I needed about six months to rip about 1200 disks, because I was spending the time proofreading the tagging and fixing it and rectifying artist data for each one and browsing the web for album art. When the data coming in off the net was accurate things could go pretty quickly. The CD drive was on a long USB cable so I could rip the disks directly on the music server without having to get up from my desk to do it. Since then I've ripped a hundred or so more, but now that the intense batch copy session is done, I usually do the work on my own computer and uploaded to the server manually.
There are two external hard drives for my music collection, one exists only as a duplicate of the other. Both are connected to an old computer I use as the household music server. Every so often, after adding a few albums, I boot up the second drive, open a shell session and rsync them. When the duplication's not going on, the second drive is powered down. Everything (plus the network hardware) is on a hefty UPS so that power problems don't mess things up. All files are ALAC, since it was the path of least resistance.
I buy music on CD whenever possible. For the time being there's still nothing better than owning hard copies. If the future of cheap bandwidth, cloud storage and ubiquitous flexibly high-res audio files become the norm, I'll make the switch.