I have enabled HDFS name node ha(nn1, nn2), and the ranger is installed on original name node(nn1).
Now I launch a hdfs namenode failover operation, the active namenode switches to the nn2 successfully. But the user sync plugin of Ranger is installed on nn1, cause the auth problem.
Although I solve the problem by copy user and group info from nn1 to nn2, but, I think there should be another easier way to cover the hdfs ha switch situation.
Is there any better solution for this?