Mailing-List: contact users-help@qpid.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@qpid.apache.org
Received-SPF: pass (nike.apache.org: domain of bpedman@gmail.com designates
 209.85.213.42 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <CAAiPYj08CYfBVFx4R2ooDQk_Hj9Rj1VCwD=yYh6uBiPWpLbB1g@mail.gmail.com>
References: 
 <CAAiPYj3RXskG+KbLQrmgMRa3ZMJXv4H3AG5pHNrMF2qqRNsdNg@mail.gmail.com>
 <4E78F998.1000704@redhat.com>
 <CAAiPYj0MCj3CUAesR9r7k44Fe7TRSVJ6Hrbp1tCOR2pmfK_34A@mail.gmail.com>
 <4E791165.9040409@redhat.com>
 <CAAiPYj08CYfBVFx4R2ooDQk_Hj9Rj1VCwD=yYh6uBiPWpLbB1g@mail.gmail.com>
From: Brandon Pedersen <bpedman@gmail.com>
Date: Tue, 20 Sep 2011 16:42:13 -0600
Message-ID: 
 <CAAiPYj3W=oeZko7D3iNwM8z_-cH_V18RySHNVuckrFthGF5fag@mail.gmail.com>
Subject: Re: routing messages to another broker that has lost connection
To: users@qpid.apache.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Is there any way to verify that it is indeed ack'ing those messages?

On Tue, Sep 20, 2011 at 4:40 PM, Brandon Pedersen <bpedman@gmail.com> wrote=
:
> Yeah, I used 1 for ack
>
> On Tue, Sep 20, 2011 at 4:19 PM, Ted Ross <tross@redhat.com> wrote:
>> Did you use a large number for the --ack argument in the route? =A0Try u=
sing
>> 1.
>>
>> It appears that the ack-batching is quite literal and the downstream bro=
ker
>> *only* acks on modulo-N messages (i.e. it doesn't ack gratuitously after=
 a
>> period of inactivity).
>>
>> -Ted
>>
>> On 09/20/2011 05:22 PM, Brandon Pedersen wrote:
>>>
>>> Thank you! This almost works ;) I was worried about doing this because
>>> I didn't want the local source queue to fill up because I wouldn't be
>>> popping messages off it but it looks like it handles it properly and
>>> removes them from the source queue when they have been pulled to the
>>> destination exchange....awesome.
>>>
>>> However, I shutdown the destination broker, use spout to generate a
>>> message, and then restart the destination broker. The message that was
>>> missed while the broker was down is received but so is every past
>>> message that was already received before the broker went down but with
>>> redelivered set to true. I don't even know where these messages come
>>> from because the queue on the source broker does not have any of those
>>> messages. How can I get it to not receive the messages it has already
>>> received?
>>>
>>> Thanks,
>>>
>>> -Brandon
>>>
>>> On Tue, Sep 20, 2011 at 2:37 PM, Ted Ross<tross@redhat.com> =A0wrote:
>>>>
>>>> Brandon,
>>>>
>>>> You can do this with a queue route (i.e. the downstream broker subscri=
bes
>>>> to
>>>> an existing queue).
>>>>
>>>> Static exchange routes may be durable, meaning they will reappear afte=
r a
>>>> restart, but their queues are always transient. =A0Furthermore, when t=
he
>>>> route
>>>> is down, there is no queue to hold the messages produced.
>>>>
>>>> If you create a queue, then establish a queue route to draw from that
>>>> queue,
>>>> the messages will accumulate there when the downstream broker is
>>>> disconnected.
>>>>
>>>> When you set up the route in qpid-route, use the --ack N option with a=
n N
>>>> greater than zero. =A0This causes the brokers to acknowledge messages =
and
>>>> will
>>>> ensure that in-doubt messages (i.e. in flight during the failure) will=
 be
>>>> re-transmitted upon reconnection.
>>>>
>>>> The qpid-route syntax is:
>>>>
>>>> =A0 =A0qpid-route [OPTIONS] queue add<dest-broker> =A0<src-broker> =A0=
<exchange>
>>>> <queue> =A0[mechanism]
>>>>
>>>> where<exchange> =A0is the name of the exchange on<dest-broker> =A0wher=
e
>>>> messages
>>>> will be delivered to and<queue> =A0is the name of the queue on<src-bro=
ker>
>>>> where messages will be received from.
>>>>
>>>> -Ted
>>>>
>>>> On 09/20/2011 03:46 PM, Brandon Pedersen wrote:
>>>>>
>>>>> I am trying to get something like this working:
>>>>>
>>>>> - publish a message to a local exchange
>>>>> - have the message get pushed to a remote exchange (or
>>>>> pulled...preferably pushed)
>>>>> and
>>>>> - if the remote broker goes down, when it comes back up it will
>>>>> receive all messages it missed while it was down
>>>>>
>>>>> I can get the first 2 working with a simple static route. However,
>>>>> when the remote broker goes down it does not receive any of the
>>>>> messages that were sent while it was down. Is there any way to do
>>>>> this? And everything is being marked as durable (the queue, the
>>>>> exchange, the binding, and the route)
>>>>>
>>>>> Thanks,
>>>>>
>>>>> -Brandon
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> Apache Qpid - AMQP Messaging Implementation
>>>>> Project: =A0 =A0 =A0http://qpid.apache.org
>>>>> Use/Interact: mailto:users-subscribe@qpid.apache.org
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> Apache Qpid - AMQP Messaging Implementation
>>>> Project: =A0 =A0 =A0http://qpid.apache.org
>>>> Use/Interact: mailto:users-subscribe@qpid.apache.org
>>>>
>>>>
>>> ---------------------------------------------------------------------
>>> Apache Qpid - AMQP Messaging Implementation
>>> Project: =A0 =A0 =A0http://qpid.apache.org
>>> Use/Interact: mailto:users-subscribe@qpid.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> Apache Qpid - AMQP Messaging Implementation
>> Project: =A0 =A0 =A0http://qpid.apache.org
>> Use/Interact: mailto:users-subscribe@qpid.apache.org
>>
>>
>

---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project:      http://qpid.apache.org
Use/Interact: mailto:users-subscribe@qpid.apache.org