Rohit,      Sorry for not being so specific in my post. is there any way to find out the process(job) which is causing most threads to Hive Meta Store in a 1000 node cluster?   Regards, Tej ... View more

Rohit,     Thanks for your response.  output of  set datanucleus.connectionPool.maxPoolSize;  is 80 ; HDP version 2.6.5.4 As you suggested lsof -p PID | grep ESTABLISHED    will provide the connections made to HMS. My question is if your cluster has 1000 nodes. If you run a any hive or spark job it will get the resources from many nodes and run on many nodes.  If you want to find out what process causing the more Threads to the HMS. How could i find out? I believe if we run a job it may cause many number of threads depends on the job. is there any way to find out what process which is causing more threads to HMS?   Regards, Tej ... View more

Let me explain my question very clearly, if Hive Meta Store getting connection failure alerts because of the more load (more connections) , how could we find what process causing the more connections to HMS.? ... View more