Support Questions
Find answers, ask questions, and share your expertise

Issue when trying to reach HBase thrift server from my edge server

Rising Star


I use a kerberized 2.6.3 HDP cluster comprising 3 HBase master nodes (named p21, p22 & p23).

On one of these master nodes (p23) , I started up a thrift server following explanations from :

Everything went fine and when I launch (being "hbase" user) the DemoClient from one of the 3 master nodes ("hbase org.apache.hadoop.hbase.thrift.DemoClient p23 9090 true"), it executes successfully (even when launching from the 2 master nodes that don't host the thrift server)

Anyway, when I try to execute some client code from an edge node (and from another "non hbase" user) to access my thrift server, it fails to executes properly :

I can see many instances of following error in thrift server logs :

2018-01-25 14:12:40,051 INFO [thrift-worker-0] client.RpcRetryingCaller: Call exception, tries=10, retries=35, started=48576 ms ago, cancelled=false, org.apache.hadoop.hbase.exceptions.ConnectionClosingException: Call to p21/10.XXX.XXX.XXX:16000 failed on local exception: org.apache.hadoop.hbase.exceptions.ConnectionClosingException: Connection to p21/10.XXX.XXX.XXX:16000 is closing. Call id=10, waitTime=7

And after a while, my client fails and return same kind of error...

It's worth saying that I can successfully execute "hbase shell" queries from this edge node, so I guess my "hbase-site.xml" file is correct...

Any idea why this error can happen only for thrift based queries ?

Thanks for your help