Solved: Question re debug output of EIGRP receiving summary route

CriscoSystems · ‎10-30-2009

Alright then, Router R2 (172.31.1.6) has an EIGRP neighborsip with R1 over a serial link (172.31.1.5). R1 and its neigbor R5 have many many 24-bit loopback networks config'd, I wanted R1 to send one 20-bit summary address to its neighbor R2. It went rather well, but a couple of things I don't understand in the debug output (I had "debug eigrp packets terse" activated on R2 while I configured the summary-address on R1).

The debug output is in the attached .txt file.

R2 receives the UPDATE but says "Neighbor not yet found." ...well, where was it looking? The neighborship had been up for several minutes and all of the 24-bit networks that R1 knew about were listed in R2's routing table - you can't have more of an EIGRP neighborship than that!

Then after it receives the update it says "enqueueing update on Serial0" - the same interface the update arrived on, and R2 has no other neighbors. I know that EIGRP multicasts updates to 224.0.0.10, but does it not observe split-horizon?

Can anyone also decrypt some of the other output, for instance, is there a chart somewhere that tells me what each "Flag" means? What about "iidbQ" and "serno 17-17" etc...

Peter Paluch · ‎10-30-2009

Hello Seth,

Whenever you configure a summarization in EIGRP, it will drop and reestablish adjacency with its neighbors so that the un-summarized networks can be purged from the network quickly. The message "Neighbor not yet found" is therefore related to the restarted adjacency. Slightly more information can be found here:

http://www.google.sk/url?sa=t&source=web&ct=res&cd=5&ved=0CCAQFjAE&url=https%3A%2F%2Fcisco.hosted.jivesoftware.com%2Fservlet%2FJiveServlet%2Fdownload%2F3739-3-3214%2FTroubleshooting%2520EIGRP%2520CCNP_CLNv2_Posted_1_7_09.pdf&rct=j&q=EIGRP%3A+Neighbor...

Regarding the "enqueueing update" - yes, the EIGRP does observe split horizon but what it actually does is a split horizon with a poisoned reverse. If it learns a route from its neighbor, it will advertise it back to that neighbor as unreachable.

Regarding the other information in the debug output, I am afraid I don't know a definitive compendium for that. Regarding the flags, you can have a look on this description:

http://www.cisco.com/en/US/docs/ios/debug/command/reference/db_d1.html#wp1031719

Also this document is worth reading:

http://www.cisco.com/en/US/tech/tk365/technologies_white_paper09186a0080094cb7.shtml

The serno talks about the sequence number of the EIGRP packet (as you know, the EIGRP uses reliable multicasting that needs sequence numbers). Unfortunately, that is all I know about.

Best regards,

Peter

View solution in original post

Peter Paluch · ‎10-30-2009

Hello Seth,

Whenever you configure a summarization in EIGRP, it will drop and reestablish adjacency with its neighbors so that the un-summarized networks can be purged from the network quickly. The message "Neighbor not yet found" is therefore related to the restarted adjacency. Slightly more information can be found here:

http://www.google.sk/url?sa=t&source=web&ct=res&cd=5&ved=0CCAQFjAE&url=https%3A%2F%2Fcisco.hosted.jivesoftware.com%2Fservlet%2FJiveServlet%2Fdownload%2F3739-3-3214%2FTroubleshooting%2520EIGRP%2520CCNP_CLNv2_Posted_1_7_09.pdf&rct=j&q=EIGRP%3A+Neighbor...

Regarding the "enqueueing update" - yes, the EIGRP does observe split horizon but what it actually does is a split horizon with a poisoned reverse. If it learns a route from its neighbor, it will advertise it back to that neighbor as unreachable.

Regarding the other information in the debug output, I am afraid I don't know a definitive compendium for that. Regarding the flags, you can have a look on this description:

http://www.cisco.com/en/US/docs/ios/debug/command/reference/db_d1.html#wp1031719

Also this document is worth reading:

http://www.cisco.com/en/US/tech/tk365/technologies_white_paper09186a0080094cb7.shtml

The serno talks about the sequence number of the EIGRP packet (as you know, the EIGRP uses reliable multicasting that needs sequence numbers). Unfortunately, that is all I know about.

Best regards,

Peter

CriscoSystems · ‎10-30-2009

Wow, you DID set me up!

Thanks, I'll definitely have a look at those documents.

Peter Paluch · ‎10-30-2009

Hi Seth,

Well, I found another reference worth having look at: the great Jeff Doyle's TCP/IP Routing Vol.1 here:

http://books.google.sk/books?id=JjdF2yWqJAwC&pg=PA280&lpg=PA280&dq=iidbq&source=bl&ots=CdXFrh5yyc&sig=G_PIYkKhafm7Zz_9fihoeJTFou0&hl=sk&ei=rpXrStWCJIu-mQPK29GZDw&sa=X&oi=book_result&ct=result&resnum=8&ved=0CCYQ6AEwBzgK#v=onepage&q=iidbq&f=false

And immediately I found that I have mystified you: the "serno" is not a sequence number but rather a some kind of pointer to an internal IOS/EIGRP structure.

Sorry.

Best regards,

Peter