activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alan Conway (JIRA)" <j...@apache.org>
Subject [jira] [Created] (AMQ-5453) AMQP interop: Active broker does not respect the "drain" flag.
Date Mon, 24 Nov 2014 20:36:13 GMT
Alan Conway created AMQ-5453:
--------------------------------

             Summary: AMQP interop: Active broker does not respect the "drain" flag.
                 Key: AMQ-5453
                 URL: https://issues.apache.org/jira/browse/AMQ-5453
             Project: ActiveMQ
          Issue Type: Bug
          Components: AMQP
    Affects Versions: 6.0.0
            Reporter: Alan Conway


The drain flag on the AMQP flow performative allows a client to request confirmation that
it has received the last available message that it has credit to receive.

To reproduce using the qpid-send, qpid-receive clients from http://svn.apache.org/repos/asf/qpid/trunk/qpid/.
Create a JMS queue 'foo' on the active broker then run:

$ qpid-send -a jms.queue.foo -b localhost:5455 --content-string XXX --connection-options='{protocol:amqp1.0}'
$ qpid-receive -a jms.queue.foo -b localhost:5455 --connection-options='{protocol:amqp1.0}'
--log-enable trace+:Protocol

qpid-receive hangs, the  last line of output is:
2014-11-24 15:15:46 [Protocol] trace [58e8ee08-0f33-426b-b77a-450f7c3d976c]: 0 -> @flow(19)
[next-incoming-id=2, incoming-window=2147483647, next-outgoing-id=0, outgoing-window=0, handle=0,
delivery-count=1, link-credit=1, drain=true]

This shows that qpid-receive sent a flow with drain=true but never received a response.

Why is this important? Without the drain flag it is impossible for a client to implement the
simple behavior "get the next message" correctly. The flow response tells the client immediately
"there are no more messages available for you". Without it the client can only use a timeout
which is unreliable (if too short the client may give up while the message is in flight) and
inefficient (if too long the client will wait needlessly for messages that the broker knows
are not presently available)

The spec 2.6.7 is a little ambiguous about whether this is a SHOULD or a MUST behavior but
without it it is impossible to implement the use cases described in the following section.

AMQP 1.0 specification 2.7.6
drain
The drain flag indicates how the sender SHOULD behave when insufficient messages are available
to consume the current link-credit. If set, the sender will (after sending all available messages)
advance the delivery-count as much as possible, consuming all link-credit, and send the flow
state to the receiver. Only the receiver can independently modify this field. The sender's
value is always the last known value indicated by the receiver.

If the link-credit is less than or equal to zero, i.e., the delivery-count is the same as
or greater than the delivery-limit, a sender MUST NOT send more messages. If the link-credit
is reduced by the receiver when transfers are in-flight, the receiver MAY either handle the
excess messages normally or detach the link with a transfer-limit-exceeded error code.

Figure 2.40: Flow Control

     +----------+                                    +----------+
     |  Sender  |---------------transfer------------>| Receiver |
     +----------+                                    +----------+
      \        / <----------------flow--------------- \        /
       +------+                                        +------+
          |
          |
          |
 if link-credit <= 0 then pause 
      
If the sender's drain flag is set and there are no available messages, the sender MUST advance
its delivery-count until link-credit is zero, and send its updated flow state to the receiver.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message