Strange behaviour on XR2 (lab 4) with MPLS forwarding

ANother post on 4.1 in lab 4, I dont think this particular quirk has been covered yet, sorry if it has, there is a forest of 4.1 posts out there. So I had issues with R7 getting to R3/R4 FOO loopbacks, and only those, everything else was fine including R1 and R6. The lfib on XR2 was weird, saying R3/R4 loopback 0 (GRT) was aggregated in VRF FOO. I thought this might be the POS link being an eejit as usual so I added in a new Gige link and costed out the POS one, still the same issue.  I ended up doign a commit/rollback to get it working.  Anyone seen this behaviour before. its new to me.  All I can say is I pity anyone who runs GSRs with IOX in production.

 


RP/0/3/CPU0:XR2(config-bgp-nbr)#do sho mpls for | i 10.0.0.*FOO
Fri Sep 27 11:15:05.111 UTC
20015  Aggregate   10.0.0.4/32        FOO                          0          
20016  Aggregate   10.0.0.3/32        FOO                          0  
        
RP/0/3/CPU0:XR2(config-bgp-nbr)#do sho ip ro 10.0.0.3
Fri Sep 27 11:18:35.206 UTC

Routing entry for 10.0.0.3/32
  Known via "isis 1", distance 115, metric 10, type level-2
  Installed Sep 27 10:51:15.449 for 00:27:19
  Routing Descriptor Blocks
    10.19.20.19, from 10.0.0.19, via POS0/7/0/0.1920
      Route metric is 10
  No advertising protos.
RP/0/3/CPU0:XR2(config-bgp-nbr)#do sho ip ro 10.0.0.4
Fri Sep 27 11:18:37.824 UTC

Routing entry for 10.0.0.4/32
  Known via "isis 1", distance 115, metric 10, type level-2
  Installed Sep 27 10:51:15.448 for 00:27:22
  Routing Descriptor Blocks
    10.19.20.19, from 10.0.0.19, via POS0/7/0/0.1920
      Route metric is 10
  No advertising protos.
RP/0/3/CPU0:XR2(config-bgp-nbr)#do sho cef 10.0.0.3
Fri Sep 27 11:18:44.752 UTC
10.0.0.3/32, version 3996, internal 0x4000001 (ptr 0xae3f2b50) [1], 0x0 (0xae3c1378), 0x0 (0x0)
 Updated Sep 27 10:51:15.452
 local adjacency point2point
 Prefix Len 32, traffic index 0, precedence n/a, priority 1
   via 10.19.20.19, POS0/7/0/0.1920, 9 dependencies, weight 0, class 0 [flags 0x0]
    path-idx 0 [0xaeea1058 0x0]
    next hop 10.19.20.19
    local adjacency
RP/0/3/CPU0:XR2(config-bgp-nbr)#do sho cef 10.0.0.4
Fri Sep 27 11:18:46.965 UTC
10.0.0.4/32, version 3995, internal 0x4000001 (ptr 0xae3f2da8) [1], 0x0 (0xae3c1558), 0x0 (0x0)
 Updated Sep 27 10:51:15.452
 local adjacency point2point
 Prefix Len 32, traffic index 0, precedence n/a, priority 1
   via 10.19.20.19, POS0/7/0/0.1920, 9 dependencies, weight 0, class 0 [flags 0x0]
    path-idx 0 [0xaeea1058 0x0]
    next hop 10.19.20.19
    local adjacency

Comments

  • I had strange issues with XR2 on rack2 :

     

    1) Everybody could ping XR2 VRF FOO loopback but when the ping came from XR2 towards other devices it would fail, further troubleshooting showed me that it was trying to send packets out a non-existant interface.  Solution was to enable interfaces (gig0/4/0/0.207 and .208), after enabling them if appeared on running config with vrf forwarding on VRF FOO , ipv4/v6 addresses and dot1q vlan, wich did not explain at all why it was trying to forward packets to that interface nonetheless, i deleted the vrf config and everything worked

     

    2) doing 6VPE XR2 would not populate CEF under VRF accordingly, did a clear CEF, wich apparently fixed the CEF table, but still no connectivity, a reload did the job.

     

    It seems that this device is broken, INE should take a closer look.

     

    just my 2 cents.

  • The XR I'm working on at the moment is my own one, grey market from a carrier so previosuly one careful owner. You would expect problems with rentable routers, like with a hire car or something, different people drive them different ways and they wear out. My one hasn't had that level of abuse so I am sticking to my theory that they're pieces of... :D Great for practicing, I mean would we have learned this much about their pitfalls without having them!? I'll try and convince my boss to buy me a CRS toplay with ha ha

  • Hi, same thing here on rack10 , just finished the first 4 sections of Lab 4 and got issues bw XR1 and R3. And in 6VPE R7 loopback reachability issue until a commit replace/rollback.

     

Sign In or Register to comment.