Re: java.io.NotSerializableException: redis.clients.jedis.BinaryJedis$SetFromList

Hi Gergely, Thanks for using storm-redis, and report issue. The thing Storm was trying to serialize is BinaryJedis$SetFromList, not BinaryJedis. https://github.com/Fogetti/fogetti-phish-storm/blob/master/src/main/java/fogetti/phish/storm/relatedness/GoogleSemBolt.java#L74 Jedis.smembers() retur

Re: Thread[main,5,main] died

Could you share your Storm version? And could you check that all cluster nodes & topologies are using same version? Best, Jungtaek Lim (HeartSaVioR) 2015-12-09 23:57 GMT+09:00 askr askor : > can anybody help ? > > On Wed, Dec 9, 2015 at 5:04 AM, askr askor wrote: > >> i'm new to storm , i faced

Re: HBase bolt preparation failed

Could you share your worker's classpath? Looks like HBase libraries messed up in classpath. Best, Jungtaek Lim (HeartSaVioR) 2015-12-09 19:05 GMT+09:00 Venkatesh Bodapati : > i am working with storm-hbase, i got this error while running storm jar > > > java.lang.RuntimeException: HBase bolt prep

Re: Storm worker becoming zombie process

Hi Seungjin, I didn't see/hear this behavior, so we may want to have more evidences to find out issues, for example, thread dump via jstack. Best, Jungtaek Lim (HeartSaVioR) 2015-11-26 10:40 GMT+09:00 이승진 : > These days I'm seeing a weird problem of zombie worker process. > > > > When I restar

Re: error when submit the topology

Please set scope to "provided" from your storm-core dependency. Thanks, Jungtaek Lim (HeartSaVioR) 2015-11-03 8:29 GMT+09:00 researcher cs : > storm > storm 0.8.2 > compile junit > junit 3.8.1 test > colt > colt 1.2.0 > com.fasterxml.jackson.core > jackson-core 2.0.0 > org.twitter4j >

Re: Yet another Storm Cassandra connector

I'm not familiar with Cassandra, but I've seen some users in mailing group are using Storm + Cassandra. So I think this feature would be helpful to users. Before going into the details, great job, Florian! Best, Jungtaek Lim (HeartSaVioR) 2015-10-28 21:47 GMT+09:00 Florian Hussonnois : > Hi Par

Re: Storm at Stackoverflow

Nice work Matthias! Upvoted. Jungtaek Lim (HeartSaVioR) 2015-10-22 1:16 GMT+09:00 Matthias J. Sax : > 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 questi

Re: why storm is written in clojure

FYI, Storm has been using Disruptor from 0.8.0. http://storm.apache.org/2012/08/02/storm080-released.html And Nathan loves Clojure. https://twitter.com/nathanmarz/status/639137146620784641 I don't think Storm should be implemented as Clojure. Storm could be implemented as pure Java (see JStorm, w

Re: Ack not being called

Hi, Stuart. Apache Storm is a "Framework" to provide a distributed realtime computation, so you may need to understand the Concept of Apache Storm to successfully adopt your topologies to production. http://storm.apache.org/documentation/Concepts.html >From this page, section "Spouts", It is imp

Re: one worker per machine per topology, is it still recommended?

In addition to Matthias's opinion, it's true that inter-process communication cost is high, since it also skips serialization / deserialization when destination is same process. And having less workers reduce ZK load. But you need to configure more heap memory to each worker when using less worker

Re: Storm Trident - persistentAggregate With HBase 0.98

oetz <https://github.com/ptgoetz>" code > which is giving me the exception. > > Do you or anyone have used the persistanceAggregate with trident using new > version of HBase. > > Regards, > Nilesh Chhapru. > > > On Thursday 06 August 2015 03:35 AM, 임정택 wrote

Re: Storm Trident - persistentAggregate With HBase 0.98

Hi, I didn't look into it, but it seems to use old or unmanaged module. Official extern module for HBase is here, https://github.com/apache/storm/tree/master/external/storm-hbase. It has been released with Storm 0.9.3 and onwards. Please refer its README.md on Github repository, and below link to

Re: STORM REST API - Topology ID

Hi, You can get topology ID by sending */api/v1/topology/summary (GET)* API and find your target topology by name. https://github.com/apache/storm/blob/0.9.x-branch/STORM-UI-REST-API.md Hope this helps. Best, Jungtaek Lim (HeartSaVioR) 2015-07-28 22:57 GMT+09:00 Hong Jeon : > Hi, > > I'm tryin

Re: Process Latency < Execute latency for BaseBasicBolt?

Hi, If you use Storm 0.9.x then https://issues.apache.org/jira/browse/STORM-788 may be a reason to process latency < execute latency. It's merged to master and 0.10.0-beta, but not backported yet. Hope this helps. Thanks, Jungtaek Lim (HeartSaVioR) 2015-07-28 4:55 GMT+09:00 Kashyap Mhaisekar :

Re: Gathering opinion: changing multilang heartbeat mechanism

U intensive tasks, as people can just bump up the wait > time. > > -Dan > > On Jul 9, 2015, at 11:51 PM, 임정택 wrote: > > Thinking GIL once more, current approach can't deal with GIL, too. > If one of tuple takes more time then heartbeat timeout processing CPU > intensiv

Re: The supervisor cannot launch the workers

it is up to you. Thanks, Jungtaek Lim (HeartSaVioR) 2015-07-10 15:49 GMT+09:00 Chuanlei Ni : > The version of storm is 0.8.1 > And the jstack result is attached. > Thanks for responding. > > 2015-07-10 14:07 GMT+08:00 임정택 : > >> Hi, Chuanlei. >> >> Could you share

Re: The supervisor cannot launch the workers

Hi, Chuanlei. Could you share your Storm version? And is it reproducible? I'd like to see thread dump while supervisor and workers are freezing. I can't find out your issue deeply with only symptom. Thanks, Jungtaek Lim (HeartSaVioR) 2015-07-10 13:36 GMT+09:00 Chuanlei Ni : > Maybe the supervis

Re: Gathering opinion: changing multilang heartbeat mechanism

rs us whatever we're checking heartbeat from subprocess. Only thing we can avoid this situation is multiprocessing, which is too complex so I'm afraid we have to follow. Best, Jungtaek Lim (HeartSaVioR) 2015-07-10 11:19 GMT+09:00 임정택 : > Dan, > > I experimented about python&

Re: Gathering opinion: changing multilang heartbeat mechanism

nion. I don't have strong opinion to apply newer heartbeat mechanism since I found another constraint. I'd like to hear any opinions, objections, suggestions so please don't hesitate to tell. Thanks, Jungtaek Lim (HeartSaVioR) 2015-07-10 8:04 GMT+09:00 임정택 : > Thanks Dan for givin

Re: Gathering opinion: changing multilang heartbeat mechanism

GIL issues) just to accommodate a change to the heartbeat > protocol. It seems to me that the workaround you proposed in STORM–742 > <https://issues.apache.org/jira/browse/STORM-742> (where any command the > ShellBolt receives counts as a heartbeat) should be sufficient. > > Thank

Re: Problem to recept massive tuples

Hi, Charlie. Your Spout has the issue. Spout shouldn't stay longer in nextTuple() cause Spout takes care of events (including calling ack, fail, nextTuple) in event loop with just one thread. In other words, back pressure cannot work in your Spout cause how max spout pending works is checking pen

Re: Storm Async loop problem and Netty-Client connection failed

Hi. Seems like these are logs which occurs in progress of connecting. (We may want to lower its log level cause it doesn't mean we give up to connect, just attempting.) Your worker should print either "Giving up to connect to ", or "connection established to ". Former means that worker gave up c

Re: How to configure Storm log level

- Did your workers restarted after changing log level? - Could you check your workers to have -Dlogback.configurationFile option, and it's pointing to modified file? 2015-07-02 9:04 GMT+09:00 张炜 : > Thanks a lot, 임정택. > > It's my my bad: This is also set but

Re: How to configure Storm log level

Hi, You have to modify root logger tag's level to WARN to make your worker write less logs. Normally you don't need to adjust access / metrics logger's level cause it is relatively small. Hope this helps. Thanks, Jungtaek Lim (HeartSaVioR) 2015-07-02 8:50 GMT+09:00 张炜 : > Hi all, > Storm

Re: storm ui problem

Hi! What version do you use? Thanks, Jungtaek Lim (HeartSaVioR) 2015-07-01 12:31 GMT+09:00 徐鹏 : > hi: > I found a weird problem which shown in the picture. > Base on my understanding, the number of tuple emited and the number of > tuple transfered of the spout should be swaped. > Correct me if

Re: ShellBolts and Heartbeats

Hi, ShellSpout / ShellBolt will suicide itself when there's a heartbeat issue from the subprocess, and Supervisor or Nimbus will relaunch ShellSpout / ShellBolt eventually. Btw, ShellSpout treats any messages from subprocess as heartbeat, so piling up is not an issue. But, ShellBolt didn't take c

Re: Tried moving .0.9.4 to 0.9.5 but logs are still going to 0.9.4

Hi! Please check following things, - Could you check your worker process that where it is logging now by "-Dstorm.log.dir"? - Did you shutdown your whole topology and relaunch it? Thanks, Jungtaek Lim (HeartSaVioR) 2015-06-30 3:36 GMT+09:00 Dillian Murphey : > I'm trying to move to .0.9.5. >

Re: Thrift Error while submitting Topology - Message Length Exceeded

Could you provide stack traces about submitting and nimbus? Regards, Jungtaek Lim (HeartSaVioR) 2015-06-27 0:00 GMT+09:00 Jaime Solano : > Hi guys, > > I'm new in the list, so I'm not sure if this question has already been > answered. I'm getting the following error while trying to submit a topo

Re: Is there an upgrade process for storm?

Please note that it is just our team's approach. We install Storm by Ansible, which sets up configuration files, too. Based on Ansible, we're maintaining symbolic link which points last installed version of Storm. It means that we install clean version of Storm for each upgrade, and change symbolic

Gathering opinion: changing multilang heartbeat mechanism

Hi! Since it's about multilang feature and you can use your own implementation of multilang (and I believe multilang library developers are subscribing user group), I wanna get opinion about changing multilang heartbeat mechanism. At Storm 0.9.3, Storm introduces multilang heartbeat feature. http

Re: increasing timeout in Test (STORM_TEST_TIMEOUT_MS)

tests with adjusted TEST-TIMEOUT-MS. 2015-06-22 17:24 GMT+09:00 Tousif : > Jungtaek, > > Is there a way to override timeout without rebuilding ? > > On Sat, Jun 20, 2015 at 5:36 AM, 임정택 wrote: > >> export STORM_TEST_TIMEOUT_MS=3 >>> mvn clean install >> &g

Re: increasing timeout in Test (STORM_TEST_TIMEOUT_MS)

> > export STORM_TEST_TIMEOUT_MS=3 > mvn clean install should work, cause we already use it via Travis CI, and it's documented to DEVELOPER.md. Hope this helps. Thanks. Jungtaek Lim (HeartSaVioR) 2015-06-18 18:14 GMT+09:00 Tousif : > Hello, > > I tried passing STORM_TEST_TIMEOUT_MS as env

Re: Load balance biased shuffle grouping?

Hi! In shuffle grouping, you cannot assume that some executors continuously receive tuples which points to hotspot region since tuples are distributed randomly. Worst case is that hotspot region tuples are going to specific executors. Tuples are shuffled so sometimes it can be happened, but it wou

Re: Does Storm keep sending tick tuples to bolts when a topology is deactivated?

Hi! AFAIK, "setup-tick!" is called from start of executor (which schedules tick timer for each executor), and tick tuples will be emitted unless worker is going to be shutdown. In short, your fellow is correct. Thanks! Jungtaek Lim (HeartSaVioR) 2015-06-08 21:52 GMT+09:00 Dan Blanchard : > >

Re: ShellBolt Error - Anchored onto “” after ack/fail

And i saw a strange case in my cppBolt running, my spout sent the tuple > to my cppBolt and from a time, the last one did not receive the tuple. > > > -- > *De :* 임정택 > > *Envoyé :* mercredi 3 juin 2015 15:13 > *À :* user@storm.apache.org >

Re: ShellBolt Error - Anchored onto “” after ack/fail

e quillard : > Thanks for answering, i will add message timeout sec and will keep you > updated. > > > ------ > *De :* 임정택 > *Envoyé :* mercredi 3 juin 2015 12:33 > *À :* user@storm.apache.org > *Objet :* Re: ShellBolt Error - Anchored onto “” a

Re: ShellBolt Error - Anchored onto “” after ack/fail

Hi. IMO there seems to be race condition between BoltReaderRunnable.run() and ShellBolt.cleanup(). ShellBolt.cleanup() interrupts ReaderThread but actually few of methods inside BoltReaderRunnable.run() throws InterruptedException. (Accessing _pendingWrites throws InterruptedException, but I canno

Re: Applying a patch

applying patch > fails? > > > > The command I used was “apply patch 527.patch” > > > > *From:* 임정택 [mailto:kabh...@gmail.com] > *Sent:* Tuesday, June 02, 2015 5:41 PM > > *To:* user@storm.apache.org > *Subject:* Re: Applying a patch > > > > I think there w

Re: Trident Not Counting Properly

Hi! Trident is a Transactional Topology coordinated by micro-batch. In this case you set MAX_BATCH_SIZE_CONF to 100, batch size is set up by 100. So first 5 lines represent first batch (100 datas emitted from CSVSpout), and rest lines represent second batch. You may want to use State to persist a

Re: Applying a patch

I think there would be no problem applying this patch to 0.9.2 and upper, since mk-transfer-fn has changed to current via 0.9.2. But Storm 0.9.5 also resolves some more puzzling errors so I recommend you to apply 0.9.5. Thanks! Jungtaek Lim (HeartSaVioR) 2015-06-03 5:58 GMT+09:00 Srividhya Shan

Re: [DISCUSS] Drop Support for Java 1.6 in Storm 0.10.0

Great for dropping support 1.6. Thanks. Jungtaek Lim (HeartSaVioR) 2015년 6월 2일 화요일, P. Taylor Goetz님이 작성한 메시지: > 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 have been very > conservative in terms of

Re: Load /Cache Data inside Bolt

If it is really static thing, you can serialize it and pass to Bolt's constructor. If you don't want to spend serialization cost, you can store it to file (and include it to resources of jar) and load it from preparation of Bolt. If you want to have small jar, you can use external storage. (RDB, HD

Re: RotatingMap and Spout

; ExpiredCallBack implementation. So the message time out delta is not > evaluated when rotate method is invoked here. All those logic actually > happens in the executor’s mk-thread method for spout. > > Let me know if my understanding is different. > > > > *From:* 임정택 [mai

Re: 1 spout emit 2 tuples to 2 different bolt

Hi. I also suggest you to read https://storm.apache.org/documentation/Concepts.html before reading my comment. You can define multiple streams from Spout, and emit tuple A and B to different streams. (But you should emit tuples with stream nam

Re: RotatingMap and Spout

t can be run from different workers, and though it runs with same worker, synchronization for the map could introduce worse performance. > > Thanks again, > > Srividhya > > > > Thanks! Jungtaek Lim (HeartSaVioR) > > > > > *From:* 임정택 [mailto:kabh...@

Re: RotatingMap and Spout

Hi. AFAIK, Spout sets up timer to emit tick tuples each TOPOLOGY-TICK-TUPLE-FREQ-SECS. (Please refer setup-ticks!) Spout changes TOPOLOGY-TICK-TUPLE-FREQ-SECS to TOPOLOGY-MESSAGEE-TIMEOUT-SECS for the first time, so timer is scheduled to each TOPOLOGY-MESSAGE-TIMEOUT-SECS. (Please refer add-acker

Re: Multiple Workers in local mode gives error

ll see one single > process handling all the bolts and when running 6 workers for 5 bolts and 1 > spout than i will see 6 separate processes running each bolt and spout? M i > right.If not then what does it mean? > > > On Thu, May 14, 2015 at 2:46 PM, 임정택 > wrote: > >&g

Re: Multiple Workers in local mode gives error

Hi. Storm serializes tuples when tuple should be sent to other (remote) worker. In other words, Storm doesn't serialize tuples when destination is local task. That's why you didn't meet error when testing with 1 worker. MapEventBean seems to be not serializable, so you need to convert to other da

Re: Converting LinkedHashMap to a Tuple

confirm processing > _collector.ack(tuple); > } > > The out come of the println displayed in bold is: [{action=monitor, > loc=49, time=2015-04-29 00:02:19, orig=10.26.107.214, i/f_dir=inbound, > i/f_name=eth2}] > > So you were right about the serializable

Re: Who Calling nextTuple() in spout?

Hi! First of all, you want to compare Spark streaming and Storm Trident, not Storm Spout-Bolt topology. It's not same. Generally batching makes more throughputs while it loses latency. Answering to your question, you can refer https://github.com/apache/storm/blob/master/storm-core/src/clj/backtyp

Re: Online docs for newbies or build/environment for storm-starter/WordCountTopology broken?

Lim (HeartSaVioR) 2015-05-12 9:20 GMT+09:00 Richard Huber : > 임정택 - > > Thank you very much for the quick reply I thought so. The native or > clojure code runs fine, this is strictly a packaging for either local > directory or preparation of the environment issue for Python

Re: Converting LinkedHashMap to a Tuple

Hi. Seems like you're confusing emit and ack. Ack is for guaranteeing message processing, not for sending something new. http://storm.apache.org/documentation/Guaranteeing-message-processing.html So when you converted sth. and want to send to next bolt, use emit. LinkedHashMap is serializable so

Re: Help with implementing storm topology for streaming tweets

Hi. You can refer storm-starter's PrintSampleStream to make your topology work with Twitter Streaming API. Also you can use storm-hdfs to interacting with HDFS. Hope this helps. Thanks. Jungtaek Lim (HeartSaVioR) 2015-05-10 0:27 GMT+09:00 mani kandan : > Hi > > I'm new to Storm, and I would l

Re: Unable To Perform Topology Fast

Hi. Please keep in mind that Spout uses single thread to follow control flow. For example, it would be better to return fast within nextTuple() cause Spout cannot check acker's response while in nextTuple() so already acked tuples can be timed out. Reading 100 mb files and emit whole lines in one

Re: Online docs for newbies or build/environment for storm-starter/WordCountTopology broken?

Hi! Recently Storm packages multilang files into Maven artifacts, and Supervisor extracts fat jar's resources to resource directory. Here's supervisor's log. 4584 [Thread-3] INFO backtype.storm.daemon.supervisor - Extracting resources from jar at ./target/flux-examples-0.11.0-SNAPSHOT.jar to /v

Re: Storm worker died - Client is being closed, and does not take requests any more

AFAIK Nimbus doesn't reassign task to other worker with normal circumstance. Maybe worker didn't send heartbeat well, or Nimbus detected worker's dead. Actually two things are same since Nimbus detects worker's dead by checking heartbeat. Hope this helps. Thanks. Jungtaek Lim (HeartSaVioR) 2015-

Re: Storm worker died - Client is being closed, and does not take requests any more

If one or more workers were moved to other slots, you may hit this issue. https://issues.apache.org/jira/browse/STORM-737 Hope this helps. 2015-05-08 18:53 GMT+09:00 SuXingyu : > I am running my cluster using Storm 0.9.3. This exception was caused by > sending msg while Netty client is closin

Re: Conflict with json-simple

com.googlecode.json-simple version 1.1.1 > ( If they had not already died due to an un-hanlded exception ). The > supervisor would then restart the worker, which would pick up the new, > 1.1.1 jar. > > This will only work if the com.google.code.json-simple library, 1.1.1 is >

Re: Conflict with json-simple

Hi. Unfortunately, Storm adds all jars in /lib directory to classpath first. Recently Storm relocates some libraries to reduce dependency conflict (please see http://storm.apache.org/2014/11/25/storm093-released.html) but json-simple is not in. You can try replacing json-simple-1.1.jar to json-si

Re: Storm halts "Shutting down master"

e.storm.testing - Shutting down in process > zookeeper > 36909 [main] INFO backtype.storm.testing - Done shutting down in process > zookeeper > 36909 [main] INFO backtype.storm.testing - Deleting temporary path > /tmp/d5babc7f-2c83-4e2f-9489-05fe73157fed > 36910 [main] INFO backty

Re: Storm halts "Shutting down master"

Hi. Could you elavorate it? - What's your Storm version? - Did you run your topology with local mode or cluster mode? - Is there no Exception and Stack trace logs from Nimbus? - Could you share more log lines, or full Nimbus log file? Regards Jungtaek Lim (HeartSaVioR) 2015-05-03 2:23 GMT+09:0

Re: Storm UI Internal Server Error

Did you install Zookeeper to nimbus node? Seems like nimbus cannot connect Zookeeper. 2015-04-18 15:16 GMT+09:00 Chun Yuen Lim : > The nimbus log is as attached. > -- Name : 임 정택 Blog : http://www.heartsavior.net / http://dev.heartsavior.net Twitter : http://twitter.com/heartsavior LinkedIn :

Re: Workaround until https://issues.apache.org/jira/browse/STORM-573 is fixed in a release

'm not sure it makes sense to kill the entire JVM just for > an exception. > > Thanks! > Mark > > On Thu, Apr 16, 2015 at 9:00 AM, 임정택 wrote: > >> Hello. >> >> You can try set system environment "STORM_TEST_TIMEOUT_MS" and run your &g

Re: Workaround until https://issues.apache.org/jira/browse/STORM-573 is fixed in a release

Hello. You can try set system environment "STORM_TEST_TIMEOUT_MS" and run your test. It'll bind to TEST_TIMEOUT_MS and it's default timeout value of complete-topology. Hope this helps. Regards. Jungtaek Lim (HeartSaVioR) 2015-04-16 1:08 GMT+09:00 Mark Tomko : > Hi, > > I'm working on some test

Re: Timeout twice more than I set up.

Hello! AFAIK, Spout uses tick tuple to trigger check tuple timeout. Tick interval is 'topology.message.timeout.secs'. Spout handles it to override Storm conf 'topology.tick.tuple.freq.secs' to 'topology.message.timeout.secs' on the fly. Spout also uses RotatingMap, which seems to be like a LRU. I

Re: Cannot connect to socket in spout in cluster mode

Hello. Maybe it's up to serverSocket. Could you expose whole code for verifying? Thanks Regard. Jungtaek Lim (HeartSaVioR) 2015-04-07 12:27 GMT+09:00 Junfeng Chen : > I intend to establish Socket connection in spout. Here is my code : > > @Override > public void nextTuple() { > Socket soc

Re: Worker process shows higher than expected CPU usage on an idle system

Did you disable THP? If you didn't, it would be better to give it a try. (Sorry, but I don't find a relationship between these, but THP shows CPU issues from many places.) https://twitter.com/mattsta/status/529327782351097856 2015-03-16 8:03 GMT+09:00 Dennis Cammer : > I haven't heard anything ba

Re: Kafka spout seems act like a bolt

As I understood correctly, you're turning on tcpflow and send a command to Redis, right? Then maybe Redis protocol could be captured by tcpflow and published to Kafka. Please check your tcpflow setting. 2015-03-11 21:10 GMT+09:00 Mithun Kalan : > Hi > I've got something weird going on. > > [my se

Re: Storm 0.9.3 ShellBolt stops processing tuples after a while

What version of Pyleus do you use? AFAIK Pyleus doesn't have any release versions which support heartbeat introduced from 0.9.3. It's on develop branch. Could you check your worker / shellbolt subprocess log, and paste it if you don't mind? On 2015년 3월 11일 (수) at 오전 4:39 Victor Godoy Poluceno < vi

Wish for Storm users/committers to review storm-redis integration

Hello. dashengju and me are trying to add a feature to Storm, storm-redis integration. There is already a JIRA issue and Github PR regarding this. But it could become more valuable when we collect our requ

Re: How to set up STORM_TEST_TIMEOUT_MS programatically in 0.9.3?

It's for modifying default timeout between storm tests to avoid failure of testing & building storm-core. It's for slow machines. If you need to modify default timeout to use by your own test, completeTopology in clojure side, trackedWait for java & clojure would be have timeout parameter. If func

Re: STORM-513

Hello! It's due to differences between spout and bolt. In spout, nextTuple is automatically called from Storm, so depending on your configuration but nextTuple in subprocess would be called periodically. That's great for heartbeat. But bolt is based on event, bolt never calls subprocess when ther

Re: [ANNOUNCE] Apache Storm 0.9.3 Released

It may be better to notify that multilang protocol's changeset (with documentation). Cause some users are directly implemented multilang protocol - including php users at mailing list - and they will have an issue of upgrading Storm to 0.9.3 - idle bolt is being killed without reason, and heartbea

Re: We want to use Zookeeper like service for synchronization which can be deployed on single machine on CFS.

Disk shared between > two or more nodes. For minimal change in the present code base we are going > for this design. > > Regards > Tarkeshwar > > > On Wed, Nov 19, 2014 at 12:21 PM, 임정택 wrote: > >> Hello. >> >> Just a curious, some question regarding on t

Re: We want to use Zookeeper like service for synchronization which can be deployed on single machine on CFS.

Hello. Just a curious, some question regarding on this. 1. What's the reason your team decides to find alternatives other than Zookeeper? 2. What / how do your new project synchronizes? 3. What's CFS, and what if your single machine running synchronization service has failure? Your distributed co

"graceful" multilang failures

Spout doesn't need additional heartbeat tuple, since sync message from subprocess (PHP) side should be sended continuously if it's working correctly. ShellSpout uses any messages from subprocess to check heartbeat, so I recommend you to change log level to DEBUG and check log message related to mul

Re: Can message successful exit from any bolt?

Yes, you can skip emit to that messages. Be sure that it should be "ack"ed when you use tracking tuple mode. Hope this helps. 2014-11-09 11:39 GMT+09:00 swapnil joshi : > Hi Storm Friends, > I have some question about topology execution. I have sample topology > it contain 10 bolts and in no

Re: Set default log levels in a topology jar

Sorry, it's logback/cluster.xml. 2014년 11월 7일 금요일, 임정택님이 작성한 메시지: > Storm uses its logback configuration, located on /cluster. > AFAIK, you should modify it because slf4j picks it first so your log > configuration file cannot bind. > Hope this helps. > > 2014년 11월 7일 금요일, Bryan Baugher >님이 작성한 메시

Re: Set default log levels in a topology jar

Storm uses its logback configuration, located on /cluster. AFAIK, you should modify it because slf4j picks it first so your log configuration file cannot bind. Hope this helps. 2014년 11월 7일 금요일, Bryan Baugher님이 작성한 메시지: > Hi everyone, > > We recently switched to Storm 0.9.2 which moved to logback

Re: multilang + failure modes

It can be related to STORM-513. But it fixes only python, ruby, nodejs multilang. Hope this helps. 2014년 10월 24일 금요일, William Oberman님이 작성한 메시지: > I'm thinking I'm getting hit with: > https://issues.apache.org/jira/browse/STORM-442 > > On Thu, Oct 23, 2014 at 3:10 PM, William Oberman > wrote: >

Re: What is the purpose of timing out tuples?

AFAIK, Storm guarantees "tuple runs at least once" when spout can replay failed tuple. If tuple should run only once, Trident supports it. 2014년 10월 24일 금요일, Sam Mati님이 작성한 메시지: > Thanks for the response. > > My takeaway is that "timing out" is a last ditch effort, and that your > timeout time

Re: Read from another Topology

How about merging topologies into one? Though tuple timeout should be set to max processing time into all of topologies, there's only way to work without adding other components. Btw, ideally supporting pub-sub between topology seems great, but AFAIK there're many hurdles to realize. 1. Subscribin