Zeppelin "sql interpreter not found"

Mon Nov 9 09:54:28 UTC 2015

Hi Andrew

I do not seem to be able to replicate the behavior where spark is not the
top interpreter in the binding list. If I encounter this again, I'll dig a
little deeper to see how this happened.

Kind regards
Merlijn

2015-11-06 19:28 GMT+01:00 Andrew Mcleod <andrew.mcleod at canonical.com>:

> Sorry Merlijn, I seem to have partially misread your reply, so to further
> clarify:
>
>
> - I have deployed (apache-zeppelin and realtime syslog analytics bundle)
> ~4 times today and each time spark has been the top interpreter in the
> binding list.
>
> - In each of our zeppelin tutorials I see an interpreter defined for
> everything that does not use the spark interpreter (with special
> consideration given to the sql paragraphs)
>
> Can you please let me know if either of these points is not consistent for
> you after a redeploy?
>
> Thanks!
>
>
> On Fri, Nov 6, 2015 at 5:22 PM, Andrew Mcleod <andrew.mcleod at canonical.com
> > wrote:
>
>> That's strange, I haven't seen anything other than spark as the
>> top/default interpreter in the list after binding (as it is the top in the
>> binding list also). Are you able to replicate this and take a screenshot?
>> Spark should definitely be the default...
>>
>>
>> Andrew
>>
>> On Fri, Nov 6, 2015 at 4:56 PM, Merlijn Sebrechts <
>> merlijn.sebrechts at gmail.com> wrote:
>>
>>> Yes, this is what I see. The default interpreter is the top one in the
>>> list (if you click on the gear icon). What the top one is after
>>> installation seems to be more or less random...
>>>
>>> 2015-11-06 16:53 GMT+01:00 Andrew Mcleod <andrew.mcleod at canonical.com>:
>>>
>>>> Hi Merlijn,
>>>>
>>>> Can you tell me if this is what you see:
>>>> http://pasteboard.co/1W9RYYPF.png
>>>>
>>>> If so, the default interpreter for the top paragraph is %md, but the
>>>> others, without a specification, will use %spark as the default (once the
>>>> save button is clicked to bind the interpreters, that is)
>>>>
>>>>
>>>> Andrew
>>>>
>>>>
>>>>
>>>> On Fri, Nov 6, 2015 at 2:38 PM, Merlijn Sebrechts <
>>>> merlijn.sebrechts at gmail.com> wrote:
>>>>
>>>>> Hi Andrew
>>>>>
>>>>>
>>>>> Thanks again for your help. The problem was that the code to create
>>>>> the table didn't specify which interpreter had to be used. The default
>>>>> interpreter was markdown, so it just printed out the lines. I'll see if I
>>>>> can create patch to the charm.
>>>>>
>>>>>
>>>>>
>>>>> Kind regards
>>>>> Merlijn
>>>>>
>>>>> 2015-11-06 15:09 GMT+01:00 Andrew Mcleod <andrew.mcleod at canonical.com>
>>>>> :
>>>>>
>>>>>> Hi Merlijn,
>>>>>>
>>>>>> I have seen this problem, but don't know the exact cause - I think it
>>>>>> has to do with the default spark contexts which are created by zeppelin
>>>>>> when it starts the interpreter, specifically the SQLContext, see   (
>>>>>> https://zeppelin.incubator.apache.org/docs/interpreter/spark.html)
>>>>>> for more details.
>>>>>>
>>>>>> Try the following in a new paragraph (no %sql interpreter) and if it
>>>>>> works, its probably an issue with the %sql interpreter context/binding:
>>>>>>
>>>>>> sqlContext.sql("SELECT * COUNT(1) from bank")
>>>>>>
>>>>>> Try restarting the interpreter (interpreters tab, then the restart
>>>>>> button next to the interpreter) - then re-run the job that creates the temp
>>>>>> table.
>>>>>>
>>>>>> i.e. this line: "bank.toDF().registerTempTable("bank")"
>>>>>>
>>>>>> If that doesn't work, try restarting zeppelin completely
>>>>>> (/usr/lib/zeppelin/bin/zeppelin-daemon.sh restart)
>>>>>>
>>>>>>
>>>>>> Andrew
>>>>>>
>>>>>> On Fri, Nov 6, 2015 at 1:59 PM, Merlijn Sebrechts <
>>>>>> merlijn.sebrechts at gmail.com> wrote:
>>>>>>
>>>>>>> Hi Andrew
>>>>>>>
>>>>>>>
>>>>>>> Thanks for your help!
>>>>>>>
>>>>>>> I just figured out my problem: for some reason I thought "blue"
>>>>>>> meant unselected and "white" meant selected. After selecting the spark
>>>>>>> interpreter the queries execute, but now I get another error.
>>>>>>>
>>>>>>> When running the hdfs tutorial notebook, I get the error "no such
>>>>>>> table List(bank);". This is strange since the "load data into Table" note
>>>>>>> executed without any errors. I get the same error when I execute the
>>>>>>> tutorial notes one by one. Any idea what I'm doing wrong now?
>>>>>>>
>>>>>>>
>>>>>>> Kind regards
>>>>>>> Merlijn
>>>>>>>
>>>>>>> 2015-11-06 14:42 GMT+01:00 Andrew Mcleod <
>>>>>>> andrew.mcleod at canonical.com>:
>>>>>>>
>>>>>>>> Hi Merlijn,
>>>>>>>>
>>>>>>>> Have you bound the interpreters to the notebook? The first time you
>>>>>>>> use the notebook, the top paragraph will be a list of interpreters. You
>>>>>>>> have to save this to be able to run anything which requires an interpreter
>>>>>>>> definition.
>>>>>>>>
>>>>>>>>
>>>>>>>> Andrew
>>>>>>>>
>>>>>>>> On Fri, Nov 6, 2015 at 1:00 PM, Merlijn Sebrechts <
>>>>>>>> merlijn.sebrechts at gmail.com> wrote:
>>>>>>>>
>>>>>>>>> Hi all
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> I'm trying to get Zeppelin working. Installing works fine, but
>>>>>>>>> when I run the hdfs notebook, the query parts fail with following error:
>>>>>>>>> "sql interpreter not found".
>>>>>>>>>
>>>>>>>>> I basically deployed the realtime rsyslog bundle
>>>>>>>>> <https://jujucharms.com/realtime-syslog-analytics/11> without the
>>>>>>>>> rsyslog and flume parts. I thought this bundle was working since I saw it
>>>>>>>>> at a demo during the summit. Any ideas to what might be wrong here?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Kind regards
>>>>>>>>> Merlijn
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Bigdata mailing list
>>>>>>>>> Bigdata at lists.ubuntu.com
>>>>>>>>> Modify settings or unsubscribe at:
>>>>>>>>> https://lists.ubuntu.com/mailman/listinfo/bigdata
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/bigdata/attachments/20151109/5462092e/attachment.html>