On connecting stuff...
Not my error, but I've seen it more than once. Shared public sector offices between different organisations. both using RFC 1918 addressing, both using OSPF area 0 (reasonable enough with smallish networks) sharing a common wiring system, but with separate hubs/switches for each organisation.
The two networks get patched together. Routing tables suddenly double in size, routing becomes extremely random, default routes go all over the place, chaos ensues.
The first time I saw it was deliberately caused by a devolved departmental "IT manager" who thought it was a good idea to have shared access to files between organisations and didn't see fit to inform anyone. As one of the networks he linked up was NHS this was spectacularly poor thinking.
The other time it was just a careless outsourced engineer who didn't know there were two organisations on the site and said to himself "funny, those two switches aren't connected to the router, I'll just patch them in".