> I also had tried to increase Metron node memory, 8GB to 16GB directlty
through VBox interface

When changing the memory available, I would start over completely.

1. Run `vagrant destroy`.
2. Then edit the Vagrantfile here [1] and increase the memory from 8192 to
your desired value.
3. Then try it again; `vagrant up`.


> Otherwise, what logs i can check to see if Kafka correctly receive data
from NiFi ?

Use the `kafka-console-consumer` to check if any data is landing in Kafka
from Nifi.


[1]
https://github.com/apache/metron/blob/master/metron-deployment/vagrant/full-dev-platform/Vagrantfile#L48





On Wed, Jul 19, 2017 at 3:22 PM, Sebastien Lehuede <[email protected]>
wrote:

> Nick,
>
> I had tried to install NiFi on an other server, but result is the same. I
> can see my data pass through NiFi, but Elasticsearch index is not created.
> (My server successfully reach my Metron node on port 6667).
>
> I also had tried to increase Metron node memory, 8GB to 16GB directlty
> through VBox interface. « top » command show 16GB memory on server but
> Ambari show only 8GB, there is others actions to do to increase Metron node
> memory ?
>
> Otherwise, what logs i can check to see if Kafka correctly receive data
> from NiFi ?
>
> Sebastien.
>
> (Sorry for my English)
>
> Le 19 juil. 2017 à 16:56, Nick Allen <[email protected]> a écrit :
>
> That being said, I would highly suggest experimenting in a cluster; for
> instance using our automated AWS deployment mechanism [1].  This will
> spin-up a full cluster (a totally open, unsecured cluster that you need to
> secure yourself) in AWS.
>
> Of course, this will cost you money, but should you have the funds, this
> is the way to go.
>
> [1] https://github.com/apache/metron/tree/master/metron-depl
> oyment/amazon-ec2
>
> On Wed, Jul 19, 2017 at 10:52 AM, Nick Allen <[email protected]> wrote:
>
>> Yes, you can try to increase the amount of memory, as long as you have
>> enough physical memory on your machine.  I don't know of any hard memory
>> limits on that VM.
>>
>> On Wed, Jul 19, 2017 at 9:44 AM, Lehuede sebastien <[email protected]>
>> wrote:
>>
>>> Hi Nick,
>>>
>>> Yes, i run Nifi directly in "Full Dev VM". I will try to reinstall and
>>> reconfigure Nifi in an other VM or is it possible to simply increase amount
>>> of Memory for "Full Dev VM ?
>>>
>>> Thanks,
>>>
>>> Sebastien.
>>>
>>> 2017-07-19 15:13 GMT+02:00 Nick Allen <[email protected]>:
>>>
>>>> Are you also running Nifi in your "Full Dev" VM?  If so, I would assume
>>>> that you just don't have enough resources to run all that on one VM.  Just
>>>> running Metron's components alone in "Full Dev" has taken considerable
>>>> effort to tune.  Adding Nifi would probably cause some weird errors due to
>>>> resource constraints.
>>>>
>>>> On Wed, Jul 19, 2017 at 8:50 AM, Lehuede sebastien <[email protected]
>>>> > wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> I have successfully deployed Metron 0.4.0 full dev platform. After
>>>>> that, i have tried to add a new telemetry data source with this tutorial :
>>>>> https://cwiki.apache.org/confluence/display/METRON
>>>>> /Adding+a+New+Telemetry+Data+Source. But when i start NiFi processor,
>>>>> i don't see any logs in Storm UI and no Elasticsearch index are created.
>>>>>
>>>>> I had no error, in NiFi interface when i start processor everything
>>>>> seem to be OK and in my Ambari interface all services are started, except
>>>>> "Metron REST".
>>>>>
>>>>> I have setting up a very simple grok parser
>>>>> ("%{DATA:toto}\:%{DATA:tata}") for testing purpose. My goal is to parse a
>>>>> simple text file who contains things like "bob:alice".
>>>>>
>>>>> How can i troubleshoot this problem ? With which logs ?
>>>>> Has anyone ever had this problem ?
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Sebastien.
>>>>>
>>>>>
>>>>
>>>
>>
>
>

Reply via email to