I wish I had my camera so I could share with you the difference between the last two server rooms I was in.
I will try to describe. The first was an organizational dream. Color coordinated patch cables, wire management actually used. You could actually see the port numbers on the Cisco 4507. Even more the configs were labeled. So when you were logged into the switch you knew what was going on. Just a brief picture but the rest of the room was the same, the blades, SAN and everthing else was how I wish everyone would be.
In contrast, the next day, I was trying to upgrade some ESX servers. The cables had no rhyme or reason. The switches had no labels and the trunks and access ports were done in the most backward way I could think of. It would be too hard to describe. I vMotioned all the vm’s to another host. Shutdown the host to upgrade to 32GB of Memory. After unplugging everything and trying to slide the Dell 2950 out of the rack I discover the fibre to the server above it is running through the “handle” on the back of the 2950. So the server will only slide about 1 foot. Great, so I hook everything back up and vMotion everything off of the host above. Move the troublesome fibre, and decide to upgrade that server. I start to slide it out for the memory upgrade. The server on top of it is not on rails and starts to slide with the 2950. So I now need to get downtime (that is a physical server) to take that server down so I can upgrade the memory of the ESX servers then upgrade them to ESX 3.5 update 2.
So I would guess one of the quickest ways to eat all of my billable hours is to spend hours fighting poor datacenter work and eventually not get anything done.