The first thing that happened was that the floodwalls and levees were overtopped.  With the floodwalls the water fell down the onshore side and eroded away the support for the wall, in most places chewing down a cavity that was about 3 ft wide on the land side of the wall.  The loss of resistance, and local factors would then cause one segment of the wall to move in as the land-side resistance failed, and the sheet piling bent into a spout shape in most places.

The major difference at the Lower Ninth Ward (where all the videos are usually shot) is that the wall was breached by a barge than ran along the wall as it went through it, knocking out and moving some 200 yards or more of wall, almost instantaneously.  Thus, instead of just a slow rising water level, in that case there was a wall of water than entered the city.  It removed all the dwellings for about four blocks carrying everything before it until it started to run out of power and dumped most of the debris (house bits, cars etc), which caused barriers that redirected some of the flow then laterally. This sudden failure and large damage zone was atypical of the city damage.  On the other hand down the delta whenever the water overtopped the levees the damage was almost total as it was whipped up by the force of the hurricane.  

Quite frankly wrong, at least for the 17th Street and Orleans Canal breaches, which flooded a majority of New Orleans.

The debris line was over a foot from the top.  Terrible Corps of Engineers design did not sink pilings deep enough, water went underneath and bubbled up on the other side.  Eventually this "tunneled" and undermined  the levees there.

In Placquemines and St. Bernard, the Corps of Engineers built Mr GO (Mississippi River Gulf Outlet) was also a Gulf of Mexico Gulf Inlet and, after having destroyed the swamps around it with salt water intrusion, let in a wave of water that over topped the levees in that side.

Just got back from dinner @ GW Fins with two Corps of Engineers engineers.  A good friend is a City of New Orleans Public Works engineer who got back in on Wednesday after Katrina and took photos before evidence was bulldozed.
 

Not to be controversial but I do have a picture from the 17th St breech, showing the overtopping scar as it was filled in during the creation of the new wall.
Without seeing that specific photo, I have talked with those engineers on the scene.  In the undermine scenario, the concrete plates would fall underneath the waves as the dirt disappeared underneath.  The intact concrete plates next to the breach showed debris lines a foot from the top.  Since all plates were within an inch of each other in height before Katrina, if it did not overtop next to the breach AND there is clear evidence of a sand boil in the yard of a neighbor, then this supports the undermining explanation.

Note, more than one breach on the 17th Street Canal (love the Quality control at the Corps of Engineers !), so there may be more than one mechanism for failure.

This is a typical overtopping scour (from the Industrial Canal). The width is about 3-4 ft and you can see the sheet piles exposed under the concrete wall.  The sheet piles are then moved forward and tilt.

This was from down at Port Sulfur. Note that the water flow has gouged out down to bedrock in front of the sheet piles.
The concrete slabs on top are then thrown forward

This is from the center of the 17th St failure. You can see the slabs fell forward and not back,

We have many photos that show similarities across the floodwall failures around NOLA and the delta, but these I can use, since they are mine, the rest will only be available when the report  is released.

I can likley get you additional photos taken on onsite Wednesday after Katrina fro 17th Street & Orleans failures.

eMail with directions (friend of mine with Public Works took them) Alan_Drake@Juno.com

I didn't know Louisiana had any "bedrock".  Most of my projects, we only hit stiff clay and oyster shells sometimes if we were lucky.   I worked a lot up at Taft and then down in Homa, Lake Charles, Lafayette, Morgan City and a lot we did only on piles set deep into muck with a build up of shell on the surface.  I thought most all of its the kind of gumbo you can't eat, clay gumbo; kinda' too thick to drink, but too thin to plow.
I may have got that wrong - I wasn't about to go down and look, though it did look like a rock slab it could have been a hard layer of clay, and it was fairly deep below the surface.
The core of levees should be clay, in order to stop water migration.
Everything in LA and East Texas is Clay Clay Clay, in varying degrees of "stiffness".  Sometimes you do see a pretty stiff layer that you'd almost think is rock, but usually its just clay with a lot of micro coral shells well embedded.