https://community.cloudera.com/t5/Support-Questions/hdfs-client-takes-20s-to-failover-to-alternate-namenode/m-p/269564#M206928 <P>Hi,</P><P>&nbsp;</P><P>Have you find a solution ? I have the same problem with my production cluster.</P><P>HDFS HA is implemented from several years without problem.</P><P>&nbsp;</P><P>But recently,&nbsp;we realized that a hdfs client have to wait 20 secondes when the server hosting the nn1 is shutdown. Exemple when I set the debug mode :</P><P><EM>19/08/29 11:03:05 DEBUG ipc.Client: Connecting to XXXXX/XXXXX:8020</EM></P><P><EM>19/08/29 11:03:23 DEBUG ipc.Client: Failed to connect to server: XXXXX/XXXXX:8020: try once and fail.</EM></P><P><EM>java.net.NoRouteToHostException: No route to host</EM></P><P>&nbsp;</P><P>Few informations :&nbsp;</P><P>Hadoop version : 2.7.3</P><P>dfs.client.retry.policy.enabled :&nbsp;false</P><P>dfs.client.failover.sleep.max.millis :&nbsp;15000</P><P>ipc.client.connect.timeout :&nbsp;20000</P><P>&nbsp;</P><P>Thanks for your help !&nbsp;</P><P>&nbsp;</P> Mon, 02 Sep 2019 13:42:57 GMT dupuy_gregory 2019-09-02T13:42:57Z