CISCO-OSPF non-backbone area direct-Link re-release (direct-attached) Link redundancy deployment and maintenance experience sharing

Recently has been tuning some of the more detailed parameter configuration, there are some relatively simple but not often encountered in the network environment requirements. Coincides with the Mid-Autumn Festival, I did not intend to go out, determined to properly update a blog. At the same time review their own only system. I'm not going to tell this to the party one day:).

Well, not much to say, the efforts of the people will eventually meet Bole, not hard people will eventually be social elimination, this is the technical industry rules, this is a very good engineer culture of young people, why has been steadfast for the technology, push off the party, push off the date, push off the game, anyway, I am such a person, With bigotry and focus to solve all the problems encountered in the work, feelings will not be split.

Into the topic, see below, we are relatively familiar with, because the previous article is also used in this environment to do, in fact, is different needs, so naturally there will be a number of experience and configuration of the article appeared, or thank you see here.

Demand:

1. Using OSPF protocol Group networking

2. Loop to make it a backup line

The 3.area area is separated, rejecting the entire network of a region.

4 ...

5 ...

6. Other put to the next introduction, too much demand, will be the focus today "OSPF non-backbone area direct re-release (direct connection) Link redundancy deployment and maintenance"

650) this.width=650; "Src=" Http://s2.51cto.com/wyfs02/M01/87/49/wKiom1faJNDQ-D1CAADMfVUH5EQ372.jpg-wh_500x0-wm_3 -wmp_4-s_3726147970.jpg "title=" demo topology. jpg "alt=" wkiom1fajndq-d1caadmfvuh5eq372.jpg-wh_50 "/>

OSPF configuration is temporarily unavailable, because I am in the home, debugging needs Ssl-vpn dial into the room, so the demo will appear SSL network segment, please ignore it. I will mark the direction of the beginning in the direction of the entire flow. After all, they are professional!!

Demand 1, loop implementation is as follows:

650) this.width=650; "src=" http://s2.51cto.com/wyfs02/M01/87/49/wKiom1faJZOxzrqVAACkXuzg7n8003.jpg "title=" loop. jpg "alt=" Wkiom1fajzoxzrqvaackxuzg7n8003.jpg "/>

Key configuration:

Core routes:

router OSPF 110

Router-id 172.16.107.2

Redistribute connected subnets

Redistribute static subnets

Network 172.16.106.2 0.0.0.0 Area 0

Network 172.16.106.5 0.0.0.0 Area 0

Network 172.16.106.17 0.0.0.0 Area 1

Network 172.16.107.2 0.0.0.0 Area 0

Route B:

router OSPF 110

Router-id 172.16.107.4

Redistribute connected metric subnets

Redistribute static subnets

Network 172.16.106.18 0.0.0.0 Area 1

Network 172.16.107.4 0.0.0.0 Area 1

Route C:

router OSPF 110

Router-id 172.16.107.3

Redistribute connected metric subnets

Redistribute static

Network 172.16.106.6 0.0.0.0 Area 0

Network 172.16.107.3 0.0.0.0 Area 0

At this point, the entire network is interoperable, and the routing output of the 172.16.106.52/30 we learned in the core route is as follows:

O E2 172.16.106.52/30 [110/25] via 172.16.106.18, 01:57:17, Vlan2052

[110/25] via 172.16.106.6, 01:57:17, Vlan2051

Another: Note that at this point, the default cost of redistribution is 20, I here manually designated as 25, when the simulation interrupt test, traffic can be forwarded normally. For a core route, Route B and Route c any link failure, the OSPF routing table will automatically unload the link routing table, the principle is not described in detail. This explains the phenomenon only on demand.

My notebook test results are as follows:

650) this.width=650; "src=" http://s2.51cto.com/wyfs02/M02/87/49/wKiom1faMNyxCrtdAABZNOwWiOA564.jpg "style=" float: none; "title=" Route b.jpg "alt=" Wkiom1famnyxcrtdaabznowwioa564.jpg "/>

650) this.width=650; "src=" http://s2.51cto.com/wyfs02/M02/87/46/wKioL1faMN2DEnG8AABMqvtG6Fw141.jpg "style=" float: none; "title=" Route c.jpg "alt=" Wkiol1famn2deng8aabmqvtg6fw141.jpg "/>

conclusion, although it is to see that the route represents load balancing, the traffic will still be forwarded within the AREA0 with the intra-zone priority. The access to Route B and Route C goes as shown:

650) this.width=650; "Src=" Http://s2.51cto.com/wyfs02/M00/87/49/wKiom1faMdehzxlbAADVSWb9HUM971.jpg-wh_500x0-wm_3 -wmp_4-s_2877549540.jpg "title=" qq20160915133124.jpg "alt=" Wkiom1famdehzxlbaadvswb9hum971.jpg-wh_50 "/>

Of course, in order to test the redundancy, I manually modified the cost of Route B to make it a bit more priority, to achieve the purpose of hot switching.

There are many kinds of test methods, but basically it is the result: the neighbor down, (the neighbor turned here does not show, the effect is more egg pain)

The configuration is as follows:

rack1_106.18 (config-router) # Redistribute connected subnets metric

The next hop learns to Route B ( the link IP of Route b is 172.16.106.18), the SH IP route, the output is as follows:

650) this.width=650; "src=" Http://s4.51cto.com/wyfs02/M02/87/46/wKioL1faM7ihjWoHAACgaV_D3WM038.jpg "title=" Qq20160915134006.jpg "alt=" Wkiol1fam7ihjwohaacgav_d3wm038.jpg "/>

Then test the route trace, the output is as follows: and the above-mentioned route tracking is completely different, you can carefully observe, and then look down.

650) this.width=650; "src=" http://s1.51cto.com/wyfs02/M00/87/49/wKiom1faNCyBPO1ZAABUSgS0XE8824.jpg "style=" float: none; "title=" 111.jpg "alt=" Wkiom1fancybpo1zaabusgs0xe8824.jpg "/>

650) this.width=650; "src=" http://s4.51cto.com/wyfs02/M00/87/46/wKioL1faNCyCbNNKAABdOAXYThA864.jpg "style=" float: none; "title=" 222.jpg "alt=" Wkiol1fancycbnnkaabdoaxytha864.jpg "/>

Ping value comparison chart for link eager swap:

Before switching:

650) this.width=650; "src=" http://s2.51cto.com/wyfs02/M01/87/46/wKioL1faNsbjUOPIAALxBekxf_8054.jpg "style=" float: none; "title=" switch before-.jpg "alt=" Wkiol1fansbjuopiaalxbekxf_8054.jpg "/>

After switching:

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/87/49/wKiom1faNsfzXkabAAL0yAICrj0945.jpg "style=" float: none; "title=" after switching. jpg "alt=" wkiom1fansfzxkabaal0yaicrj0945.jpg "/>

Conclusion: The flow has been cut from the original area0 backbone area to the other side, note here is the hot switch, delay and no jitter, here why seems to delay there is jitter because of the VPN relationship (I tested it locally after the festival, to supplement. Offend everyone). Not cold cut. For you and intuitive to see the flow trend, please refer to

650) this.width=650; "src=" http://s2.51cto.com/wyfs02/M01/87/46/wKioL1faNLKx1z-sAADnEYpd4EY071.jpg "title=" 333. JPG "alt=" wkiol1fanlkx1z-saadneypd4ey071.jpg "/>

Maintenance experience, here to give you a bit, say too much, we will feel too thick, later to try to make a short story. Hey

1. Pay more attention to the quality of physical interfaces between point-to-point transmissions, for example: optical decay values of optical fibers

2. Monitoring must be done enough

3. Be sure to have an independent out-of-band management line that rejects business and manages line multiplexing.

Another: the technical details as described in this article are very much used in daily use. Because many operators need to put their own network looping deployment, to ensure the reliability, so there must be a main link because of municipal construction, causing link cutting switch. So these are the operators of the network workers are commonplace, if you are determined to run to operators network, please pay attention to some of the high-availability protocols, especially the high-level application of the agreement, otherwise go to the interview will be a variety of contempt!!!!!

There are missing or not rigorous places, looking forward to you crossing the great God more guidance. Do the technology seriously, focus on enough, do not need to be too gifted, because diligence is the entry of the network workers must have the quality.

---------The sharing of network workers from a two-tier operator

This article is from "Allen on the road-from zero to one" blog, please be sure to keep this source http://allen686.blog.51cto.com/6758434/1852939

CISCO-OSPF non-backbone area direct-Link re-release (direct-attached) Link redundancy deployment and maintenance experience sharing