Having a worker and nimbus on the same host should not be a problem. We have seen this exact thing happen when two nimbus daemons are operating on the same file system location (sometimes an NFS mount).
The other thing that comes to mind is ZooKeeper configuration. Make sure all of the ZooKeeper hosts are part of the same ZooKeeper cluster. -- Derek ________________________________ From: S. Zhou <[email protected]> To: Derek Dagit <[email protected]> Cc: "[email protected]" <[email protected]> Sent: Wednesday, March 25, 2015 12:55 PM Subject: Re: storm failed to activate topology Only one nimbus is running. Another thing: I run nimbus and worker on the same machine using "distributed" mode. I am not sure if it could be an issue. On Wednesday, March 25, 2015 10:13 AM, Derek Dagit <[email protected]> wrote: Check for a second nimbus process running somehow. There should be only one at a time. -- Derek ________________________________ From: S. Zhou <[email protected]> To: "[email protected]" <[email protected]>; yuanjun Li <[email protected]>; S. Zhou <[email protected]> Sent: Wednesday, March 25, 2015 12:03 PM Subject: Re: storm failed to activate topology Another thing I observed while investigating is: when I run "storm jar" command, it creates a directory "Mytopology-2-1427302545" under "nimbus/stormdist". In that directory it has 3 files "stormcode.ser stormconf.ser stormjar.jar" But after a while, these three files and the directory "Mytopology-2-1427302545" are disappeared. I tried to manually create that directory and copied back these three files, they are deleted (I guess by storm) again. On Wednesday, March 25, 2015 7:52 AM, S. Zhou <[email protected]> wrote: output from "storm list" 0 [main] INFO backtype.storm.thrift - Connecting to Nimbus at localhost:6627 No topologies running. On Tuesday, March 24, 2015 9:53 PM, yuanjun Li <[email protected]> wrote: use ./storm list to list your top status. 2015-03-25 9:52 GMT+08:00 S. Zhou <[email protected]>: I used "storm jar" first actually. Then the topology was active for a short period time then it dies. I checked log files from nimbus, ui, supervisor and worker and found no error message. > > > > > >On Tuesday, March 24, 2015 5:43 PM, Nathan Leung <[email protected]> wrote: > > > >Is the topology already running? Given the exception I suspect not. You will >need to load it to the cluster using "storm jar ...". Activate should only be >used if the topology is already loaded to the cluster and deactivated (i.e. >spouts are not calling nextTuple). > > >On Tue, Mar 24, 2015 at 6:07 PM, S. Zhou <[email protected]> wrote: > >I was running this command "./storm activate MyTopology" and got the following >exception. I looked through storm log files and could not see any clues. Any >hints? Thanks! >> >> >>We are using storm 0.8.0 on Ubuntu Lucid (10.04). >> >> >>0 [main] INFO backtype.storm.thrift - Connecting to Nimbus at >>localhost:6627 >>Exception in thread "main" NotAliveException(msg:MyTopology) >> at backtype.storm.generated.Nimbus$activate_result.read(Nimbus.java:4188) >> at org.apache.thrift7.TServiceClient.receiveBase(TServiceClient.java:78) >> at backtype.storm.generated.Nimbus$Client.recv_activate(Nimbus.java:210) >> at backtype.storm.generated.Nimbus$Client.activate(Nimbus.java:197) >> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) >> at >> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57) >> at >> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) >> at java.lang.reflect.Method.invoke(Method.java:622) >> at clojure.lang.Reflector.invokeMatchingMethod(Reflector.java:93) >> at clojure.lang.Reflector.invokeInstanceMethod(Reflector.java:28) >> at backtype.storm.command.activate$_main.invoke(activate.clj:7) >> at clojure.lang.AFn.applyToHelper(AFn.java:161) >> at clojure.lang.AFn.applyTo(AFn.java:151) >> at backtype.storm.command.activate.main(Unknown Source) >> >> >> > > >
