Recently with the Spark Streaming real-time computing, on Yarn run for a long time to find Storage Memory has been slow growth, each batch may rise several M to M

Use memory leak check to get the following results


Where would like to ask about the

CodePudding user response:

Not clear, help you UP a friendship,

CodePudding user response:

I met this problem before, because at that time had not closed connection

CodePudding user response:

reference 1st floor yangjuanli response:

is not clear, help you UP a friendship,

thank you very much

CodePudding user response:

Do have write HBase program, but is the singleton pattern, has closed the connection pool, see the results of the analysis is netty created many buffersize, this is not how to understand