01-08-2011 11:02 AM - edited 03-06-2019 02:53 PM
So, I'm still learning eigrp and routing, so go easy on me!
This morning, I woke up to a bunch of emails that some remote systems were down. After looking deeper into it, it appears that some routes aren't being exchanged between my core switch and my edge switch.
Topology: two 6506 cores running vss, which connect to a 3750 edge switch, and run eigrp between them. The 6506 is connected with a 3650 with hsrp for failover to the internet.
ED01-3750#sh ip route
Codes: C - connected, S - static, R - RIP, M - mobile, B - BGP
D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area
N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2
E1 - OSPF external type 1, E2 - OSPF external type 2
i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2
ia - IS-IS inter area, * - candidate default, U - per-user static route
o - ODR, P - periodic downloaded static route
Gateway of last resort is 10.201.1.2 to network 0.0.0.0
172.16.0.0/24 is subnetted, 1 subnets
S 172.16.75.0 [1/0] via 10.74.193.1
10.0.0.0/8 is variably subnetted, 52 subnets, 6 masks
D 10.13.5.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.3.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.200.199.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.13.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.3.15.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.13.15.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.3.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.200.1.201/32 [90/1536] via 10.200.253.2, 3w0d, Port-channel1
D EX 10.0.0.0/8 [170/145800448] via 10.200.253.2, 03:44:21, Port-channel1
D 10.200.1.200/32 [90/1536] via 10.200.253.2, 3w0d, Port-channel1
D 10.1.0.0/16 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.3.5.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.13.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.5.2.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.5.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.21.15.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.20.15.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.22.15.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.19.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.31.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.20.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.21.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.22.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.19.15.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.19.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.22.5.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.31.15.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.21.5.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.20.5.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.19.5.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.22.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.21.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.31.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.20.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D EX 10.32.0.0/16
[170/256256] via 10.200.253.66, 03:45:24, GigabitEthernet2/0/14
C 10.200.253.0/30 is directly connected, Port-channel1
S 10.75.0.0/16 [1/0] via 10.200.253.74
D 10.100.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.100.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
C 10.200.253.72/29 is directly connected, GigabitEthernet2/0/21
C 10.200.253.64/29 is directly connected, GigabitEthernet2/0/14
S 10.74.193.0/24 [1/0] via 10.200.253.74
S 10.150.10.0/24 [1/0] via 10.201.1.2
D 10.200.2.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.202.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.200.3.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
C 10.201.1.0/24 is directly connected, Vlan511
D 10.200.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.205.1.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.200.10.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
D 10.200.1.9/32 [90/1536] via 10.200.253.2, 3w0d, Port-channel1
D 10.200.20.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
198.xx.xx.0/30 is subnetted, 2 subnets
D 198.xx.xx.x [90/429312] via 10.200.253.66, 3w0d, GigabitEthernet2/0/14
D EX 198.xx.xx.x
[170/256256] via 10.200.253.66, 3w0d, GigabitEthernet2/0/14
D 192.168.100.0/24 [90/3072] via 10.200.253.2, 3w0d, Port-channel1
S* 0.0.0.0/0 [1/0] via 10.201.1.2
VCORE-6506#sh ip route eigrp
10.0.0.0/8 is variably subnetted, 51 subnets, 6 masks
D EX 10.32.0.0/16 [170/256512] via 10.200.253.1, 03:43:06, Port-channel100
D EX 10.0.0.0/8 [170/145800192] via 10.200.199.3, 03:43:12, Vlan999
D 10.200.253.72/29 [90/28416] via 10.200.253.1, 3d20h, Port-channel100
198.18.35.0/30 is subnetted, 2 subnets
D 198.18.35.4 [90/429568] via 10.200.253.1, 3w0d, Port-channel100
D EX 198.18.35.8 [170/256512] via 10.200.253.1, 3w0d, Port-channel100
D*EX 0.0.0.0/0 [170/256256] via 10.200.253.1, 7w0d, Port-channel100
01-08-2011 11:43 AM
Ryan
You tell us that the edge switch is running EIGRP with the core. But you do not tell us how the edge switch is configured. In particular you do not tell us whether the edge switch is redistributing static routes into EIGRP.
I see in the route table of the edge switch that the 2 networks you mention are in the route table as static routes. If the edge switch is not redistributing static then that would explain why the core is not learning the routes.
HTH
Rick
01-08-2011 11:53 AM
thanks for the response. How do I tell if it's configured to redistribute static routes? What baffles me is that it was working fine for about the last month and all of a sudden it quit working this morning. (along with that other network to the remote site that came back up)
here's the config for the edge switch
router eigrp 100
network 0.0.0.0
redistribute static metric 10000 0 255 1 1500 route-map STATIC>EIGRP
passive-interface default
no passive-interface Port-channel1
no passive-interface GigabitEthernet2/0/14
!
ip classless
ip route 0.0.0.0 0.0.0.0 10.201.1.2 name DEFAULT-TO-ASA
ip route 10.74.193.0 255.255.255.0 10.200.253.74 name to_DR_net
ip route 10.75.0.0 255.255.0.0 10.200.253.74 name to_DR_Net
ip route 10.150.10.0 255.255.255.0 10.201.1.2
ip route 172.16.75.0 255.255.255.0 10.74.193.1
01-08-2011 12:07 PM
Ryan
Yes the information provided in this post does show that the edge switch is configured to redistribute static routes, but that is controlled by the route map STATIC>EIGRP. So what is configured in that route map?
Also is it possible that anyone made any config changes this morning? If it was working and stopped working this morning then it seems logical that either someone made a change this morning or that there was some network event that impacted things.
In your original post you gave the route table from the edge switch. Was this taken while the problem was actively being experienced or is this from a time when the network was working ok?
HTH
Rick
01-08-2011 12:12 PM
Hello,
In addition to Rick's spot-on comments, it would be very interesting to see the output of these commands both on the core and on the edge switch:
Especially the "show ip eigrp topology" should reveal whether the redistributed static routes are present in the EIGRP's topology database - the main data storage of EIGRP. If they are not present there (especially on the edge switch), then the redistribution is not working as expected.
Best regards,
Peter
01-08-2011 12:43 PM
Thanks for the continued help. Here's what's in the map:
route-map STATIC>EIGRP permit 10
match ip address prefix-list STATIC>EIGRP
01-08-2011 01:10 PM
Ryan,
The route-map references a prefix-list called STATIC>EIGRP - can you please post its configuration as well? Look for lines ip prefix-list STATIC>EIGRP in your configuration - we need to see all of them.
In any case, the static routes do not seem to be redistributed into your EIGRP process in this moment.
Best regards,
Peter
01-08-2011 01:32 PM
Thanks for the reply. It probably got buried in the last post, but the static map was at the top of my last reply. Thanks!
01-08-2011 01:50 PM
Ryan,
Don't confuse the route-map called "STATIC>EIGRP" with a prefix-list of the same name. You have pasted the route-map but not the prefix-list. Please read my requests more carefully
Best regards,
Peter
01-08-2011 02:05 PM
Oh, ok. I'm out on an errand right now, so I'll get that when I get back shortly. Thanks!
01-08-2011 03:53 PM
ok, so these are the two lines with that configuration:
ip prefix-list STATIC>EIGRP seq 5 permit 0.0.0.0/0
ip prefix-list STATIC>EIGRP seq 10 permit 10.150.10.0/24
01-08-2011 06:56 PM
Ryan
Thanks for the additional information. The first line of the prefix list is permitting redistribution of the default route and the second line permits redistribution of only a single specific network (10.150.10.0/24). No other static routes will be redistributed. So this means that the current configuration would not redistribute either 10.75.0.0 or 10.74.193.0 and therefore the edge switch would not advertise these networks to the core.
At first I thought that this was the problem. But as I examined the information available I realize that the default route on the core seems to have the edge switch as the next hop. Is that the case?
If the default route of the core does use the edge as the next hop, then the static routes from the edge router do not need to be advertised to the core since the default route on the core will send to the edge to get to those networks. So I believe that now the question becomes - why was the default route on the core not sending traffic for those networks to the edge and why did you need to add static routes to make things work?
Is it possible that there is a command in the core config for "ip classless" or for "no ip classless"?
HTH
Rick
01-08-2011 09:31 PM
Thanks a lot for the reply. Yes, on the core the gateway of last resort does go to the edge switch, so that makes sense the edge would not need to advertise those static routes to it. So, I guess that's why this is so confusing, why these two networks are not routing to that edge switch, but routing to the hsrp partner to the core, 10.200.199.3, as you saw in the traceroute above. And the weird thing, it all of a sudden just started happening this morning, and has been down ever since.
What's also weird, though, is that I got a bunch of emails at about 6:15 - 6:30 tonight, that made it look like the route started to work for a second, and then quit working again. I get email alerts from my Compellent SAN when it can't replicate to the remote volumes, and since it's been down since this morning it basically tells you once when they're down and that's it. Tonight at 6:15, I got more emails from it, which basically means that it connected again, then went down again, so that makes me think the route worked for a second, then quit again. Man, I wish I knew what was happening!
Also, to answer your question, both the core switch and the edge switch has ip classless in its config. Thanks a lot for sticking this out with me!
01-08-2011 10:13 PM
ok, so believe it or not I got it working again. I decided to do a reload on the the HSRP partner switch, the one that those networks kept trying to route to, and when it went down the routing got figured out and started working again! (I'm running HSRP with the core to a 3650 on another floor that is supposed to route out to the internet in case we lose the data center where the core is. This way, users can access the internet from their offices and potentially get to our DR site via citrix over the internet. Management requirement... ) Anyways, since no one uses that switch, I figured it couldn't hurt to reboot it.
So, that made it work again, but I kind of need to figure out how this happened so it doesn't happen again. As I was typing the above paragraph, it got me thinking that maybe the core switching failed over to that 3650, and that gateway of last resort is the internet, and the gateway of last resort on the 6500 is the edge switch. But, what I don't understand is, if that's what happened, why did the other network to my remote site that went down this morning come back, when these didn't? These were the only two that were affected. Weird.
Any insight is greatly appreciated. Thanks!
01-09-2011 05:15 AM
Ryan
I am glad that you got it working again. Here is my current best guess at the reason it happened. In the morning something happened that caused HSRP to make the 3650 the lead/active router and the core to be the standby. So traffic from devices at your site would forward to the 3650 rather than the core. I am guessing that the 3650 had routes to most of the subnets in the network but not to those two networks. And the 3650 has a default route that goes toward the Internet and not toward the 3750 as the core does. So traffic from PCs at your site would go through the 3650 and most of your networks work but these two fail.
Perhaps you can check the content of the routing table of the 3650 and verify whether it has most of the networks but not these two?
HTH
Rick
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide