I am even having the almost same versions of storm (0.9.1) and kafka. And
my topologies were also facing the same issue. When I ran the worker
command directly, I came to know that somehow hostname was wrong in the
configuration passed to the workers. So, I fixed that in storm config and
my topology worked after that. However, now again it has stuck with same
"still hasn't started" error message and in my case now the error in
running the worker command is "Address already in use" for supervisor port.
So, what is the error when you directly run the worker command?
On Tue, Aug 26, 2014 at 9:39 PM, Naga Vij <nvbuc...@gmail.com> wrote:
> I fail to understand why that should happen, as testing with LocalCluster
> goes through fine.
>
> I did a clean fresh start to figure out what could be happening, and here
> are my observations -
>
> - fresh clean start: cleanup in zk (rmr /storm), and /bin/rm -fr {storm's
> tmp dir}
> - used local pseudo cluster on my mac
> - nimbus process started fine
> - supervisor process started fine
> - ensured toplogy works fine with (the embedded) LocalCluster
> - topology was then submitted to local pseudo cluster on my mac ; that's
> when I see ``still hasn't started`` messages in supervisor terminal
> window
>
> When submitting topology to local pseudo cluster, had to add jars to
> overcome these ...
>
> Caused by: java.lang.ClassNotFoundException: storm.kafka.BrokerHosts
> Caused by: java.lang.ClassNotFoundException: kafka.api.OffsetRequest
> Caused by: java.lang.ClassNotFoundException: scala.Product
>
> Above were overcome by adding these to lib dir -
>
> storm-kafka-0.9.2-incubating.jar
> kafka_2.10-0.8.1.1.jar
> scala-library-2.10.1.jar
>
> I have tried the command in log as well ; hasn't helped.
>
> What am I missing?
>
>
> On Mon, Aug 25, 2014 at 11:41 PM, Vikas Agarwal <vi...@infoobjects.com>
> wrote:
>
>> >> dd7c588e-5fa0-4c4b-96ed-de0d420001e9 still hasn't started<<
>>
>> This is the clue. One of your topology is failing to start. You must see
>> the worker command before these logs in the same log file. Just try to run
>> those directly on console and it would show the exact error.
>>
>>
>> On Tue, Aug 26, 2014 at 11:45 AM, Naga Vij <nvbuc...@gmail.com> wrote:
>>
>>> Hello,
>>>
>>> I am trying out Storm 0.9.2-incubating pseudo cluster (on just one box)
>>> on these two systems -
>>>
>>> > cat /etc/redhat-release
>>> CentOS release 6.3 (Final)
>>>
>>> and
>>>
>>> > sw_vers
>>> ProductName: Mac OS X
>>> ProductVersion: 10.9.2
>>> BuildVersion: 13C64
>>>
>>> After starting supervisor, I notice it is not listening on the
>>> configured port (6700) -
>>>
>>> > nc -zv localhost 6700
>>> nc: connectx to localhost port 6700 (tcp) failed: Connection refused
>>>
>>> When I submit topology, I see this scrolling message in the terminal
>>> window for supervisor -
>>>
>>> 23:11:44.532 [Thread-2] INFO backtype.storm.daemon.supervisor -
>>> dd7c588e-5fa0-4c4b-96ed-de0d420001e9 still hasn't started
>>>
>>> I don't see any worker id in UI. No error in logs.
>>>
>>> Any idea what could be happening?
>>>
>>> Thanks in advance.
>>>
>>> Naga
>>>
>>>
>>
>>
>> --
>> Regards,
>> Vikas Agarwal
>> 91 – 9928301411
>>
>> InfoObjects, Inc.
>> Execution Matters
>> http://www.infoobjects.com
>> 2041 Mission College Boulevard, #280
>> Santa Clara, CA 95054
>> +1 (408) 988-2000 Work
>> +1 (408) 716-2726 Fax
>>
>>
>
--
Regards,
Vikas Agarwal
91 – 9928301411
InfoObjects, Inc.
Execution Matters
http://www.infoobjects.com
2041 Mission College Boulevard, #280
Santa Clara, CA 95054
+1 (408) 988-2000 Work
+1 (408) 716-2726 Fax
