Reply
Highlighted
New Contributor
Posts: 3
Registered: ‎02-22-2017
Accepted Solution

python thrift TSocket read 0 bytes

Hi every one.

First we are using amazon EC2 for our hbase infrastructure.

We use 25 small instance to run our script (4 process in each instance), the script scan a table and make some computing before putting the new data to some other tables.

The script take some minutes before falling down saying : Exception thrift.transport.TTransport.TTransportException: TTransportException('TSocket read 0 bytes',) in <bound method Scanner.__del__ of <libs.bnet.thrift.hbase.scanner.Scanner instance at 0x7f439dddf7a0>> ignored

 

I had enabled TRACE logging in thrift server and i searched for somme error in the same instant when error happend, without success.

I don't have any idea about what is causing this. if you can help me thanks a lot.

New Contributor
Posts: 3
Registered: ‎02-22-2017

Re: python thrift TSocket read 0 bytes

Hi every one.

the problem was that thrift close the connection socket after 1 min, than i needed to augment the thrift config key ( on ms) :

hbase.thrift.server.socket.read.timeout

 

New Contributor
Posts: 3
Registered: ‎10-14-2017

Re: python thrift TSocket read 0 bytes

Could you please inform how to modify this timeout?

New Contributor
Posts: 3
Registered: ‎02-22-2017

Re: python thrift TSocket read 0 bytes

New Contributor
Posts: 3
Registered: ‎10-14-2017

Re: python thrift TSocket read 0 bytes

[ Edited ]

I cannot use the Cloudera interface as I'm running Hive server inside docker container and trying to connect to the Hive server from outside the container through python (PyHive v0.5, python v2.7.13)

New Contributor
Posts: 3
Registered: ‎10-14-2017

Re: python thrift TSocket read 0 bytes

adding these configuration settings as mentioned here (https://issues.apache.org/jira/browse/HIVE-2006):
1. hive.server.read.socket.timeout=1000
2. hive.server.read.socket.timeout=1000

in the file hive-site.xml also didn't work
Announcements