storm-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Itai Frenkel <>
Subject RE: Re: storm upgrade issue
Date Fri, 18 Jul 2014 19:10:28 GMT
I'm guessing here...

The old Serialized class came from either files on disk, or from code that had access to the
old class.

Have you tried completely removing all traces of the old version ? (all data files and jar

From your details flow below I assume both nimbus and supervisor are running on the same machine
with access to the same deployment folder.

If not, make sure you have deleted all traces from both machines.


From: 唐思成 <>
Sent: Friday, July 18, 2014 9:41 AM
To: user
Subject: Re: Re: storm upgrade issue

pretty sure that I run the same version on worknode as I do on the maternode

发件人: Harsha
发送时间: 2014-07-18  13:21:08
收件人: user
主题: Re: storm upgrade issue
Does your worker node also have the same storm version installed make sure your older STORM_HOME
is not in PATH.

On Thu, Jul 17, 2014, at 06:39 PM, 唐思成 wrote:
the step i took listed below

1. kill -9 all storm process
2. remove storm directory on zookeeper
3. change storm local dir
4. start nimbus and ui (is fine)
5. start supervisor on a worknode( the nimbus goes down)



发件人: Itai Frenkel
发送时间: 2014-07-18  00:16:21
收件人: storm_user
主题: RE: storm upgrade issue

The message says that SupervisorInfo that your code was compiled with is not compatible with
the SupervisorInfo that was received over the network.

That happens when you have a Serializable class that changes and there is no explicit backwards
compatibility in place.

I would first check that all of your Storm instances are running the same version.

If that does not help I would check that you code is compiled against the correct code version.

Please report your findings, since it's interesting :)

full disclosure - I'm a Storm newbie,


From: 唐思成 <>
Sent: Thursday, July 17, 2014 2:23 PM
To: storm_user
Subject: storm upgrade issue

Hi all:
I try to upgrade storm from 0.9.1 to 0.9.2-incubating, and when the worknode supervisor startup,
the nimbus process goes down, here is what the nimbus.log say:

Before upgrade, I already change storm.local.dir: to a new location and remove storm node
in zookeeper using, however that dont help.


2014-07-17 19:15:29 b.s.d.nimbus [ERROR] Error when processing event
java.lang.RuntimeException: backtype.storm.daemon.common.SupervisorInfo;
local class incompatible: stream classdesc serialVersionUID = 7648414326720210054, local class
serialVersionUID = 7463898661547835557
        at backtype.storm.utils.Utils.deserialize( ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at backtype.storm.cluster$maybe_deserialize.invoke(cluster.clj:200) ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at backtype.storm.cluster$mk_storm_cluster_state$reify__2284.supervisor_info(cluster.clj:299)
        at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) ~[na:na]
        at sun.reflect.NativeMethodAccessorImpl.invoke( ~[na:na]
        at sun.reflect.DelegatingMethodAccessorImpl.invoke(
        at java.lang.reflect.Method.invoke( ~[na:na]
        at clojure.lang.Reflector.invokeMatchingMethod( ~[clojure-1.5.1.jar:na]
        at clojure.lang.Reflector.invokeInstanceMethod( ~[clojure-1.5.1.jar:na]
        at backtype.storm.daemon.nimbus$all_supervisor_info$fn__4715.invoke(nimbus.clj:277)
        at clojure.core$map$fn__4207.invoke(core.clj:2487) ~[clojure-1.5.1.jar:na]
        at clojure.lang.LazySeq.sval( ~[clojure-1.5.1.jar:na]
        at clojure.lang.LazySeq.seq( ~[clojure-1.5.1.jar:na]
        at clojure.lang.RT.seq( ~[clojure-1.5.1.jar:na]
        at clojure.core$seq.invoke(core.clj:133) ~[clojure-1.5.1.jar:na]
        at clojure.core$apply.invoke(core.clj:617) ~[clojure-1.5.1.jar:na]
        at clojure.core$mapcat.doInvoke(core.clj:2514) ~[clojure-1.5.1.jar:na]
        at clojure.lang.RestFn.invoke( ~[clojure-1.5.1.jar:na]
        at backtype.storm.daemon.nimbus$all_supervisor_info.invoke(nimbus.clj:275) ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at backtype.storm.daemon.nimbus$all_scheduling_slots.invoke(nimbus.clj:288) ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at backtype.storm.daemon.nimbus$compute_new_topology__GT_executor__GT_node_PLUS_port.invoke(nimbus.clj:580)
        at backtype.storm.daemon.nimbus$mk_assignments.doInvoke(nimbus.clj:660) ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at clojure.lang.RestFn.invoke( ~[clojure-1.5.1.jar:na]
        at backtype.storm.daemon.nimbus$fn__5210$exec_fn__1396__auto____5211$fn__5216$fn__5217.invoke(nimbus.clj:905)
        at backtype.storm.daemon.nimbus$fn__5210$exec_fn__1396__auto____5211$fn__5216.invoke(nimbus.clj:904)
        at backtype.storm.timer$schedule_recurring$this__1134.invoke(timer.clj:99) ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at backtype.storm.timer$mk_timer$fn__1117$fn__1118.invoke(timer.clj:50) ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at backtype.storm.timer$mk_timer$fn__1117.invoke(timer.clj:42) [storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        at [clojure-1.5.1.jar:na]
        at [na:na]
Caused by: backtype.storm.daemon.common.SupervisorInfo; local
class incompatible: stream classdesc serialVersionUID = 7648414326720210054, local class serialVersionUID
= 7463898661547835557
        at ~[na:na]
        at ~[na:na]
        at ~[na:na]
        at ~[na:na]
        at ~[na:na]
        at ~[na:na]
        at backtype.storm.utils.Utils.deserialize( ~[storm-core-0.9.2-incubating.jar:0.9.2-incubating]
        ... 29 common frames omitted

View raw message