Friday was a bad hair day. Let us not talk about Friday.
…
Actually, lets talk about Friday.
Okay, I'm not going into details about what happened Friday (it was my fault for not fully testing the configuration [1], but really, what did Cisco [2] do to break DHCP (Dynamic Host Configuration Protocol) in IOS [3] 12.3?) but I do want to talk about the [DELETED-insane-DELETED] [DELETED-incipid-DELETED] [DELETED-stupid?-DELETED] [DELETED-ill-thought out-DELETED] … I'm not sure how to diplomatically state what I think this is.
I think this stems from an incident last week when F (a company that's also a customer of ours) experienced an outtage. I suspect that everyone suspected their consumer grade router that was suspected of going out (resetting it seemed to have fixed the issue, which points to the base problem, but it's not what you think it is) and Smirk thought that a Cisco router (that would connect us to them via Ethernet) that could switch over to a DSL (Digital Subscriber Line) connection as a backup was a good solution (I'm guessing here; I don't know the actual details that lead to this solution).
And that right there is the problem I have with this scenario.
Let's see … our Ethernet connection to F goes down (I can't see how it can, seeing how Cisco routers rarely fail, and we have a cable that runs down the hall from our Data Center to their office). Their Cisco router sees this, and thus the default network route through us (populated via OSPF (Open Shortest Path First)) is removed, and the backup default route through the DSL (static route with a higher metric than the OSPF default route) kicks in. But since the DSL is being provided via another party, F's public IP (Internet Protocol) address suddenly changes. Any ongoing TCP (Transmission Control Protocol) connections they might have had are now gone.
I'm not even going to mention the email issues that come up.
The change-over is noticable, and annoying enough to possibly prompt a call to us asking us what the aitch-ee-double-hocky sticks is going on (or even worse, what the fudge is going on).
And that's the best case scenario!
Most likely is a routing issue with our upstreams that cause the outtage (and affecting not only F, but us and all our customers). F's connection to us is still fine and thus, the Cisco router sitting in their office has no reason to switch over to the DSL connection. It would have to be done manually and well … that's a call to us.
Then there's the fact that they have 8 static IPs through their DSL provider because, hey, why not? They only have one static IP through us. I don't see this as being a Good Thing™.