According to the tripartite view of knowledge, knowledge is equivalent to justified true belief. Formulaically, it is the view that S knows P if and only if:
S believes PP is trueS is justified in believing P
Gettier-style cases are examples of beliefs that are both justified and true, but that we would nonetheless not want to deem as knowledge. An example would be the following:
'Alice walks into a bar to watch the Wimbledon final between Alan and Steven. She sees Alan win the final point, and lift the trophy. She therefore forms the belief that Alan won the final. When she returns home, she watches Alan lift the trophy again, but this time Steven hits the ball into the net. In the bar, she saw Alan hit a winner. As it turns out, the people in the bar were, for whatever reason, watching a recording of last year's final. The recording ended a few minutes before the current year's final ended.' In this example, Alice had a belief that was true (Alan did win the final), and justified (on the television, she saw Alan win the final point and lift the trophy). Nonetheless, we wouldn't want to say that she knew that Alan won the final. After all, she formed the belief on the basis of a recording of the previous year's final.
All Gettier cases have the same format: 1. S believes proposition P based on justification J. P turns out to be true, but J is flawed.