|Previous ||Table of Contents ||Next |
Through the course of troubleshooting this case study, we found the value of several key IOS commands to help resolve OSPF and performance-related problems:
- show ip ospf neighbors
- show ip interfaces
- debug ip ospf events
- ip accounting and show IP accounting
We discovered that OSPF stub area configurations have some dangers if there is a misconfigured network interface or router port on the local LAN that the router attaches to. In most cases, the benefit of having a smaller routing table generally outweighs the drawbacks of using stub area, but it is important to understand the implications that can be caused by OSPF stub area misconfigurations. We also found that it is also critical to have the correct authentication key in order to form an adjacency. For Cisco router performance problems, we examined why output drops were occurring and how to control them by raising output hold queues and turning off weighted fair queuing.
In conclusion, this case study demonstrates that if you follow a structured troubleshooting methodology with solid fundamentals, and dig deep into the trenches, you can divide and conquer any networking problem.
Frequently Asked Questions (FAQs)
- QMy routers are experiencing some performance problems and I can see that its interfaces are experiencing input and/or output drops. What is happening and how do I fix it?
- AWhat are the causes of output queue drops? If the router interface is not able to clear up the queue as fast as the router is sending them, this will result in output drops.
How do you fix this problem? You can modify the output queue using the interface subcommand hold-queue nn out, where nn is a value. You should increment it by a value of 50 and monitor.
- QWhen doing a debug ip ospf events I get an error message that is telling me a bad checksum has occurred.
- AChecksum messages indicate that the checksum within a packet did not compute properly which does happen sometime.
- QWhen doing a debug ip ospf events I get an error message that says the following:
OSPF: Router/Net Link state Update timer in area 1 OSPF: No full nbrs to build net LSA
- AThis error message indicates that the router is the OSPF Designated Router (DR) and there are no other routers on the segment to build an adjacency with. Therefore no network LSAs will be created.
- QHow do I produce a stable OSPF network with serial links flapping?
- AThere is nothing you can do in the OSPF process to stabilize a network with flapping serials. You need to address the real problem and fix the flapping serial lines. If you want to Band-Aid the problem for a while, you could set the keepalives on the serial interfaces higher than 10 seconds, especially in point-to-point environments. Summarization is also an alternative to this problem.
- QWhat are the well-known class D IP addresses?
- ATable 8-17 documents the well known class D IP addresses.
Table 8-17 Well known class D IP addresses.
|Class D IP Address ||Purpose |
|126.96.36.199 ||All hosts on a subnet |
|188.8.131.52 ||All routers on a subnet |
|184.108.40.206 ||All DVMRP routers |
|220.127.116.11 ||All OSPF Designated Routers |
|18.104.22.168 ||RIP v2 |
|22.214.171.124 ||Network Time Protocol (NTP) |
|126.96.36.199 ||SGI Dogfight |
|188.8.131.52 ||Audio news |
|184.108.40.206 ||IETF audio |
|220.127.116.11 ||IETF video |
- QIn Priority Queuing, are RIPs in the high queue? When we set up priority queuing, do IP RIP packets or OSPF LSAs automatically go into the high queue, or is there something that needs to be done to make this happen. I see that in Custom Queuing, the hello packets go into queue 0, but what about the update table packets?
- AThe four priority queueshigh, medium, normal, and loware listed in order from highest to lowest priority. Keepalives sourced by the network server are always assigned to the high-priority queue; all other management traffic (such as IGRP updates) must be configured. Packets that are not classified by the priority list mechanism are assigned to the normal queue.
To get IP RIP packets and OSPF LSA packets to go into the high queue, you need to specify the protocol or port number for the protocol. A list of protocol and port numbers can be found in RFC 1340 at: http://www.internic.com/RFC/rfc1340.txt. RIP packets can be specified as high priority directly using a priority-list command:
router(config)#priority-list 1 protocol ip high udp rip,
OSPF as a protocol needs to be specified through an access list before it can be assigned to the priority-list:
router(config)#access-list 100 permit ospf any any router(config)#access-list 100 permit udp any any eq rip /* for rip */ router(config)#priority-list 1 protocol ip high list 100
- QI have a Cisco 7010 router that is running OSPF and IGRP. OSPF is used to route within my network, and IGRP is used to my service provider. I redistribute between them. I also have two static routes. I have a requirement to advertise one of my static routes out all of my OSPF interfaces and my IGRP interface. The second static route needs to be advertised out all of my OSPF interfaces, but needs to be filtered from being advertised to my service provider via IGRP. I think I need to use the distribute-list and access-list commands in my IGRP process, but I am not sure how to do it.
- AThe best way is to use route-map. Here is an example:
! router ospf 333 redistribute static subnets route-map ospf-static network 18.104.22.168 0.0.255.255 area 0.0.0.0 ! router igrp 187 redistribute static subnets route-map igrp-static network 22.214.171.124 ! ip route 126.96.36.199 255.255.255.0 188.8.131.52 ip route 184.108.40.206 255.255.255.0 220.127.116.11 ip route 18.104.22.168 255.255.0.0 22.214.171.124 ! access-list 7 permit 126.96.36.199 0.0.0.255 access-list 7 permit 188.8.131.52 0.0.0.255 access-list 7 deny any access-list 8 permit 184.108.40.206 0.0.255.255 access-list 8 deny any ! route-map ospf-static permit 10 match ip address 7 ! route-map igrp-static permit 10 match ip address 8
In the preceding example, there are three static routes defined. I want the first two static routes to be redistributed into my OSPF domain only, and the very last static route to be redistributed into my IGRP domain only.
- QWhen is a topology change from RIP, IGRP or EIGRP redistributed into OSPF? My router learns of a change of topology by its RIP, IGRP, or EIGRP process. It has another OSPF routing process running on another interface. The information from the first protocol is redistributed into OSPF. Does this redistribution occur every time a RIP, IGRP or EIGRP update is received? Or does OSPF compare the first routing DB with its own DB every dead interval?
- AAn OSPF update goes out every time a RIP, IGRP, or EIGRP update is received and the router notices a change in the routing information. Changes in the routing tables are caused by incoming routing updates or interface state changes.
- QUsing OSPF, how do you generate a summary route? When I redistribute one OSPF process into another one (without the subnets parameter), a subnetted network is not redistributed and no summary route for that network is generated. I understand that OSPF is a classless protocol, but I expect OSPF to auto-summarize the subnets during the redistribution. The area range command didnt help either since summarization only takes place between areas only. What should I do?
- AWhen OSPF redistributes routes from other protocols (including another OSPF process), you have to explicitly describe the summarization. In Cisco IOS 10.2 and later, they added the summary-address router subcommand to assist you.
- QWhat are the differences in memory and router performance requirements for OSPF and EIGRP? Which protocol (OSPF or EIGRP) requires less router CPU and memory to run? We are looking at implementing LS routing primarily with Cisco 2500 series routers with some Cisco 4000s mixed in. Also, we will be running DECNET and OSPF on routers running 10.0 and later code.
- AThere are no benchmarks, but I think EIGRP uses less memory than OSPF when the network is stable. Regarding CPU time, EIGRP will not use more than 50 percent of the CPU time when there are processes waiting for the CPU. If no one else is competing for the resource, then it might use 100 percent if there is a need (this is good).
If you are using Cisco 2500s in a remote location situation, you can use default routes out so you dont need to burden these routers. Or you can try making the areas stub areas, etc. Use techniques such as these to eliminate too many LSAs going around. Do route summarizations too.
- QDo Cisco routers have any quick commands that can be executed from the keyboard?
- AYes. Appendix A provides a list of the keyboard commands that can be executed for Cisco routers.
- QWhy does OSPF redistribute a static route, but not the default static route? What is the recommended way to get the default route redistributed in an OSPF-only network?
- AYou have to specify at least default-info originate and the router will announce the default into OSPF if it has one. The default-info originate always command announces the default into OSPF, even if the originator does not have a default route.
- QDoes OSPF support load balancing among serial lines?
- AYes, load balancing works in OSPF with up to four equal-cost paths, serial or otherwise.
- QWhat does it mean that OSPF is in a 2-Way/drother state?
- AIt means that you recognize that router as a neighbor, but because it isnt a DR or BDR, you didnt form an adjacency with it. Routers sharing a common segment, an Ethernet for example, that sends out hello packets and through this a designated and backup designated router are elected. The show ip ospf interface command will show the result of this election which can be DR (Designated Router), BDR (Backup Designated Router), or DROTHER (Designated Router Other).
Drother implies that on this segment a DR and BDR have already been elected. Therefore, all other routers on the segment will be in the state DROTHER. Two-way indicates that the router has seen itself in the hello packet but will not become a DR or BDR.
- QWill OSPF pick the default route based upon cost metrics? My OSPF network has connection to two different ISPs. I am generating the default route by using the default-information originate OSPF command on both of my ASBRs. Will my OSPF network routers pick the default route based on cost metrics? For example, will some of my OSPF network routers pick one of the ASBR as the default route and some the other ASBR based on the existing OSPF conditions at that given time? Please note that I am not using the always option in the default-information originate command because I want one of the ASBRs to be the default candidate when the ISPs link on one of the ASBRs goes down.
- AYes. Unless you are changing the default route to type 1 from type 2, the lower cost default route will be chosen by your entire OSPF domain over your higher cost default route.
- QI am having a problem using OSPF and subinterfaces over Frame Relay. I would like to use a multipoint subinterface over FR. Is it necessary to use fully-meshed topology?
- ANo, it is not necessary to fully mesh a FR multipoint to use OSPF. In a full mesh, OSPF can use the ip ospf network broadcast command to treat the FR cloud like a LAN (broadcast) media (9.21 and later). A partial mesh or hierarchical design can use the neighbor command to establish OSPF across non-broadcast media.
- QOSPF RFC 2178 was just released and it is not supported by Cisco yet. When will it be?
- AAccording to Cisco, the only extensions that are not supported as in 11.2/11.3 are G.5, G.7, G.8, and G.9. They should all be supported by 12.0.
|Previous ||Table of Contents ||Next |