Robert, Running a topology in local mode launches in-process zookeeper which by default binds to 2000. https://github.com/apache/storm/blob/master/storm-core/src/jvm/org/apache/storm/zookeeper/Zookeeper.java#L284
@Rodrigo It is possible that you are killing the local cluster just after you spawn it. It is difficult to say without having code at hand. On Thu, Feb 25, 2016 at 2:37 AM, Bobby Evans <ev...@yahoo-inc.com.invalid> wrote: > Inside the log for your worker it shows that you have the config > "storm.zookeeper.port" 2000, > > Not sure where it came from, but it is in your config. > - Bobby > > On Wednesday, February 24, 2016 3:04 PM, Bobby Evans < > ev...@yahoo-inc.com> wrote: > > > the master branch has been going through a lot of changes lately. We are > in the process of merging with the JStorm project and as such we expect > there to be some instability. We are trying to keep the build so it > continues to pass all tests and could be released if need be, but as > expected with this type of things bugs do creep in. > > I am not sure what/why some of the code may be trying to attach to port > 2000. I have not seen anything with that lately, and I just ran a few > tests this morning on a single node distributed cluster that showed no > problems. > The existing problems that I am aware of, that I still need to file JIRA > for is that the UI has a casting exception, and that performance has > degraded quite a lot, and we need to look into it (most likely a reflection > issue with clojure, but I don't know for sure). > > Another thing to be aware of is that between 0.10, and 1.0, which is still > being worked on, the packages all moved to org.apache.storm, from > backtype.storm. This does impact your config in a few places that you > might need to update for your cluster to run. Also if you don't recompile > your topology against the new APIs, or use our scary hack to rewrite your > topology jar, your topology being submitted will not run. > > - Bobby > > On Wednesday, February 24, 2016 1:58 PM, Rodrigo Valladares < > rodrigovallada...@gmail.com> wrote: > > > I was trying to build the current github version and run it to do some > experiments with the new LoadAwareGrouping, but I am having some problems > running topologies with it. Nimbus and supervisor start and from the UI you > see workers and executors are all allocated, but no tuple is processed. I > tried to run this build in local mode and I am also having problems. I am > attaching the worker log when I executed it in local mode. Looks like there > is a problem in communicating with the zookeeper when launching the worker, > but I do not understand very well why it is trying to connect to port 2000 > in "20599 [Thread-13-SendThread(localhost:2000)] INFO > o.a.s.s.o.a.z.ClientCnxn - Socket connection established to localhost/ > 127.0.0.1:2000, initiating session" since zookeeper is in port 2181. I do > not know how to interpret this log. I downloaded the latest 0.10.0 and run > it in the same cluster with the same configurations without a problem. > Is this a bug in the recent version? Do anyone have any idea how to solve > this? > > Thank you,Rodrigo Valladares Cotta > Master's Student, Computer Science > University of Nebraska-Lincoln > > > > > > -- Regards, Abhishek Agarwal
