This is true for 7.7 as well as beyond:
For the first case when many replication jobs are running at once, but few appear to be making progress, the system may have reached a state of "gridlock." The solution is simply to decrease the Max I/O strea...