Yes KDB logged the same block twice.

It's to do with the fact that KDB always sees each block twice (since I changed it last year) - once from the block record from the pool software and once from it's own calculations after it gets the share from the pool software.
Normally it sees that it already had it, but in this case KDB was faster than the pool and it ended up in the ram table twice.
I'll add it to the todo list to fix that bug

The disk DB can't store exactly the same block twice so it rejected the duplicate records.
A KDB restart will fix it later - it calculates all the block stats from the data, they're not stored in the disk DB - but until then the stats will show incorrectly.