Any Link which can tell extensive details of storm grouping ?
Hello friends, I need some documentation on storm grouping and how it is internally working. Can you please help me? Regards tarkeshwar
Storm not processing topology without logs
Hi, I have started to explore the Storm for distributed processing for our use case which we were earlier fulfilling by JMS based MQ system. Topology worked after some efforts. It has one spout (KafkaSpout from kafka-storm project) and 3 bolts. First bolt sets context for other two bolts which in turn do some processing on the tuples and persist the analyzed results in some DB (Mongo, Solr, HBase etc). Recently the topology stopped working. I am able to submit the topology and it does not throw any error in submitting the topology, however, nimbus.log or worker-6701.log files are not showing any progress and eventually topology does not consume any message. I don't have doubt on KafkaSpout because if it was the culprit, at least some initialization logs of spout and bolts should have been there in nimbus.log or worker-.log. Isn't it? Here is the snippet of nimbus.log after uploading the jar to cluster Uploading file from client to /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Finished uploading file from client: /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Received topology submission for aleads with conf {topology.max.task.parallelism nil, topology.acker.executors nil, topology.kryo.register nil, topology.kryo.decorators (), topology.name aleads, storm.id aleads-3-1408964869, modelId ut, topology.workers 1, topology.debug true} 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Activating aleads: aleads-3-1408964869 2014-08-25 07:07:50 b.s.s.EvenScheduler [INFO] Available slots: ([e56c2cc7-d35a-4355-9906-506618ff70c5 6701] [e56c2cc7-d35a-4355-9906-506618ff70c5 6700]) 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Setting new assignment for topology id aleads-3-1408964869: #backtype.storm.daemon.common.Assignment{:master-code-dir /hadoop/storm/nimbus/stormdist/aleads-3-1408964869, :node-host {e56c2cc7-d35a-4355-9906-506618ff70c5 hdp.ambari}, :executor-node+port {[2 2] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [3 3] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [4 4] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [5 5] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [6 6] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [7 7] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [8 8] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [9 9] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [1 1] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701]}, :executor-start-time-secs {[1 1] 1408964870, [9 9] 1408964870, [8 8] 1408964870, [7 7] 1408964870, [6 6] 1408964870, [5 5] 1408964870, [4 4] 1408964870, [3 3] 1408964870, [2 2] 1408964870}} Can anyone guess what I have done wrong and why Storm is not giving any error log anywhere. Storm version is 0.9.1.2.1.3.0-563 (Installed via HortonWorks) Kafka version is 2.10-0.8.1.1 Storm-Kafka version 0.9.2-incubating -- 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
Re: Storm not processing topology without logs
Found the fix. I was stuck in this problem from last 3-4 days and it was waiting just for joining the storm mailing list to be resolved. :) I found something in supervisor.log this time, it was dumping jar_UUID still hasn't started with the actual worker java command which is failing. However, it was not showing any error. So, I copied the command from the logs and directly run on console and it showed me the root cause. Somehow, localhost was get appended to hdp.ambari (which was my host name) and due to it was not able to find the server to run the command on. :( On Mon, Aug 25, 2014 at 5:25 PM, Vikas Agarwal vi...@infoobjects.com wrote: Hi, I have started to explore the Storm for distributed processing for our use case which we were earlier fulfilling by JMS based MQ system. Topology worked after some efforts. It has one spout (KafkaSpout from kafka-storm project) and 3 bolts. First bolt sets context for other two bolts which in turn do some processing on the tuples and persist the analyzed results in some DB (Mongo, Solr, HBase etc). Recently the topology stopped working. I am able to submit the topology and it does not throw any error in submitting the topology, however, nimbus.log or worker-6701.log files are not showing any progress and eventually topology does not consume any message. I don't have doubt on KafkaSpout because if it was the culprit, at least some initialization logs of spout and bolts should have been there in nimbus.log or worker-.log. Isn't it? Here is the snippet of nimbus.log after uploading the jar to cluster Uploading file from client to /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Finished uploading file from client: /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Received topology submission for aleads with conf {topology.max.task.parallelism nil, topology.acker.executors nil, topology.kryo.register nil, topology.kryo.decorators (), topology.name aleads, storm.id aleads-3-1408964869, modelId ut, topology.workers 1, topology.debug true} 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Activating aleads: aleads-3-1408964869 2014-08-25 07:07:50 b.s.s.EvenScheduler [INFO] Available slots: ([e56c2cc7-d35a-4355-9906-506618ff70c5 6701] [e56c2cc7-d35a-4355-9906-506618ff70c5 6700]) 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Setting new assignment for topology id aleads-3-1408964869: #backtype.storm.daemon.common.Assignment{:master-code-dir /hadoop/storm/nimbus/stormdist/aleads-3-1408964869, :node-host {e56c2cc7-d35a-4355-9906-506618ff70c5 hdp.ambari}, :executor-node+port {[2 2] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [3 3] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [4 4] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [5 5] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [6 6] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [7 7] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [8 8] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [9 9] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [1 1] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701]}, :executor-start-time-secs {[1 1] 1408964870, [9 9] 1408964870, [8 8] 1408964870, [7 7] 1408964870, [6 6] 1408964870, [5 5] 1408964870, [4 4] 1408964870, [3 3] 1408964870, [2 2] 1408964870}} Can anyone guess what I have done wrong and why Storm is not giving any error log anywhere. Storm version is 0.9.1.2.1.3.0-563 (Installed via HortonWorks) Kafka version is 2.10-0.8.1.1 Storm-Kafka version 0.9.2-incubating -- 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
Re: Location of last error details seen in storm UI
Click on the link of the bolt/spout that is all the way on the left side. On Sun, Aug 24, 2014 at 11:19 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason
Re: Location of last error details seen in storm UI
Better would be to view log files under /var/log/storm. Any issue with worker would be logged into /var/log/storm/worker-6700.log and /var/log/storm/worker-6701.log. On Mon, Aug 25, 2014 at 8:00 PM, Vincent Russell vincent.russ...@gmail.com wrote: Click on the link of the bolt/spout that is all the way on the left side. On Sun, Aug 24, 2014 at 11:19 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason -- 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
Re: Location of last error details seen in storm UI
Thanks for the help. Unfortunately, the content shown when the spout/bolt is clicked does not show any information about any errors. From: Vincent Russell vincent.russ...@gmail.com To: user@storm.incubator.apache.org; Jason Kania jason.ka...@ymail.com Sent: Monday, August 25, 2014 10:30:32 AM Subject: Re: Location of last error details seen in storm UI Click on the link of the bolt/spout that is all the way on the left side. On Sun, Aug 24, 2014 at 11:19 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason
Re: Location of last error details seen in storm UI
Thanks for the response. Unfortunately, I have no /var/log/storm on my system. Where is the path to these logs specified. I am guessing it is pointing somewhere else by default. Thanks, Jason From: Vikas Agarwal vi...@infoobjects.com To: user@storm.incubator.apache.org Cc: Jason Kania jason.ka...@ymail.com Sent: Monday, August 25, 2014 10:34:00 AM Subject: Re: Location of last error details seen in storm UI Better would be to view log files under /var/log/storm. Any issue with worker would be logged into /var/log/storm/worker-6700.log and /var/log/storm/worker-6701.log. On Mon, Aug 25, 2014 at 8:00 PM, Vincent Russell vincent.russ...@gmail.com wrote: Click on the link of the bolt/spout that is all the way on the left side. On Sun, Aug 24, 2014 at 11:19 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason -- 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
Re: Location of last error details seen in storm UI
Jason, Default is under your storm installation check for logs dir. -Harsha On Mon, Aug 25, 2014, at 07:54 AM, Jason Kania wrote: Thanks for the response. Unfortunately, I have no /var/log/storm on my system. Where is the path to these logs specified. I am guessing it is pointing somewhere else by default. Thanks, Jason __ From: Vikas Agarwal vi...@infoobjects.com To: user@storm.incubator.apache.org Cc: Jason Kania jason.ka...@ymail.com Sent: Monday, August 25, 2014 10:34:00 AM Subject: Re: Location of last error details seen in storm UI Better would be to view log files under /var/log/storm. Any issue with worker would be logged into /var/log/storm/worker-6700.log and /var/log/storm/worker-6701.log. On Mon, Aug 25, 2014 at 8:00 PM, Vincent Russell [1]vincent.russ...@gmail.com wrote: Click on the link of the bolt/spout that is all the way on the left side. On Sun, Aug 24, 2014 at 11:19 PM, Jason Kania [2]jason.ka...@ymail.com wrote: Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason -- Regards, Vikas Agarwal 91 – 9928301411 InfoObjects, Inc. Execution Matters [3]http://www.infoobjects.com 2041 Mission College Boulevard, #280 Santa Clara, CA 95054 +1 (408) 988-2000 Work +1 (408) 716-2726 Fax References 1. mailto:vincent.russ...@gmail.com 2. mailto:jason.ka...@ymail.com 3. http://www.infoobjects.com/
RE: Storm not processing topology without logs
Are you able to see the topology in storm UI or with storm list command ?? And worker mentioned in the UI doesn't have any log ?? -Original Message- From: Vikas Agarwal Sent: 25-08-2014 PM 05:25 To: user@storm.incubator.apache.org Subject: Storm not processing topology without logs Hi, I have started to explore the Storm for distributed processing for our use case which we were earlier fulfilling by JMS based MQ system. Topology worked after some efforts. It has one spout (KafkaSpout from kafka-storm project) and 3 bolts. First bolt sets context for other two bolts which in turn do some processing on the tuples and persist the analyzed results in some DB (Mongo, Solr, HBase etc). Recently the topology stopped working. I am able to submit the topology and it does not throw any error in submitting the topology, however, nimbus.log or worker-6701.log files are not showing any progress and eventually topology does not consume any message. I don't have doubt on KafkaSpout because if it was the culprit, at least some initialization logs of spout and bolts should have been there in nimbus.log or worker-.log. Isn't it? Here is the snippet of nimbus.log after uploading the jar to cluster Uploading file from client to /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Finished uploading file from client: /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Received topology submission for aleads with conf {topology.max.task.parallelism nil, topology.acker.executors nil, topology.kryo.register nil, topology.kryo.decorators (), topology.name aleads, storm.id aleads-3-1408964869, modelId ut, topology.workers 1, topology.debug true} 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Activating aleads: aleads-3-1408964869 2014-08-25 07:07:50 b.s.s.EvenScheduler [INFO] Available slots: ([e56c2cc7-d35a-4355-9906-506618ff70c5 6701] [e56c2cc7-d35a-4355-9906-506618ff70c5 6700]) 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Setting new assignment for topology id aleads-3-1408964869: #backtype.storm.daemon.common.Assignment{:master-code-dir /hadoop/storm/nimbus/stormdist/aleads-3-1408964869, :node-host {e56c2cc7-d35a-4355-9906-506618ff70c5 hdp.ambari}, :executor-node+port {[2 2] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [3 3] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [4 4] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [5 5] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [6 6] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [7 7] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [8 8] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [9 9] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [1 1] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701]}, :executor-start-time-secs {[1 1] 1408964870, [9 9] 1408964870, [8 8] 1408964870, [7 7] 1408964870, [6 6] 1408964870, [5 5] 1408964870, [4 4] 1408964870, [3 3] 1408964870, [2 2] 1408964870}} Can anyone guess what I have done wrong and why Storm is not giving any error log anywhere. Storm version is 0.9.1.2.1.3.0-563 (Installed via HortonWorks) Kafka version is 2.10-0.8.1.1 Storm-Kafka version 0.9.2-incubating -- 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
RE: Location of last error details seen in storm UI
The log would be in the corresponding worker node for that bolt or spout . Ssh onto the node , inside storm installation directory there are logs like worker-6700.log and so on . You can get all the logs from there. -Original Message- From: Jason Kania Sent: 25-08-2014 AM 08:52 To: user@storm.incubator.apache.org Subject: Location of last error details seen in storm UI Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason
Re: Storm not processing topology without logs
Yes, I was able to see the topology in Storm UI and nothing was logged into worker logs. However, as I mentioned, I am able to resolve it by finding an hint in supervisor.log file this time. On Mon, Aug 25, 2014 at 8:58 PM, Georgy Abraham itsmegeo...@gmail.com wrote: Are you able to see the topology in storm UI or with storm list command ?? And worker mentioned in the UI doesn't have any log ?? -- From: Vikas Agarwal Sent: 25-08-2014 PM 05:25 To: user@storm.incubator.apache.org Subject: Storm not processing topology without logs Hi, I have started to explore the Storm for distributed processing for our use case which we were earlier fulfilling by JMS based MQ system. Topology worked after some efforts. It has one spout (KafkaSpout from kafka-storm project) and 3 bolts. First bolt sets context for other two bolts which in turn do some processing on the tuples and persist the analyzed results in some DB (Mongo, Solr, HBase etc). Recently the topology stopped working. I am able to submit the topology and it does not throw any error in submitting the topology, however, nimbus.log or worker-6701.log files are not showing any progress and eventually topology does not consume any message. I don't have doubt on KafkaSpout because if it was the culprit, at least some initialization logs of spout and bolts should have been there in nimbus.log or worker-.log. Isn't it? Here is the snippet of nimbus.log after uploading the jar to cluster Uploading file from client to /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Finished uploading file from client: /hadoop/storm/nimbus/inbox/stormjar-31fe068b-337b-428f-8ae2-fe13c706b2ab.jar 2014-08-25 07:07:49 b.s.d.nimbus [INFO] Received topology submission for aleads with conf {topology.max.task.parallelism nil, topology.acker.executors nil, topology.kryo.register nil, topology.kryo.decorators (), topology.name aleads, storm.id aleads-3-1408964869, modelId ut, topology.workers 1, topology.debug true} 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Activating aleads: aleads-3-1408964869 2014-08-25 07:07:50 b.s.s.EvenScheduler [INFO] Available slots: ([e56c2cc7-d35a-4355-9906-506618ff70c5 6701] [e56c2cc7-d35a-4355-9906-506618ff70c5 6700]) 2014-08-25 07:07:50 b.s.d.nimbus [INFO] Setting new assignment for topology id aleads-3-1408964869: #backtype.storm.daemon.common.Assignment{:master-code-dir /hadoop/storm/nimbus/stormdist/aleads-3-1408964869, :node-host {e56c2cc7-d35a-4355-9906-506618ff70c5 hdp.ambari}, :executor-node+port {[2 2] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [3 3] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [4 4] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [5 5] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [6 6] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [7 7] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [8 8] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [9 9] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701], [1 1] [e56c2cc7-d35a-4355-9906-506618ff70c5 6701]}, :executor-start-time-secs {[1 1] 1408964870, [9 9] 1408964870, [8 8] 1408964870, [7 7] 1408964870, [6 6] 1408964870, [5 5] 1408964870, [4 4] 1408964870, [3 3] 1408964870, [2 2] 1408964870}} Can anyone guess what I have done wrong and why Storm is not giving any error log anywhere. Storm version is 0.9.1.2.1.3.0-563 (Installed via HortonWorks) Kafka version is 2.10-0.8.1.1 Storm-Kafka version 0.9.2-incubating -- 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
Re: Location of last error details seen in storm UI
current version of storm doesn't have a way to define storm log dir. One way to do is to edit logback/cluster.xml under storm installation. Upcoming release will have a config option storm.log.dir to redirect the logs from default dir. -Harsha On Mon, Aug 25, 2014, at 08:16 AM, Jason Kania wrote: Thanks for that. I looked to find which property or configuration parameter sets it but could not find it. Is there such a parameter? Thanks, Jason __ From: Harsha st...@harsha.io To: user@storm.incubator.apache.org Sent: Monday, August 25, 2014 11:10:17 AM Subject: Re: Location of last error details seen in storm UI Jason, Default is under your storm installation check for logs dir. -Harsha On Mon, Aug 25, 2014, at 07:54 AM, Jason Kania wrote: Thanks for the response. Unfortunately, I have no /var/log/storm on my system. Where is the path to these logs specified. I am guessing it is pointing somewhere else by default. Thanks, Jason __ From: Vikas Agarwal vi...@infoobjects.com To: user@storm.incubator.apache.org Cc: Jason Kania jason.ka...@ymail.com Sent: Monday, August 25, 2014 10:34:00 AM Subject: Re: Location of last error details seen in storm UI Better would be to view log files under /var/log/storm. Any issue with worker would be logged into /var/log/storm/worker-6700.log and /var/log/storm/worker-6701.log. On Mon, Aug 25, 2014 at 8:00 PM, Vincent Russell [1]vincent.russ...@gmail.com wrote: Click on the link of the bolt/spout that is all the way on the left side. On Sun, Aug 24, 2014 at 11:19 PM, Jason Kania [2]jason.ka...@ymail.com wrote: Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason -- Regards, Vikas Agarwal 91 – 9928301411 InfoObjects, Inc. Execution Matters [3]http://www.infoobjects.com 2041 Mission College Boulevard, #280 Santa Clara, CA 95054 +1 (408) 988-2000 Work +1 (408) 716-2726 Fax References 1. mailto:vincent.russ...@gmail.com 2. mailto:jason.ka...@ymail.com 3. http://www.infoobjects.com/
Re: Location of last error details seen in storm UI
I have used HortonWorks distribution for Storm and other Hadoop ecosystem and it has property for setting storm.log.dir. It is using storm 0.9.1. On Mon, Aug 25, 2014 at 9:02 PM, Harsha st...@harsha.io wrote: current version of storm doesn't have a way to define storm log dir. One way to do is to edit logback/cluster.xml under storm installation. Upcoming release will have a config option storm.log.dir to redirect the logs from default dir. -Harsha On Mon, Aug 25, 2014, at 08:16 AM, Jason Kania wrote: Thanks for that. I looked to find which property or configuration parameter sets it but could not find it. Is there such a parameter? Thanks, Jason -- *From:* Harsha st...@harsha.io *To:* user@storm.incubator.apache.org *Sent:* Monday, August 25, 2014 11:10:17 AM *Subject:* Re: Location of last error details seen in storm UI Jason, Default is under your storm installation check for logs dir. -Harsha On Mon, Aug 25, 2014, at 07:54 AM, Jason Kania wrote: Thanks for the response. Unfortunately, I have no /var/log/storm on my system. Where is the path to these logs specified. I am guessing it is pointing somewhere else by default. Thanks, Jason -- *From:* Vikas Agarwal vi...@infoobjects.com *To:* user@storm.incubator.apache.org *Cc:* Jason Kania jason.ka...@ymail.com *Sent:* Monday, August 25, 2014 10:34:00 AM *Subject:* Re: Location of last error details seen in storm UI Better would be to view log files under /var/log/storm. Any issue with worker would be logged into /var/log/storm/worker-6700.log and /var/log/storm/worker-6701.log. On Mon, Aug 25, 2014 at 8:00 PM, Vincent Russell vincent.russ...@gmail.com wrote: Click on the link of the bolt/spout that is all the way on the left side. On Sun, Aug 24, 2014 at 11:19 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I am trying to get more detail on an error that is being displayed in the Storm UI under the Last Error column but unfortunately, I am not seeing it captured anywhere else. Does anyone know where this text could be seen? The problem is that the error text is insufficient to diagnose the problem. Thanks, Jason -- 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
Preventing storm from caching classes
Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason
Re: Preventing storm from caching classes
Are you killing the topology before uploading new one? On Mon, Aug 25, 2014 at 9:26 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason -- 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
Re: Preventing storm from caching classes
That sounds like a very odd issue, as each storm worker runs within an independent JVM. Are you putting all your classes in a single jar file (that you upload to storm)? Is there any chance your build system is not including the newest versions? Can you open the jar file and verify that the correct versions are there? --Tom On Mon, Aug 25, 2014 at 10:02 AM, Vikas Agarwal vi...@infoobjects.com wrote: Are you killing the topology before uploading new one? On Mon, Aug 25, 2014 at 9:26 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason -- 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
Re: Preventing storm from caching classes
Tom, All the classes are in a single jar and I know the build system is including the newest versions only. I can see this because I have had to keep renaming the classes each time I upload and the only things I see in the jar are my latest classes. Unfortunately, it is old classes that are not in the jar that storm would attempt to run without the renaming. I am running with the 1.7 update 67 jdk. Thanks, Jason From: Tom Brown tombrow...@gmail.com To: user@storm.incubator.apache.org user@storm.incubator.apache.org Sent: Monday, August 25, 2014 12:27:03 PM Subject: Re: Preventing storm from caching classes That sounds like a very odd issue, as each storm worker runs within an independent JVM. Are you putting all your classes in a single jar file (that you upload to storm)? Is there any chance your build system is not including the newest versions? Can you open the jar file and verify that the correct versions are there? --Tom On Mon, Aug 25, 2014 at 10:02 AM, Vikas Agarwal vi...@infoobjects.com wrote: Are you killing the topology before uploading new one? On Mon, Aug 25, 2014 at 9:26 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason -- 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
Re: Preventing storm from caching classes
I am killing the topology. I can even kill storm and the result remains the same once storm is relaunched. Jason From: Vikas Agarwal vi...@infoobjects.com To: user@storm.incubator.apache.org; Jason Kania jason.ka...@ymail.com Sent: Monday, August 25, 2014 12:02:25 PM Subject: Re: Preventing storm from caching classes Are you killing the topology before uploading new one? On Mon, Aug 25, 2014 at 9:26 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason -- 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
Determining Topology Ready State for Continuous Deployment/Integration
Hey, I'm curious if there's a way to determine when a newly-deployed Topology is ready. I'm trying to tighten up our continuous deployment pipeline during integration tests. I've got an simple end-to-end test to ensure a certain message is delivered through the cluster, and it would be nice to know when a topology is 'ready' so I can start queuing up messages. By ready I mean something along the lines of: - topology code has been deployed throughout the cluster - all jvm processes have started - (potentially): storm's internal acker has received one ack from each component. I've looked at the Nimbus API and it's not clear how much info I can glean from this. Any help is appreciated. Thanks! Best, Aaron Levin -- Aaron Levin Data Scientist demeure.com
Question on failing ack
I have set up topology to load a very large volume of data. Recently I just loaded about 60K records and found out that there are some failed acks on few spouts but non on the bolts. Storm completed running and seem to look stable. Initially i started with a lesser amount of data like about 500 records successfully and then increased up to 60K where i saw the failed acks. Questions: 1. Does that mean that the spout was not able to read some messages from Kafka? Since there are no failed ack on the bolts as per UI, what ever the message received has been successfully processed by the bolts. 2. how do i interpret the numbers of failed acks like this acked:315500 and failed: 2980. Does this mean that 2980 records failed to be processed? Is this is the case then, how do I avoid this from happening because I will be loosing 2980 records. 3. I also see that few of the records failed to be inserted into Cassandra database. What is the best way to reprocess the data again as it is quite difficult to do it through the batch process that I am currently running. LMK, thanks. -- Kushan Maskey 817.403.7500
Re: Question on failing ack
I would suspect that at some point the rate at which the spouts emitted exceeded the rate at which the bolts could process. Maybe you could look at configuring the buffers (if you haven't yet done that). Do your records get processed at a constant rate? On Tue, Aug 26, 2014 at 4:12 AM, Kushan Maskey kushan.mas...@mmillerassociates.com wrote: I have set up topology to load a very large volume of data. Recently I just loaded about 60K records and found out that there are some failed acks on few spouts but non on the bolts. Storm completed running and seem to look stable. Initially i started with a lesser amount of data like about 500 records successfully and then increased up to 60K where i saw the failed acks. Questions: 1. Does that mean that the spout was not able to read some messages from Kafka? Since there are no failed ack on the bolts as per UI, what ever the message received has been successfully processed by the bolts. 2. how do i interpret the numbers of failed acks like this acked:315500 and failed: 2980. Does this mean that 2980 records failed to be processed? Is this is the case then, how do I avoid this from happening because I will be loosing 2980 records. 3. I also see that few of the records failed to be inserted into Cassandra database. What is the best way to reprocess the data again as it is quite difficult to do it through the batch process that I am currently running. LMK, thanks. -- Kushan Maskey 817.403.7500
Re: Question on failing ack
Hi Kushan, Depending on the Kafka spout you're using, it could be doing different things when it failed. However, if it's running reliably, the Cassandra insertion failures would have forced a replay from the spout until they had completed. Michael Rose (@Xorlev https://twitter.com/xorlev) Senior Platform Engineer, FullContact http://www.fullcontact.com/ mich...@fullcontact.com On Mon, Aug 25, 2014 at 4:42 PM, Kushan Maskey kushan.mas...@mmillerassociates.com wrote: I have set up topology to load a very large volume of data. Recently I just loaded about 60K records and found out that there are some failed acks on few spouts but non on the bolts. Storm completed running and seem to look stable. Initially i started with a lesser amount of data like about 500 records successfully and then increased up to 60K where i saw the failed acks. Questions: 1. Does that mean that the spout was not able to read some messages from Kafka? Since there are no failed ack on the bolts as per UI, what ever the message received has been successfully processed by the bolts. 2. how do i interpret the numbers of failed acks like this acked:315500 and failed: 2980. Does this mean that 2980 records failed to be processed? Is this is the case then, how do I avoid this from happening because I will be loosing 2980 records. 3. I also see that few of the records failed to be inserted into Cassandra database. What is the best way to reprocess the data again as it is quite difficult to do it through the batch process that I am currently running. LMK, thanks. -- Kushan Maskey 817.403.7500
Re: Preventing storm from caching classes
You can check if kill operation is dumping any error in nimbus.log. It might be the case that it is unable to delete the jar somehow. Further, are you always seeing the immediate previous version of your classes on deployment of new one OR it is quite old version of the same? In later case, it might be due to some previous jar being present in storm lib or somewhere else from where it is coming into the classpath. On Mon, Aug 25, 2014 at 11:45 PM, Jason Kania jason.ka...@ymail.com wrote: I am killing the topology. I can even kill storm and the result remains the same once storm is relaunched. Jason -- *From:* Vikas Agarwal vi...@infoobjects.com *To:* user@storm.incubator.apache.org; Jason Kania jason.ka...@ymail.com *Sent:* Monday, August 25, 2014 12:02:25 PM *Subject:* Re: Preventing storm from caching classes Are you killing the topology before uploading new one? On Mon, Aug 25, 2014 at 9:26 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason -- 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
Re: Preventing storm from caching classes
Thanks for the response. I checked and didn't see any errors in the nimbus log. It is hard to tell exactly, but it seems that the classes being used vary over time and are one or more versions behind the current ones. To me it makes sense that it is something sneaking into the classpath, but as of yet, I do not know how. From: Vikas Agarwal vi...@infoobjects.com To: user@storm.incubator.apache.org; Jason Kania jason.ka...@ymail.com Sent: Monday, August 25, 2014 10:57:19 PM Subject: Re: Preventing storm from caching classes You can check if kill operation is dumping any error in nimbus.log. It might be the case that it is unable to delete the jar somehow. Further, are you always seeing the immediate previous version of your classes on deployment of new one OR it is quite old version of the same? In later case, it might be due to some previous jar being present in storm lib or somewhere else from where it is coming into the classpath. On Mon, Aug 25, 2014 at 11:45 PM, Jason Kania jason.ka...@ymail.com wrote: I am killing the topology. I can even kill storm and the result remains the same once storm is relaunched. Jason From: Vikas Agarwal vi...@infoobjects.com To: user@storm.incubator.apache.org; Jason Kania jason.ka...@ymail.com Sent: Monday, August 25, 2014 12:02:25 PM Subject: Re: Preventing storm from caching classes Are you killing the topology before uploading new one? On Mon, Aug 25, 2014 at 9:26 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason -- 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
Re: Preventing storm from caching classes
If you are running topology with debug true, I guess there would be the command used to launch the workers in supervisor.log. Check it, it may contain the clue to the wrong classpath. You can try running the that command directly from console to further debug the issue. On Tue, Aug 26, 2014 at 9:00 AM, Jason Kania jason.ka...@ymail.com wrote: Thanks for the response. I checked and didn't see any errors in the nimbus log. It is hard to tell exactly, but it seems that the classes being used vary over time and are one or more versions behind the current ones. To me it makes sense that it is something sneaking into the classpath, but as of yet, I do not know how. -- *From:* Vikas Agarwal vi...@infoobjects.com *To:* user@storm.incubator.apache.org; Jason Kania jason.ka...@ymail.com *Sent:* Monday, August 25, 2014 10:57:19 PM *Subject:* Re: Preventing storm from caching classes You can check if kill operation is dumping any error in nimbus.log. It might be the case that it is unable to delete the jar somehow. Further, are you always seeing the immediate previous version of your classes on deployment of new one OR it is quite old version of the same? In later case, it might be due to some previous jar being present in storm lib or somewhere else from where it is coming into the classpath. On Mon, Aug 25, 2014 at 11:45 PM, Jason Kania jason.ka...@ymail.com wrote: I am killing the topology. I can even kill storm and the result remains the same once storm is relaunched. Jason -- *From:* Vikas Agarwal vi...@infoobjects.com *To:* user@storm.incubator.apache.org; Jason Kania jason.ka...@ymail.com *Sent:* Monday, August 25, 2014 12:02:25 PM *Subject:* Re: Preventing storm from caching classes Are you killing the topology before uploading new one? On Mon, Aug 25, 2014 at 9:26 PM, Jason Kania jason.ka...@ymail.com wrote: Hello, I would like to know if anybody knows how to prevent Storm from caching a previous version of the classes making up a topology. I am hitting problems where Storm is running old versions of my classes instead of the ones in the currently supplied jar file. I am having to include version numbers in the names of my classes to get the desired code to run. Thanks, Jason -- 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 -- 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