

8:54 here you can see that this is for spine

8:56 two and this interface is for leaf one

9:00 and so that looks good

9:02 we see that incrementing and we could

9:03 jump to leaf one we would have very

9:05 similar output actually let's just do

9:07 that real quick

9:12 and then let's do the show

9:14 interfaces vtep

9:16 you can see here that we do have we can

9:18 see that the vx

9:21 lan endpoint address is for spline one

9:23 we see that the packets aren't

9:24 incrementing that looks good

9:26 and so great so things look good there

9:29 and so let's jump back to spine1

9:32 and in spine1 let's do the

9:35 show route evpn

9:39 mac address so this is going to be the

9:41 mac address of host one

9:48 and we can look at the route table here

9:50 and see that we are getting routes for

9:52 that mac address

9:53 and we see that in the

9:55 bgpevpn.0 table as well as the default

9:58 switch evpn.0 table that's perfect

10:00 that's exactly what we want to see and

10:04 something else we can look at is the

10:06 see the show route

10:08 basically same command let's do the

10:10 extensive output though and then we're

10:12 going to match on communities

10:16 and helps if i put match in there first

10:21 and we can see the communities that are

10:22 coming through now we see this target

10:24 colon one colon one two three that is

10:26 the inter connect community

10:28 and then we see the target 65

10:32 colon one that is the community that is

10:36 added to the default switch

10:38 configuration and so we should be seeing

10:40 both of those communities show up for

10:42 or in this output

10:44 and then we can uh

10:46 we can uh match on the esi value

10:52 we can see here actually let's

10:54 get rid of

10:56 zero zero colon zero zero

10:59 and we can see here that this route is

11:01 getting the

11:03 uh

11:03 the esi value of that interconnect esi

11:07 from spine1 so that's what's being added

11:09 to the route

11:10 and that route is then being passed on

11:11 to spine2

11:14 and

11:15 let's see what next should we look at

11:17 let's uh let's go to leaf two and run

11:19 some of those not leaf two sorry let's

11:21 go to spine two and run some of those

11:23 same commands

11:24 so let's do the show

11:27 route evpn

11:28 [Music]

11:30 mac address and type in

11:33 that mac for host one

11:38 you can see here we have routes and the

11:41 uh the bgpe vpn0 and the default switch

11:44 ev pin.0 route tables it's perfect

11:48 that's what we should see

11:49 and then we can look at the extensive

11:52 command and match on

11:55 esi

11:57 and you can see here the esi values

11:59 you're seeing and this so this is

12:01 from

12:02 the

12:04 spine one and this is the esi for spine

12:06 two

12:09 and then we

12:11 another helpful

12:13 output to look at and it's going to be

12:15 the show ethernet switching flood

12:19 instant this is going to show our mesh

12:21 groups instance so we're going to say

12:23 default switch

12:26 and you see the different mesh groups

12:28 here and you can see the underscore

12:30 underscore ves and the underscore

12:32 underscore when flood mesh groups and

12:35 the ves mesh group is for the local dc

12:39 so it'll be

12:40 uh spine 2 to uh

12:43 leaf 3 and then the wan flood is for the

12:47 interconnect okay next a good command to

12:49 help verify things here is the evpn

12:53 uh instance

12:54 default switch again

12:56 extensive and we're going to match on

12:59 dci

13:00 and you can see in here that

13:02 we're getting

13:04 routes from 192.168.1

13:06 which is spine1 and we see a one mac and

13:09 ip route

13:10 uh two auto discovery routes and that's

13:13 good that's what we should be seeing

13:15 there

13:16 and then another command we can do the

13:18 show

13:19 evpn instance dci command and this will

13:23 give us some information on what's

13:25 actually being passed

13:27 and what the evpn interconnect and you

13:30 can see here we have the route

13:31 distinguisher and so this is spine two's

13:33 route distinguisher for the interconnect

13:36 uh route distinguisher and then we have

13:37 the import vrf import and the vr export

13:40 and we could look at those policies and

13:42 what that'll show it's just showing that

13:44 it's importing routes based on the vrf

13:47 target which is this target here

13:50 and that's we're using for the vrf

13:52 import and export

13:54 and so you would actually see that in

13:56 those policies themselves and something

13:58 else is the show evpn

14:01 database command

14:04 and here you can see

14:05 what we're getting we can see the

14:08 the vni we can see the mac address

14:11 so this is this one here is the mac

14:13 address for host 2. we haven't looked at

14:14 that one yet but we can see the active

14:16 source is

14:18 the loopback ip address for

14:21 leaf 3

14:22 and we can see the ip address and this

14:24 is host 2's ip address and then with

14:26 host 1 we see the mac address and then

14:28 we see under the active source we see

14:31 that

14:32 interconnect esi

14:34 that is configured on spine1 and then

14:36 you see the ip address for

14:39 host 1 there and then

14:41 another good command would be the show

14:44 evpn database mac address and let's use

14:47 host 2's mac address i'm just going to

14:50 copy it here

14:55 extensive

14:57 and you can see in here that we have

14:59 first of all we have dci created route

15:01 so this is a route that is going to so

15:03 spine2 is going to advertise this to

15:05 spine1 and it's going to be over the dci

15:08 so that's very important because if we

15:10 run that same command

15:12 for

15:13 uh the mac address for host

15:17 one

15:19 we'll see a little bit different output

15:22 we can see dc created route

15:25 and that means it was is local to the

15:27 local dc so that makes sense because

15:30 spine 2 to get to it to get to host 2

15:34 it just uses a normal ex v or vx land

15:37 tunnel

15:39 so that does bring us to the end of this

15:40 learning byte in this learning byte we

15:42 demonstrated how to verify seamless evpn

15:44 vxlan stitching operation so as always

15:47 thanks for watching

15:50 visit the juniper education services

15:52 website to learn more about courses

15:55 view our full range of classroom online

15:58 and e-learning courses

16:00 learning paths

16:02 industry segment and technology specific

16:04 training paths

16:06 juniper networks certification program

16:09 the ultimate demonstration of your

16:11 competence and the training community

16:14 from forums to social media join the

16:16 discussion

16:23 you

Show more