access standy namenode slow

[weeei he]

access standy namenode slow

we have a hadoop cluster with 200 nodes. And use ha namenodes.

nn1 hadoop109  active

nn2 hadoop110  standby

after we switchover nms,  

hadoop110(nn2) is active

hadoop109(nn1) is standy

when we access hdfs://hadoop109:8020 ( hdfs dfs -ls hdfs://hadoop109:8020),  that get the response sometimes fast, sometimes slow.

I tuned the rpc parameters about dfs.namenode.handler.count and dfs.namenode.service.handler.count . the value from 105 to 150 ( >=20*ln(datanodes) ).

But the problem still occured.

Does someone have the same problem, and could give some suggestion ?