Backup Routers

[jwilmer]

I have a customer who purchased 6 Cisco 2610 routers to act as on-site spares for the existing routers.

What should I look at on these new routers to make sure they will work 100%. The existing routers are also 2600s

I am looking at RAM, Flash, IOS ver. Is there anything else that could be a show stopper.

Thanks

[thisisshanky]

You might want to cut and paste the configs so that they really act like plug and play routers. During a down time, you can swap out the spare router and see if everything works great..

[a.awan]

I would also verify that all hardware is working. There are rare situations when a WIC or an NM can turn out to be faulty and you definitely do not want to be troubleshooting that during a crisis situation.

[vcjones]

Step one: set proper expectations! to begin with, it is physically impossible to guarantee 100%, there is always a finite chance of failure no matter what you do. While you can reduce that probability, you cannot eliminate it.

Step two: Work out the process that will be used to execute the replacement process. Note that this must include the ability to determine which router needs to be replaced (not always obvious), how to configure the new router to match the one it is replacing, how to install it (including such issues as who has the keys to the wiring closet and how to get them at 3:00 AM on a holiday weekend). For routers installed in packed racks, can the faulty router be removed and a new router installed without disrupting the connections to other devices?

Step 3: Maximize the probability of the replacement actually working by putting them in use beforehand to make sure that all the interfaces and components work. For example, I've seen routers with NVRAM which won't hold a configuration through a power cycle, and others with interfaces which pass self test but won't drive a T1 more than 10 feet. Having a pretested, known good replacement simplifies the troubleshooting when you make the replacement and the line still does not come up. (Return to step 1: how to determine which router/component/line has actually failed).

Step 4: Design the network so that there are no single points of failure. That is, loss of a line, router, or other component may degrade performance but will not stop any critical applications. This has twin advantages: first, it takes the pressure off the replacement job so it is more likely to be done correctly. Second, it puts the "backup" spares on line where they are continuously tested and can be replaced as soon as they fail rather than waiting until they are put into use.

Step 5: If you get to step 4, you will need to also invest in the monitoring and management tools so that failures can be detected before they impact operation (you would be amazed at how many networks with redundant components are running on the backup components with no one the wiser).

Good luck and have fun!

Vincent C Jones

www.networkingunlimited.com