bigtop-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steven Núñez <>
Subject Probably Bugs in Hive Install
Date Thu, 21 Nov 2013 03:29:41 GMT

The summary of this is: the system is reporting metastore not running (nor hive server 2),
yet hive is now working. See below email where I was going to request diagnoses help, only
to discover that the example now works, for reasons unknown.

- SteveN

-------------------------------------- Original Message -----------------------------
I’ve got a bit closer. The metastore and hive server aren’t running:

org.apache.thrift.transport.TTransportException: Could not create ServerSocket on address
org.apache.thrift.transport.TTransportException: Could not create ServerSocket on address

Because something is already bound to those ports (from netstat):

tcp        0      0     *                   LISTEN      3071/java
tcp        0      0      *                   LISTEN      2827/java

Whatever is listening on those ports doesn’t speak HTTP, and 10000 just closes a telnet
connection straight away.

Grepping through /etc doesn’t produce much of value, or anything I don’t already know:

nunez$ grep -R 9083 *
alternatives/hive-conf/hive-site.xml:  <value>thrift://localhost:9083</value>
grep: alternatives/oozie-tomcat-conf/webapps/oozie/ext-2.2: No such file or directory
default/hive-hcatalog-server:export METASTORE_PORT=9083
hive/conf.dist/hive-site.xml:  <value>thrift://localhost:9083</value>
hive/conf/hive-site.xml:  <value>thrift://localhost:9083</value>
grep: oozie/tomcat-deployment.http/webapps/oozie/ext-2.2: No such file or directory
grep: oozie/tomcat-deployment.https/webapps/oozie/ext-2.2: No such file or directory
grep: oozie/tomcat-deployment/webapps/oozie/ext-2.2: No such file or directory

I modified the hive-site.xml file based on a ‘blog post I found earlier, but the port was
bound even before that change.

So, does anyone have any ideas about who is listening to these ports and why? I would have
thought them reserved for Hive, but some Java program thinks otherwise. This of course is
assuming that my Hive failure is caused by something related:


So, what was going to appear here is the Hive failure I reported yesterday<>.
However upon repeating the commands, it now works:

hive> create table doh(id int);
Time taken: 5.344 seconds

So, I guess the question now is: Why is this working when ‘service —status-all’ reports
that the services are failed, and there’s the above two socket errors in the log files?

- SteveN

P.S. Interesting that grep is reporting some kind of ooze filesystem errors. Oozie is also
failing, but I haven’t got around to looking at it yet. I wonder if these errors are part
of the reason though.

View raw message