BDR fails to retrieve Source Cluster information when adding schedule | SocketTimeoutException

0 votes
3 views
asked Aug 28, 2017 in Hadoop by admin (4,410 points)
Summary

 

Applies To
  • BDR
  • Network
Symptoms

Peer connection was added successfully, but attempting to add a new HDFS or Hive Replication Schedule fails to retrieve Source Cluster information; the Source Cluster is not available to select in the drop-down.  

The /var/log/cloudera-scm-server/cloudera-scm-server.log on the Destination Cluster Cloudera Manager Server host will log the following:

Caused by: java.net.SocketTimeoutException: SocketTimeoutException invoking https://<Source_CM_Server>:7183/api/v6/clusters?view=EXPORT: Read timed out
 

 

Cause

This issue may occur if larger response packets are not able to be transmitted from the Source Cluster Cloudera Manager Server to the Destination Cluster Cloudera Manager Server.  One possible cause is low MTU size between the two hosts.  If may be anywhere along the network.  It could be the Network interfaces, network switches, etc.

Some tests that can be used to verify MTU size might be the cause:

1) Does the following curl command from the Destination CM Server succeed?
- curl -k -u user:password https://<Source_CM_Server>:7183/api/v6/clusters?view=EXPORT
‚Äč
2) Does the following curl command from the Souce CM Server succeed?
- curl -k -u user:password https://<Source_CM_Server>l:7183/api/v6/clusters?view=EXPORT

3) Does the following curl command, for a smaller response, from the Destination CM Server succeed?
- curl -k -u user:password https://<Source_CM_Server>:7183/api/v6/clusters
 

Instructions

Contact the internal Network Team to confirm MTU size is set to 1480 or above and raise the size if necessary.

Please log in or register to answer this question.

...