As I recall, everyone was told not to make any transactions due to the fork in progress, so I guess it was a shut down of sorts.
I witnessed this pretty much as it happened, and I can tell you that the network carried on running (albeit in the forked state, so running but not really functioning). IIRC, an alert was posted with the emergency alert system in the actual Bitcoin Core client, but that probably wasn't the majority of users even back then. So some people carried on mining and transacting regardless, possibly unaware, possibly seeing whether they might be able to exploit the situation somehow .
Yeah, that was an interesting few hours. The technical scenario which led to the fork was basically bitcoin's consensus disaster scenario; such a version conflict (which is more or less what led to the fork) was theoretically possible, but unlikely and had never happened before. But it was the resolution to it which, in my opinion, contributed to the spring 2013 rally. The community came together and quickly and rationally resolved the problem through community consensus, thereby relieving the ecosystem of one of the more severe theoretical concerns.
In his tweetstorm, @patio11 seems to attack the method by which the fork was resolved. People running the big pools, Gavin and other devs, plus a bunch of industry ppl, etc, agreed that the best resolution for the integrity of the network would be if bitcoind v0.8 was not the majority version. Since BTCGuild has just upgraded to v0.8, Eleutheria agreed to downgrade to v0.7 so that the v0.7 fork would overtake v0.8. Problem solved, completely within the bounds of expected system behavior.
Nothing on the network level was "shut down"; @patio11 is being both careless and hyperbolic with his words.