What are the potential causes for the "block not found on disk" error?

We run several bitcoin core nodes and load balance requests across them all.

We occasionally hit the error "Block not found on disk" shortly after a new block gets confirmed, when looking up the contents of this new block. The error generally disappears when we retry after a few seconds.

Should I be worried about this error, or is it expected, since we are load-balancing requests across multiple nodes? My thinking is that possibly one of our nodes has seen the block and downloaded all of its contents, but another one of our nodes has not yet (and therefore throws the error)