Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
2.7.2
-
None
-
Reviewed
Description
HADOOP-10597 added support for RPC congestion control by sending retriable 'server too busy' exceptions to clients.
However every backoff results in a log message. We've seen these log messages slow down the NameNode.
2016-03-07 15:02:23,272 INFO org.apache.hadoop.ipc.Server: Socket Reader #1 for port 8020: readAndProcess from client 127.0.0.1 threw exception [org.apache.hadoop.ipc.RetriableException: Server is too busy.]
We already have a metric that tracks the number of backoff events. This log message adds nothing useful.
The IPC Server should also allow services to skip logging certain exception types altogether.