You should check the log files to see the complete stack trace of the
exceptions. That should held to identify the problem.
-Matthias
On 05/15/2015 09:52 PM, Hadi Sotudeh wrote:
> I've three Vms.
> Now, I've run the commands you've said.
> -
Google is your friend in this case ;)
The problem is not related to Storm. You need to reconfigure your OS.
https://confluence.atlassian.com/display/FISHKB/java.io.IOException%3A+error%3D12,+Cannot+allocate+memory
https://stackoverflow.com/questions/1124771/how-to-solve-java-io-ioexception-error
I just want to add, that duplicates in the result are easily possible.
Let's say, the last bolt (ie, a sink), writes some tuples to a file
before acking it. If the write is successful and the bolt fails for some
reason before the ack goes through, the tuple will be replayed and
written to the file
Hi,
Storm processes multiple tuple in an "overlapping" manner, ie, emitting
from spout, network transfer, processing at bolt is fully pipelined and
multiple tuples are in the pipeline at the same time. Additionally, this
pipeline contains multiple buffers and thread transferring the tuple
from buf
You can put your dependency jars into Storm's jar folder (eg
/opt/storm-0.9.4/lib/).
-Matthias
On 05/21/2015 04:16 PM, rajesh_kall...@dellteam.com wrote:
> *Dell - Internal Use - Confidential *
>
> I am aware of the local mode, was looking for alternatives of on the
> cluster testing.
>
>
>
Hi,
I think it is a ticky problem you want to solve. The TopologyContext
object does not give enough information to get it done. Maybe you can
get it done, by implementing a custom scheduler.
And example how to implement a custom Scheduler is given here:
https://xumingming.sinaapp.com/885/twitter
It can. You can call collector.emit() any number of times during a
single .nextTuple() call.
However, it might be bad practice to do it. In your case, I would no
recommend to read the whole file in a single .nextTuple() call, but emit
only a single tuple.
-Matthias
On 05/22/2015 06:27 AM, prasad
Dear Storm community,
we would like to share our project Aeolus with you. While the project is
not finished, our first component --- a transparent batching layer ---
is available now.
Aeolus' batching component, is a transparent layer that can increase
Storm's throughput by an order of magnitude
Put our Scheduler into a jar and put the jar into the lib folder of your
Storm installation (eg, /opt/storm-0.9.3/lib).
-Matthias
On 05/28/2015 12:40 PM, Franca van Kaam wrote:
> As indicated in the tutorial I put this in the storm.yaml file of the
> nimbus node:
> |storm.scheduler: ||"storm.Ca
Hi,
you build the jar in the wrong way:
2) You need to package the class file, not the source file
2) "CameraScheduler" in in package "storm". Thus
"CameraScheulder.class" must in in directory "storm" within the jar
The correct command would be:
jar cvf CameraScheduler.jar storm/CameraScheu
I thinks it is a good idea to drop Java 6. It reached it's life cycle
already 2 year ago.
-Matthias
On 06/01/2015 08:37 PM, P. Taylor Goetz wrote:
> CC user@
>
> I’d like to poll the community about the possibility of dropping support for
> Java 1.6 in the Storm 0.10.0 release. To date, we hav
Hi,
You are looping within "nextTuple()" to emit a tuple for each lines for
the whole file. This is "bad practice" because the spout is prevented to
take "acks" while "nextTuple()" is executing. I guess, that is the
reason why your tuples time out and fail.
You should return from "nextTuple()" af
ot;data")) {
>>
>> streamId = tuple.getStringByField("streamId");
>> time = tuple.getStringByField("timestamp");
>> value = Float.parseFloat(tuple.getStringByField("value"));
>>
>> univer
ach emit?
> Add return; after emiting tuple os something else i haven’t understood?
> Could you post a code sample?
>
>> On 5 Ιουν 2015, at 14:44, Matthias J. Sax
>> mailto:mj...@informatik.hu-berlin.de>>
>> wrote:
>>
>> Hi,
>>
>> I don
I would set the needed JVM arguments in storm.yaml file. This must be
done on every worker node.
worker.childopts: "-Xmx4096m"
( or maybe
supervisor.childopts: "-Xmx496m" )
-Matthias
On 06/05/2015 05:11 PM, Prakash Ramesh Dayaramani wrote:
> Hi All,
>
> I am trying to deploy jar
One comment: The suggestion to use a single worker to avoid overhead is
basically right. It only has the drawback of coarse grained
fault-tolerance -- if the worker JVM goes done, be one bad behaving
spout/bolt, all other spouts/bolts die, too. Also keep in mind, that a
worker will only process spo
As the name suggests (parallelism_hint), it is the number of parallel
spout instances you want to start. Of course, the UDF code must be
parallelizable, eg, different instances should emit different data. If
the UDF code is not parallelizeable (eg, the UDF reads a single file),
using a parallelism
Hi,
Storm has no support for this natively. However, you can implement a
custom scheduler. See
https://xumingming.sinaapp.com/885/twitter-storm-how-to-develop-a-pluggable-scheduler/
-Matthias
On 06/14/2015 11:06 AM, Jiaming Lin wrote:
> Hi
>
> I am just starting to learn Storm and streaming
Hi,
the idea from Mike and Nathan does not apply to your problem because in
your case the different execution times do not depend on the tuples but
on the executors. Thus, on the producer side you cannot separate "slow"
tuples from "fast" tuples.
If you can identify the "slow" executors, you can
Hi,
as far as I understand it, it is a round-robin scheduling over all
available workers.
-Matthias
On 06/14/2015 01:31 PM, Franca van Kaam wrote:
> Hello,
>
> I am trying to understand how the default scheduler distributes the
> tasks. Are they evenly distributed throughout the cluster or do
Just want to clarify: The number of task is not the number parallel
running bolt instances (called executors, which are threads). So I don't
understand why you don't want to start with the maximum number of tasks?
There should be almost no overhead if you have more tasks than executors
(executors c
rm.
>
> Thanks a lot.
>
> On 19/06/2015 6:59 pm, "Matthias J. Sax" <mailto:mj...@informatik.hu-berlin.de>> wrote:
>
> Just want to clarify: The number of task is not the number parallel
> running bolt instances (called executors, which are thread
).
-Matthias
On 06/19/2015 01:01 PM, Harshit Gupta wrote:
> That's what. I want to have an arbitrary degree of parallelism. I don't
> wish to hard code it. The current release doesn't allow that, isn't it ?
>
> On 19/06/2015 8:55 pm, "Matthias J. Sax" <
>> So, it appears the expectation is to overprovision the number of tasks,
>> start with minimal number of executors, and then grow executors to
>> achieve parallelism as workload increases. Is this right ?
Yes. This is correct.
However, over provisioning the number of tasks does not result in
ov
ess we could do something like S4 where every key got a new bolt
> instance, but then they had a lot of issues with check-pointing all
> of these bolt instances and swapping them out. They also didn't
> allow for pluggable groupings. Everything was keyed grouping.
>
I don't see any in-balance. The value of "Executed" is 440/460 for each
bolt. Thus each bolt processed about the same number of tuples.
Shuffle grouping does a round robin distribution and balances the number
of tuples sent to each receiver.
I you refer to the values "capactiy", "execute latency"
mailto:ncle...@gmail.com>> wrote:
>
> Also to clarify, unless you change the sample frequency the counts
> in the ui are not precise. Note that they are all multiples of 20.
>
> On Jun 23, 2015 7:16 AM, "Matthias J. Sax"
> <mailto:mj...@in
ne guide us on how to build such this custom grouping?
>
>
>
> On Wed, Jun 24, 2015 at 1:40 PM, Matthias J. Sax
> mailto:mj...@informatik.hu-berlin.de>>
> wrote:
>
> Worried might not be the right term. However, as a rule of thumb,
> capacity should
ount the loads of
> the bolts it sends to? Is it possible to modify it to not include a key?
>
> Alternatively, If I partialkeygroup on a unique key would that balance
> my load?
>
> On Thu, Jun 25, 2015 at 2:24 PM, Matthias J. Sax
> mailto:mj...@informatik.hu-berlin.de>>
>
t?
>
> Thanks and Regards
> Aditya Rajan
>
> On Thu, Jun 25, 2015 at 6:05 PM, Matthias J. Sax
> mailto:mj...@informatik.hu-berlin.de>>
> wrote:
>
> No. This does not work for you.
>
> PartialKeyGrouping does a count based load balancing. Thus, it is
I would recommend to write a unit test for each spout/bolt and an
integration test for the whole topology using LocalCluster.
-Matthias
On 06/29/2015 07:01 AM, Rakeshsharma PR wrote:
>
>
>
>
> Greetings!!
>
>
>
>
>
> Hi
>
>
>
> I am new to storm. I am trying to wri
Hi,
I guess you refer to Apache Storm (not Spark ;))
In Storm, there is no special support to notify your client (as far as I
know). The right place to implement the notification should be the
method Spout.ack(). I guess you are using a provided KafkaSpout. Thus,
you could extend KafkaSpout with
https://storm.apache.org/documentation/Fault-tolerance.html
On 07/06/2015 07:47 AM, Thilina Rathnayake wrote:
> I recently got to know about storm and I find it really interesting. I
> want to know
> what happens when a storm supervisor goes down.
>
> I am going through the following article to
Hi,
this sounds weird... Storm should apply back pressure an slow down the
spout if bolts cannot keep up... Have you tried in increase the dop of
the bolt? Which bolt is the problematic one? NumberAvg or TimeGlobalAvg?
You might also checkout out `max spout pending` property to solve the
problem:
Hi Charlie,
yes, if you want to use back preassure, you need to use message-ids for
tuples in spouts and in bolts, anchor emitted tuples (input tuples are
anchors) and ack processed tuples.
On the other hand, I was wondering, if you did try to increase the
parallelism of NumberAvgBolt to avoid th
ges with my example storm chain and
> when I used a complex chain (Asn.1 Decoding and AvroEncoding), i received any
> ack messages and i i don't understand why ?
>
> Best regards,
> Charlie
>
>
> De : Matthias J. Sax
> Envo
What Spout do you use? Failing tuples result in back-calls to
Spout.fail(). If you use your own Spout implementation, you need to
overwrite this method. The default implementation does nothing. Or do
you already use a (so-called) reliable Spout?
-Matthias
On 07/15/2015 07:37 AM, Rahul wrote:
>
Hi,
I recently get those error messages back, when sending to
us...@storm.apache.org
Does anyone experience the same problem?
Thanks!
-Matthias
Forwarded Message
Subject: DELIVERY FAILURE: Error transferring to GAPAR017/SRV/SOCGEN
mail.box; Maximum hop count exceeded. Messa
Your jar file contains two copies of "defaults.yaml". You need to make
sure that there is at max one.
Do you include "storm-core.jar" in your own jar? For this case, exclude
"defaults.yaml" that is contained in "storm-core.jar"
-Matthias
On 07/15/2015 02:08 PM, charlie quillard wrote:
> Hi,
>
Using Eclipe export will package "storm-core.jar" and it's dependencies
into you user jar. "storm-core.jar" contains the duplicate default.yaml
class. You should not include "storm-core.jar" and its dependencies
after all.
You might try "export -> Java -> jar". This only packages your own code.
If
Hi,
using declareStream() does not necessary declare a direct stream. There
are 4 methods:
OutputFieldsDeclarer.declare(Fields)
OutputFieldsDeclarer.declare(boolean, Fields)
OutputFieldsDeclarer.declareStream(String, Fields)
OutputFieldsDeclarer.declareStream(String, boolean, Fields)
To declare
Can anyone take care of this, please? Emails bouncing back each time on
user@... (this is kind of annoying)
Thank a lot! I really appreciate it!
-Matthias
Forwarded Message
Subject: DELIVERY FAILURE: Error transferring to GAPAR017/SRV/SOCGEN
mail.box; Maximum hop count exceed
If the call to "execute()" does loop infinitely, yes; the whole
computation stops (or might fail completely with an exception if all
buffers are full -> OutOfMemoryException)
-Matthias
On 07/21/2015 02:05 PM, Ganesh Chandrasekaran wrote:
> So let’s say we have a single threaded topology with sin
It's the ID of the task that emitted the tuple.
On 07/23/2015 12:08 AM, Kashyap Mhaisekar wrote:
> Hi,
>
> What does the number 3 in the text highlighted indicate -
> ---
> backtype.storm.daemon.executor - Processing received message
> source:* f**etchoffercount:3*, stream: defaul
>
Storm does not automatically rebalance (ie, it is kind of disabled by
default). Rebalancing has to be triggered manually (for example via
command line: "storm rebalance ")
-Matthias
On 07/23/2015 01:16 AM, Thilina Buddhika wrote:
> Hi,
>
> Is there a way to disable rebalancing in Storm? Also und
Hi,
I am not sure, but I doubt that BaseBasicBolt does ack automatically.
Process-Latency should be lower, if a tuple is acked before execute
finishes. So if you ack within execute, this should be normal.
For complete latency at spout you are right. If it should 0, it raises
the question if you en
Hi,
yes, it is correct to use Spout.ack() to throttle the ingestion rate.
However, you must also use Spout.fail(), in case tuples time-out or are
failed manually within a bolt. Maybe, that is the reason you are
"missing acks".
(I personally doubt, that Storm looses any tuples or ack/fail messages
Did you package all your classes correctly in your jar file that is
submitted to Nimbus?
-Matthias
On 07/29/2015 09:39 AM, Vamsikrishna Vinjam wrote:
> i can run the topologies which are in storm examples..but i can the
> topologies which i have written if i try to run them i getting error like
>
It the class "storm.starter.util.StormRunner" contained in the your
kafka.jar file? Of not, you need to add it.
-Matthias
On 07/30/2015 10:51 AM, Vamsikrishna Vinjam wrote:
> im trying to run my storm jar file ..but im getting error
>
> Exception in thread "main" java.lang.NoClassDefFoundError:
Having more tasks than executors is only helpful, if you want to change
the parallelism of a spout or bolt during runtime (by using command
"storm rebalance"). If you do not want to change parallelism there is no
advantage in having more tasks than executors.
-Matthias
On 07/30/2015 10:24 AM, 鄢来琼
Storm can place multiple executors into a single worker. Thus, there is
no deployment problem, as long as your hardware can handle it.
-Matthias
On 08/05/2015 03:18 AM, TSD-贾宏超 wrote:
> Hi everyone,
> What will happen if Storm doesn't have enough worker slots to run a topology?
> In my cluster, w
You an implement a custom scheduler to get this done. See an example here:
https://xumingming.sinaapp.com/885/twitter-storm-how-to-develop-a-pluggable-scheduler/
-Matthias
On 08/07/2015 08:40 AM, 이승진 wrote:
> AFAIK, there's no way to deploy a topology into certain type of machines.
>
>
>
> Let
IMHO, it's a question about fault-tolerance.
If you have a single worker per node per topology, the impact in failure
case (ie, rack going down) on a topology is low. Of course, all
topologies using this failure rack are effected.
If you use multiple workers for a single topology on the same
supe
Hi,
this is a Zookeeper setting (and not a Storm parameter). You need to
update ZK config. For example, /opt/zookeeper/conf/zoo.cfg
-Matthias
On 08/26/2015 04:38 PM, Lina FAHED wrote:
> Hello,
>
> i’m new in Apache Storm, i have a problem that the Zookeeper tick time is
> very small for the tre
Hi,
this is a Zookeeper setting (and not a Storm parameter). You need to
update ZK config. For example, /opt/zookeeper/conf/zoo.cfg
-Matthias
On 08/26/2015 04:38 PM, Lina FAHED wrote:
> Hello,
>
> i’m new in Apache Storm, i have a problem that the Zookeeper tick time is
> very small for the tre
-Matthias
On 08/27/2015 10:28 AM, Lina FAHED wrote:
> Hi,
>
> as the Zookeeper is embedded in the Storm version, so, i didn’t found the
> zookeeper in /opt/
> so, how to access it ?
>
> Thanks,
>
> Lina
>> Le 27 août 2015 à 10:19, Matthias J. Sax a
>> écri
orm
> release i installed,
> i didn’t found a path for ZK in order to change its configurations.
>
> do you think that it would be better to turn on a cluster mode ?
>
> thanks
>
> Lina
>
>> Le 27 août 2015 à 10:31, Matthias J. Sax a
>> écrit :
>>
>&
Server - Created server with
> *tickTime 2000* minSessionTimeout 4000 maxSessionTimeout 4
>
> so maybe the Client session is timed out because of the ZK « small »
> tickTime. I have a doubt that the changes i made are not really
> considered by storm.
>
> Thanks,
>
> Lina
>
The information you are looking for is provided by each incoming Tuple
of Bolt.execute(Tuple).
For example, tuple.getSourceComponent()
Have a look here for other available methods:
https://storm.apache.org/javadoc/apidocs/backtype/storm/tuple/Tuple.html
-Matthias
On 08/29/2015 02:48 PM, Marc Ro
You have two options to determine your producers:
1) each incoming tuples contains meta data you can access. For example:
Tuple.getSourceComponent(). See
https://storm.apache.org/javadoc/apidocs/backtype/storm/tuple/Tuple.html
2) in Bolt.prepare(...) one parameter is a TopologyContext object. T
Sounds reasonable. Storm does not provide any help with this. I assume
that your sensors attach a timestamp as regular attribute to each tuple.
Or do you timestamp your date in spout?
-Matthias
On 08/31/2015 09:11 AM, Marc Roos wrote:
>
> I have sensors that seem to be emitting duplicates and no
I would not do it this way... If you don't provide a custom scheduler,
you don't know if both executors will be deployed to the same worker JVM
(actually, the changes are almost zero that this happens...).
(Furthermore, you need to do proper synchronization between both
executors accessing the sam
Without any exception/error message it is hard to tell.
What is your cluster setup
- Hardware, ie, number of cores per node?
- How many node/supervisor are available?
- Configured number of workers for the topology?
- What is the number of task for each spout/bolt?
- What is the number o
4 to 41
>
> Thanks,
> Nick
>
> 2015-09-02 15:42 GMT-04:00 Matthias J. Sax <mailto:mj...@apache.org>>:
>
> Without any exception/error message it is hard to tell.
>
> What is your cluster setup
> - Hardware, ie, number of cores per node?
>
rker
> latencies are much lower than the inter-worker latencies?
>
> Thanks,
> Nick
>
> 2015-09-02 16:27 GMT-04:00 Matthias J. Sax <mailto:mj...@apache.org>>:
>
> So (for each node) you have 4 cores available for 1 supervisor JVM, 2
> worker JVMs that e
thing? Is it just a coincidence that happened in my experiments?
>
> Thank you,
> Nick
>
>
>
> 2015-09-02 17:38 GMT-04:00 Matthias J. Sax <mailto:mj...@apache.org>>:
>
> I agree. The load is not high.
>
> About higher latencies. How many ackers did you
There is no timeout or anything similar. It must be something different.
-Matthias
On 09/08/2015 09:18 AM, Chandrashekhar Kotekar wrote:
> Hi,
>
> I have 5 node storm cluster. On one of the node I am running Nimbus,
> Supervisor and UI as well. I have observed that Storm nimbus and UI
> shuts do
There is no such thing for Bolts.
The call to Spout.ack(...) happens after Storm retrieved all acks of all
(transitively) anchored tuples.
Let's say you have spout -> bolt1 -> bolt2
Spout emit t1 which is processed by bolt1.
bolt1 emits t2 (with anchor t1) and acks t1.
=> there will be no call t
Hi,
"emitted" is the number of output tuples the spout/bolt produces, while
"transfered" is the number if tuples that get shipped to consumers. If
you have two consumer bolts for a single spout/bolt, the "transfered"
count will be twice the "emitted" count, because each emitted tuple gets
transfer
Hi,
You can simple read the file directly in your Spout. This is an
implementation that reads multiple files concurrently (with respect to a
timestamp attribute that is included in the input record -- of course
you can simplify the code if you don't have a timestamp attribute and
just want to read
Hi Ankur,
If you declare a direct stream (setting the flag to true), you need to
emit tuples via
collector.directEmit(...)
methods (collector.emit(...) is not allowed for direct streams). Those
methods require to specify the consumer task ID that should receive the
tuple.
Furthermore, when co
em?
>
> Thanks again,
> Nick
>
> On Thu, Sep 10, 2015 at 5:15 PM, Matthias J. Sax <mailto:mj...@apache.org>> wrote:
>
> Hi,
>
> You can simple read the file directly in your Spout. This is an
> implementation that reads multiple files concu
If your clocks are not synchronized you will have a hard time... Why can
you not use NTP?
The only other idea I have, would be to use a custom scheduler that
ensure that all spout and sink instances are running on the same
machine, such that they can access the same local clock. But it is not
clea
Hi,
as always: it depends. ;)
Storm itself clear ups its own resources just fine. However, if the
running topology needs to clean-up/release resources before it is shut
down, Storm is not of any help. Even if there is a Spout/Bolt cleanup()
method, Storm does not guarantee that it will be called.
Let me break this down in a more simple fashion.
>
> I have a Storm Cluster named "The Quiet Storm" ;) here is what it
> consists of:
>
> **
> Server ZK1: Running Zookeeper
> Server ZK2: Running Zookeeper
> Server ZK3: Running Zookeeper
&
If you can use "partial key grouping" depends on your use case. Think
careful before you apply it...
Maybe you want to read the research paper about it. It clearly describes
when you can use it and when not:
https://melmeric.files.wordpress.com/2014/11/the-power-of-both-choices-practical-load-bala
> of my field groups to one bolt thereby causing it to be a
> bottleneck. Since I emit string, I guess the hash is on
> ArrayList(str1,str2...).hashcode(). This hashcode is coming out same
> for different string combinations...
>
> Thanks
> Kashyap
&g
>
> Thanks you!
>
> Kashyap
>
> On Sep 30, 2015 5:14 AM, "Matthias J. Sax" <mailto:mj...@apache.org>> wrote:
>
> Yes. That's right.
>
> "Values" extends ArrayList and does not overwrite .hashCode().
>
> -Matthias
>
Hi,
I never tried it, but as you add a hook in Bolt.prepare() you should be
able to create your own hook class that takes the OutputCollector as
constructor parameter and use it later on when the hook is called.
-Matthias
On 10/01/2015 12:23 AM, Raymond Conn wrote:
> Hi all,
> Just wanted to fo
This question was on SO, too, and got answered already:
https://stackoverflow.com/questions/33157303/apache-storm-classnotfoundexception-when-deploying-jar-to-remotecluster/33158308#33158308
On 10/15/2015 09:45 PM, Ankur Garg wrote:
> Hi ,
>
> I am trying to deploy my topology bundled as a fa
I would recommend to write a (bash) script.
Personally, I use the following assembly of scripts to start/stop a cluster:
The two main scripts are "start-storm.sh" and "stop-storm.sh". All other
are just helpers.
-Matthias
cat start-storm.sh
> #!/bin/bash
>
> tools=/home/tools
>
> # check if
Hi,
currently, there are two tags (apache-storm and storm) used on SO. I
just suggested "apache-storm" to be the main tag and "storm" to be a
synonym for it. This enables that all questions get tagged with a unique
tag. Old and new questions get re-tag from storm to apache-storm
automatically if t
You need to specify a cyclic dataflow:
> builder.setSpout("spout", ...);
> builder.setBolt("bolt1", ...).directGrouping("spout").directGrouping("bolt1");
> builder.setBolt("bolt2, ...).directGropuing("bolt1");
You can use the default stream.
-Matthias
On 11/04/2015 09:06 PM, Nathan Leung wrote:
Hi Raisul,
you would need to do something like this:
> BoltDeclarer agrBolt = builder.setBolt("mux-segment", new
> aggreaterProcess(framerate), 1)
> for (int i = 0; i < directoryInarr.length; i++) {
> agrBolt = argBolt.fieldsGrouping(boltName+Integer.toString(i),
> "stream_"+Integer.toStrin
In local mode, tuples are passed in memory from operator to operator
without going through the serialization stack.
I would recommend to write a couple of unit tests for your serializer.
-Matthias
On 11/09/2015 10:13 AM, Stephen Powis wrote:
> Sorry I can't answer your question, but if you've wr
There are input and output buffer for each operator. So it could sit on
the sender or on the receiver side.
Have a look here for more details:
http://www.michael-noll.com/blog/2013/06/21/understanding-storm-internal-message-buffers/
If this is not enough information, just ask again.
-Matthias
Hi,
a globally shared state is not supported by Storm. This is true, for
many (maybe even all?) large-scale distributed systems, because a shared
global stated does not scale!
However, if you just want to know the number of processed tuples, you do
not need a globally shared state. Actually, Stor
Hi Rudraneel,
find an good tutorial here:
https://xumingming.sinaapp.com/885/twitter-storm-how-to-develop-a-pluggable-scheduler/
Btw: the scheduler is called periodically.
-Matthias
On 12/09/2015 12:54 AM, Rudraneel chakraborty wrote:
> Hello ,
>
> I am trying to develop a custom storm schedu
It should be
> git clone https://github.com/apache/storm.git
-Matthias
On 12/19/2015 04:26 AM, Thirumeni, Sripriya wrote:
>
>
> Trying to get started on the Strom starter project getting the following
> error. Can you help.
>
>
>
> https://github.com/apache/storm/tree/master/examples/stor
Can you provide more information? Your code? StackTrace?
The information you provide is way too limited to give any help.
Just one thing: "supervisor still hasn't start" indicated that something
goes wrong when instantiating a Spout or Bolt. But it is unclear what
does go wrong.
Can you run your
Hi,
if you want to disable fault-tolerance, you need to emit tuples in your
spout without message IDs
-> collector.emit(new Values(...)); // no messageId provided
The parameter "topology.acker.executors" sets the number of ackers you
want to use in your topology. I am not sure, if setting it to
Are you trying to set up distributed mode on a single machine?
You might need to add a line to /etc/hosts
127.0.0.1 NameOfYourMachine
Using "localhost" as name for 127.0.0.1 confuses Storm...
-Matthias
On 12/25/2015 09:18 PM, sam mohel wrote:
> i have in /etc/hosts
> #192.168.x.xloca
Hi,
message IDs are only used if you enable fault-tolerance, ie, assign an
ID to each tuple in your spouts.
You do not assign an ID: "id: {}"
>> 6520 [Thread-14-UserContextBolt] INFO b.s.d.executor - Processing
>> received message FOR 3 TUPLE: source: *GenericEventSpout*:1, stream:
>> *GameEven
I guess you downloaded the binary and extracted it into a certain
folder. Just delete this folder to uninstall Storm.
-Matthias
On 01/02/2016 12:48 AM, sam mohel wrote:
> i know that question is simple for you but i'm new to storm followed the
> instruction to install it , but if i want to upgrad
to be processed, you can emit them as
> unanchored tuples. Since they're not anchored to any spout tuples, they
> won't cause any spout tuples to fail if they aren't acked.
>
> --John
>
>
>
>
> On Wed, Dec 23, 2015 at 5:01 PM, Matthias J. Sax <mailto:mj
I doubt it is a port problem.
0.0.0.0 is *no* valid IP address. Check your IP configuration.
-Matthias
On 01/04/2016 04:15 PM, Derek Dagit wrote:
>> org.jboss.netty.channel.ChannelException: Failed to bind to:
>> 0.0.0.0/0.0.0.0:6703
>
>
> If you see this, you can use a tool like lsof to find
If you know the number of words before hand, you can set the parallelism
accordingly. Furthermore, some custom grouping should allow you to sent
a single tuple to each instance. If you producer has only one instance,
a simple shuffle grouping would do the trick.
If you don't know the number of wor
It is absolutely ok what you are doing. No need to worry about anything.
If your tuples really time out, you can increase the timeout via
TOPOLOGY_MESSAGE_TIMEOUT_SECS (the default value is 30 seconds).
-Matthias
On 01/10/2016 06:50 PM, John Yost wrote:
> Hi Everyone,
>
> I have a topology tha
Hi,
if I remember correctly, Storm is pseudo-distibuted mode has problems to
deal with "localhost" as machine name. Thus, you need to extend
/etc/hosts with the real name of you local machine:
127.0.0.1 localhost
127.0.0.1 nameOfLocalMachine # add this line
-Matthias
On 01/
You need to start up logviewer first:
bin/storm logviewer
-Matthias
On 01/11/2016 09:46 AM, researcher cs wrote:
> when i clicked port number in storm ui i got the following message
>
> Unable to connect
> Firefox can't establish a connection to the server at nameofmachine:8000.
signature.a
1 - 100 of 164 matches
Mail list logo