Mailing-List: contact users-help@activemq.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@activemq.apache.org
Received-SPF: pass (athena.apache.org: domain of jbrook@upcbroadband.com
 designates 62.179.123.10 as permitted sender)
From: "Brook, James" <jbrook@upcbroadband.com>
To: "users@activemq.apache.org" <users@activemq.apache.org>
Date: Tue, 17 Apr 2012 12:41:10 +0200
Subject: Re: Producer blocks when trying to send async JMS messages to
 failed broker
Thread-Topic: Producer blocks when trying to send async JMS messages to
 failed broker
Thread-Index: Ac0chpzTBjFrXoLgRK24kt+ROf/S+A==
Message-ID: <DFE7C639-DEE8-4D63-964E-C7ADEE357033@upcbroadband.com>
References: <AFC3EE89-A5CA-402D-AD9D-873DFDF84622@upcbroadband.com>
 <CAH+vQmNZS-BbuD+vRy6WwCxp7cfRF=T6=6c-0PcL-UupG+wvnA@mail.gmail.com>
In-Reply-To: 
 <CAH+vQmNZS-BbuD+vRy6WwCxp7cfRF=T6=6c-0PcL-UupG+wvnA@mail.gmail.com>
Accept-Language: en-US
Content-Language: en-US
acceptlanguage: en-US
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Thanks Gary. I am still not completely sure I get it. I am using failover b=
ecause I *do* want my client to reconnect when the broker comes back up. Ho=
wever, when it's down I don't want to queue messages or block anything - th=
ey can be thrown away. My messages really are 'fire and forget'. These lose=
 their meaning if not delivered to their ultimate destination within 2 seco=
nds or so.

My expectation was that I would be able to asynchronously send messages and=
 never block. I have tried using non-persistent messages, and telling the t=
ransport not to track messages.

If I do what you suggest won't the failover transport wait 10ms, try to rec=
onnect once and then give up forever? Wouldn't that mean someone has to tak=
e manual action to reestablish the connection?

Sorry if I am missing the point.=20

On Apr 17, 2012, at 12:32 PM, Gary Tully wrote:

> configure the failover maxReconnectAttempts and the failure to connect
> will bubble up to the application where you can ignore it.
> The assumption of failover is that you want to 'transparently'
> maintain the connection and block pending reconnect
>=20
> failover:(...)? maxReconnectAttempts=3D1
>=20
> http://activemq.apache.org/failover-transport-reference.html
>=20
> On 16 April 2012 19:52, Brook, James <jbrook@upcbroadband.com> wrote:
>> I am trying to make sure my application is resilient. It sends JMS messa=
ges to a remote ActiveMQ broker. Unfortunately when that broker goes down t=
he client starts to block even though I am using the Camel asyncSendBody me=
thod. See below for the stack trace.
>>=20
>> The ActiveMQ version is 5.3.0 although I have the same problem with the =
most recent stable release.
>>=20
>> This is the argument I give to the asyncSendBody method:
>>=20
>>    jms:queue:myQueue?jmsMessageType=3DText&timeToLive=3D5000
>>=20
>> My connection factory uses this URL:
>>=20
>>    failover:(tcp://my.ip:61613,tcp://my:ip:61613)?useExponentialBackOff=
=3Dtrue&randomize=3Dfalse&timeout=3D1000
>>=20
>> As the broker starts to fail I see lots of the following errors in the l=
og:
>>=20
>>    org.apache.camel.processor.UnitOfWorkProcessor  - Caught unhandled ex=
ception while processing ExchangeId: ID-james-upcit-ds-upc-biz-64404-133460=
1186430-0-38
>> IOException: Failover timeout of 1000 ms reached
>>=20
>>   org.apache.activemq.transport.failover.FailoverTransport  - Failover t=
imed out after 29998ms
>>=20
>> Then after attempting to send quite a few more messages the blocking sta=
rts. I can't go live with my application in this state and I really don't w=
ant to put another queue in place to decouple the JMS producer. Is there so=
mething I am doing wrong? If the broker is down I am quite happy for these =
messages to be lost - they lose their meaning if not delivered immediately.
>>=20
>> Thanks
>> James
>>=20
>> at java.lang.Object.wait(Native Method)
>>        at org.apache.activemq.transport.failover.FailoverTransport.onewa=
y(FailoverTransport.java:433)
>>        at org.apache.activemq.transport.MutexTransport.oneway(MutexTrans=
port.java:40)
>>        at org.apache.activemq.transport.ResponseCorrelator.asyncRequest(=
ResponseCorrelator.java:74)
>>        at org.apache.activemq.transport.ResponseCorrelator.request(Respo=
nseCorrelator.java:79)
>>        at org.apache.activemq.ActiveMQConnection.syncSendPacket(ActiveMQ=
Connection.java:1244)
>>        at org.apache.activemq.ActiveMQConnection.ensureConnectionInfoSen=
t(ActiveMQConnection.java:1350)
>>        at org.apache.activemq.ActiveMQConnection.createSession(ActiveMQC=
onnection.java:300)
>>        at org.springframework.jms.support.JmsAccessor.createSession(JmsA=
ccessor.java:196)
>>        at org.springframework.jms.core.JmsTemplate.execute(JmsTemplate.j=
ava:457)
>>        at org.apache.camel.component.jms.JmsConfiguration$CamelJmsTempla=
te.send(JmsConfiguration.java:185)
>>        at org.apache.camel.component.jms.JmsProducer.doSend(JmsProducer.=
java:359)
>>        at org.apache.camel.component.jms.JmsProducer.processInOnly(JmsPr=
oducer.java:313)
>>        at org.apache.camel.component.jms.JmsProducer.process(JmsProducer=
.java:111)
>>        at org.apache.camel.processor.UnitOfWorkProcessor.processAsync(Un=
itOfWorkProcessor.java:150)
>>        at org.apache.camel.processor.UnitOfWorkProcessor.process(UnitOfW=
orkProcessor.java:117)
>>        at org.apache.camel.util.AsyncProcessorHelper.process(AsyncProces=
sorHelper.java:99)
>>        at org.apache.camel.processor.DelegateAsyncProcessor.process(Dele=
gateAsyncProcessor.java:86)
>>        at org.apache.camel.processor.UnitOfWorkProducer.process(UnitOfWo=
rkProducer.java:63)
>>        at org.apache.camel.impl.ProducerCache$2.doInProducer(ProducerCac=
he.java:352)
>>        at org.apache.camel.impl.ProducerCache$2.doInProducer(ProducerCac=
he.java:324)
>>        at org.apache.camel.impl.ProducerCache.doInProducer(ProducerCache=
.java:223)
>>        at org.apache.camel.impl.ProducerCache.sendExchange(ProducerCache=
.java:324)
>>        at org.apache.camel.impl.ProducerCache.send(ProducerCache.java:16=
9)
>>        at org.apache.camel.impl.DefaultProducerTemplate.send(DefaultProd=
ucerTemplate.java:111)
>>        at org.apache.camel.impl.DefaultProducerTemplate.sendBody(Default=
ProducerTemplate.java:124)
>>        at org.apache.camel.impl.DefaultProducerTemplate$14.call(DefaultP=
roducerTemplate.java:596)
>>        at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:=
303)
>>        at java.util.concurrent.FutureTask.run(FutureTask.java:138)
>>        at java.util.concurrent.ThreadPoolExecutor$CallerRunsPolicy.rejec=
tedExecution(ThreadPoolExecutor.java:1746)
>>        at java.util.concurrent.ThreadPoolExecutor.reject(ThreadPoolExecu=
tor.java:767)
>>        at java.util.concurrent.ThreadPoolExecutor.execute(ThreadPoolExec=
utor.java:658)
>>        at java.util.concurrent.AbstractExecutorService.submit(AbstractEx=
ecutorService.java:92)
>>        at org.apache.camel.impl.DefaultProducerTemplate.asyncSendBody(De=
faultProducerTemplate.java:601)
>>        at org.apache.camel.impl.DefaultProducerTemplate.asyncSendBody(De=
faultProducerTemplate.java:458)
>=20
>=20
>=20
> --=20
> http://fusesource.com
> http://blog.garytully.com