cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Bookmark
|
Subscribe
|
534
Views
5
Helpful
5
Replies

Routes lost, could it be Metric?

tahscolony
Level 1
Level 1

We have a mix of ASR and Juniper. The Junipers are strictly OSPF, the ASR's mix OSPF and BGP.  We peer with AWS and distribute their routes into our network via BGP>OSPF redistribution. This appears to be working, except for one issue, if we lose the link between the primary ASR, the routes received by the Juniper that is the gateway for all internal networks loses the routes to AWS, even though the other Juniper has them.

There is a pair of 9214 linked together, OSPF has priority set so that one 9214 is master, and there is a pair of ASR 1001-x linked together with BGP and OSPF, each one connects to a 9214, so picture a square, 9214 on the top left and right corner, ASR on the bottom left and right. The left side are the primary routers. If the link between the primary ASR to 9214 drops, all routes that are distributed from BGP into OPSF from the ASR are lost on the left 9214. The right side has all the routes, but the right 9214 is not sending the routes to the left 9214.

The only thing I can see that might be causing an issue is that the right ASR has

redistribute bgp ASN metric 2 subnets tag ASN route-map BGP-OSPF

The left ASR has the same line, but Metric is 1.

This is on the OSPF configuration.

On the 9214, the left one shows

52.36.0.0/14       *[OSPF/150] 09:23:37, metric 1, tag ASN
                    > to 172.178.251.5 via xe-0/2/3.0

The right one shows

52.36.0.0/14       *[OSPF/150] 09:40:48, metric 1, tag ASN
                      to 172.178.1.7 via irb.1000
                    > to 172.178.251.13 via xe-0/2/3.0

What I am not seeing is  to 172.178.1.9 via irb.1000 on the left 9214.   So when xe-0/2/3.0 drops off, and the neighbor goes away, it no longer has a route to 52.36.0.0.   I'm working with Juniper on those switches, but need to know if by any chance its the ASR causing it with the metric in redistribution.

Thing that is puzzling most, I can see the route in the OSPF database on all routers showing both ASR router-id's.  The left 9214 does show both in its database, and does know how to get to both ASR, so why is it not seeing a secondary route like the right one is?

5 Replies 5

Hello,

post the configs of both 9214 routers...

Left 9214

set protocols ospf export OSPF_POLICY
set protocols ospf area 0.0.0.0 interface irb.1000 priority 255
set protocols ospf area 0.0.0.0 interface xe-0/2/3.0 interface-type p2p
set protocols ospf area 0.0.0.0 interface xe-0/2/3.0 neighbor 172.178.251.5

set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_STATIC from protocol static
set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_STATIC then accept
set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_CONNECTED from protocol direct
set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_CONNECTED then accept

Right 9214


set protocols ospf export OSPF_POLICY
set protocols ospf area 0.0.0.0 interface irb.1000 priority 254
set protocols ospf area 0.0.0.0 interface xe-0/2/3.0 interface-type p2p
set protocols ospf area 0.0.0.0 interface xe-0/2/3.0 neighbor 172.178.251.13

set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_STATIC from protocol static
set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_STATIC then accept
set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_CONNECTED from protocol direct
set policy-options policy-statement OSPF_POLICY term REDISTRIBUTE_CONNECTED then accept

Hello,

can you post a schematic drawing of your setup ? Are both Junipers connected to AWS ? Post the configs of both ASRs as well...

Hello
E1 route always preferred over E2 routes
Both the left/right rtrs are receiving the same prefix into ospf - However as you have stated correctly the router on the right is seeing a better path being advertise by the left router with a metric-type 1 so this is being preferred over it own prefix of metric-type 2.


res
Paul


Please rate and mark as an accepted solution if you have found any of the information provided useful.
This then could assist others on these forums to find a valuable answer and broadens the community’s global network.

Kind Regards
Paul

tahscolony
Level 1
Level 1

Resolved. Took a while but finally tracked it down to a bad ACL in BGP.  It was not advertising the correct network back to the peer, so the routes never propagated correctly.