I thought maybe the status 33554433 denoted a public version that was in the process of editing, but further testing negates that theory. Sure would be nice to get an answer to this question. Just out of curiosity.
Did you ever find an answer for this?
We contacted ESRI Support and they walked us through manually deleting the entry from the states table. We haven't had any more issues with it.
George, here is what I got from ESRI support on what 33554433 means in sde.versions.status:
"In addition to the representing the Access type of versions [public protected, private], these numbers also reflect other versions editing processes. For example, when performing reconcile and post of one a Child version (Public access level) with the Default version, the value in the STATUS column for the Child version changes to '33554433'. It will change to a different number if working with Geodatabase Replication. Hence, the use of the '33554433x’ number sequences is an expected behavior."
With that info, our rec and post history and the values of status in our sde.versions table, my theory is that 33554433 is a public version that has been posted to the default version. 33554434 is a protected version that has been posted to the default version. A grandchild version of default that has been posted to the parent version but not the actual default version will still show the single digit status codes (0, 1, or 2) for private, public and protected.
If anyone is interested, the way to determine the access code from any of these bit codes is below... or just call the sql function, SDE_get_version_access(@status, @version_owner) returns char.
@status = @status - floor (@status / 4) * 4