issues
search
coderplay
/
dragon
Real-time computing system based on hadoop yarn
12
stars
5
forks
source link
#48 Revisits dragon state machine
#50
Closed
coderplay
closed
12 years ago
coderplay
commented
12 years ago
Dragon jobs should failed fast whenever there is a task failed. Don't like it in MapReduce, a task failure may not cause the whole job fails.
Someone may submit a job with zero tasks, we should deal with this scenario.