Re: Spark-sql USING from Zeppelin?

2016-03-11 Thread Felix Cheung
Zeppelin is always launching the interpreter group in a separate process, so that part would be he same either way. About the root cause of your issue, try putting everything into a single line, but I suspect that wouldn't be it. It might be something importing library or something. _

Re: Spark-sql USING from Zeppelin?

2016-03-11 Thread Adam Hull
Cool, I'll try that. One more hole in my understanding if you've got the patience for it: If the SQLContext.sql(SQLContext.scala:725) error were thrown from the separate SparkSubmit process, will the JVM stitch together the stack traces across processes? The org.apache.zeppelin.spark.SparkSqlInte

Re: Spark-sql USING from Zeppelin?

2016-03-11 Thread Felix Cheung
Not from the stack, I think the best way is to run jps -v You should see a process SparkSubmit if it is running the one from your spark home. _ From: Adam Hull Sent: Friday, March 11, 2016 2:06 PM Subject: Re: Spark-sql USING from Zeppelin? To: Thanks

Re: zeppelin server hangs

2016-03-11 Thread Skanda
The server does not accept new sessions. The UI page seems to be loading eternally before I killed and restarted the server. On 12-Mar-2016 12:03 AM, "Felix Cheung" wrote: > It looks like it might have timeout or disconnected? > > > > > > On Thu, Mar 10, 2016 at 3:39 AM -0800, "Skanda" < > skanda

Re: Spark-sql USING from Zeppelin?

2016-03-11 Thread Adam Hull
Thanks Felix! I've been struggling to wrap my head around *which * SQLContext.scala file is being executed in this stack. I've set SPARK_HOME="/usr/local/opt/apache-spark/libexec" in my zeppelin-env.sh file, but I also see ./spark in my zeppelin install directory. I'd image if Zeppelin is actuall

Re: Spark-sql USING from Zeppelin?

2016-03-11 Thread Felix Cheung
As you can see in the stack below, it's just calling SQLContext.sql()org.apache.spark.sql.SQLContext.sql(SQLContext.scala:725) at  It is possible this is caused by some issue with line parsing. I will try to take a look. _ From: Adam Hull Sent: Friday, March 11

Spark-sql USING from Zeppelin?

2016-03-11 Thread Adam Hull
Hi! This whole ecosystem is pretty new to me. I'd like to pull JSON files from S3 via the spark-sql interpreter. I've got code that's working when I run `spark-sql foo.sql` directly, but it fails from a Zeppelin notebook. Here's the code: ``` %sql CREATE TEMPORARY TABLE data USING org.apache.s

Re: zeppelin server hangs

2016-03-11 Thread Felix Cheung
It looks like it might have timeout or disconnected? On Thu, Mar 10, 2016 at 3:39 AM -0800, "Skanda" wrote: Hi The zeppelin server hangs after couple of days with the following exceptions in the server log. After a restart everything looks good. I understand it is a socket connection