Thank you for your reply.There is no any exception in the logs of the live
leader.The following is part of the log:
2016-05-30 21:37:07 b.s.d.nimbus [INFO] not a leader, skipping
assignments2016-05-30 21:37:07 b.s.d.nimbus [INFO] not a leader, skipping
cleanup2016-05-30 21:37:18 b.s.zookeeper [INFO] stormtest243.local gained
leadership, checking if it has all the topology code locally.2016-05-30
21:37:18 b.s.zookeeper [INFO] active-topology-ids [xxx-1-1464368723]
local-topology-ids [xxx-1-1464368723] diff-topology []2016-05-30 21:37:18
b.s.zookeeper [INFO] Accepting leadership, all active topology found
localy.2016-05-30 21:37:29 b.s.d.nimbus [INFO] [req 2] Access from: principal:
op:getClusterInfo2016-05-30 21:37:29 b.s.d.nimbus [INFO] [req 1] Access from:
principal: op:getClusterInfo2016-05-30 21:37:29 b.s.d.nimbus [INFO] [req 5]
Access from: principal: op:getClusterInfo
Regards,devopts
> From: abhishc...@gmail.com
> Date: Mon, 30 May 2016 15:00:13 +0530
> Subject: Re: Storm UI fail to find the nimbus leader sometimes!!!
> To: dev@storm.apache.org
>
> can you check the logs of live leader just after refreshing UI and see if
> there is any exception?
>
> On Mon, May 30, 2016 at 2:52 PM, devopts <devo...@outlook.com> wrote:
>
> > Hi all, Here,there are two nimbuses in the cluster(stormtest242.local
> > and stormtest243.local).The nimbus[stormtest242.local] is a leader and the
> > other one[stormtest243.local] is a follower when the cluster startup.And
> > the Storm HA is ok. But, when the storm cluster runs for a long time ,
> > the ui can't find the nimbus leader sometimes.That is the following
> > exception on the UI:backtype.storm.utils.NimbusLeaderNotFoundException:
> > Found nimbuses [NimbusSummary(host:stormtest242.local, port:6627,
> > uptime_secs:283, isLeader:false, version:0.10.0)] none of which is elected
> > as leader, please try again after some time. at
> > backtype.storm.utils.NimbusClient.getConfiguredClientAs(NimbusClient.java:85)
> > at backtype.storm.ui.core$nimbus_summary.invoke(core.clj:581) at
> > backtype.storm.ui.core$fn__10920.invoke(core.clj:1030)
> > orNo nimbus leader participant host found, have you started your nimbus
> > hosts?
> > When i view the all nimbus.log in the cluster,i found the old nimbus
> > leader[stormtest242.local] was dead and the new nimbus
> > leader[stormtest243.local] is running.But the UI always connect to dead
> > leader[stormtest242.local] rather than the live leader[stormtest243.local]
> > ,even if i restart the ui process.
> > So,what's wrong with the nimbus HA? and why the ui always connect to the
> > dead leader?
> >
> >
> >
>
>
>
>
> --
> Regards,
> Abhishek Agarwal
