I have had your problem. Here is an example: Nimbus.Client client = NimbusClient.getConfiguredClient(conf).getClient();
client.getClusterInfo().get_supervisors_size() The nimbus is trying to communicate with all the supervisors that listening to it. You need to make sure your supervisors can have port 6627 open. This is nimbus.thrift.port If you are using DHCP you need port 3772 open too. Try connecting to your Nimbus and pinging a supervisor. If you can’t it means the supervisor doesn’t have the nimbus.thrift.port open. On 3 November 2016 at 13:41:09, Mostafa Gomaa ([email protected]) wrote: Thanks for the reply, Matt. Any ideas which ports need to opened on the remote machine and the local machine. Currently 6627, 6700, 6701, 6702, 6703 are open. Is there any port that I missed? On Thu, Nov 3, 2016 at 2:37 PM, Matt Lowe <[email protected]> wrote: > Your topology is trying to communicate with itself to get the cluster > information. > Are you behind a security group with unopened ports? > > This is usually port/network related. > > > On 3 November 2016 at 13:24:01, Mostafa Gomaa ([email protected]) wrote: > > Dear all, > > I am trying to submit a topology on an azure machine. Everything is setup > and configured. However, The process takes way too long and eventually it > times out. This is the log taken from access-nimbus.log > > 2016-11-03 12:19:00.743 o.a.s.l.ThriftAccessLogger [INFO] Request ID: 39 > access from: principal: operation: getClusterInfo > > > and It just stays like that until it times out. Anyone faced this issue > with azure before? Is there any extra configuration for azure with storm > that I have to be aware of? > > Thanks. > >
