[
https://issues.apache.org/activemq/browse/AMQ-1156?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_38819
]
David Martín Clavo commented on AMQ-1156:
-----------------------------------------
Indeed Helmut. When you set socket.tcpNoDelay=true on a client (producer or
consumer) URI, the client's sockets have TCP_NODELAY properly set to true.
But when you try to do the same in activemq.xml for the broker's sockets, it
doesn't work. You have to hard code it.
I found that setting TCP_NODELAY on the producer's sockets is the most
important, but setting it on the broker's and consumer's sockets also
eliminates many timeouts.
Thank you very much for the warning on tcp_abc. I am not using any of those
kernels but it is good to know :)
> option "wireFormat.tcpNoDelayEnabled=true" is ignored
> -----------------------------------------------------
>
> Key: AMQ-1156
> URL: https://issues.apache.org/activemq/browse/AMQ-1156
> Project: ActiveMQ
> Issue Type: Bug
> Components: Documentation, Test Cases, Transport
> Affects Versions: 4.0, 4.1.0, 4.2.0
> Environment: Any.
> Reporter: David Martín Clavo
>
> *A. Description*
> Setting the wireFormat.tcpNoDelayEnabled flag to true in connection uri's
> (for example _"tcp://localhost:61616?wireFormat.tcpNoDelayEnabled"_) has no
> effect:
> The sockets created in ActiveMQ do not have the TcpNoDelay flag set to true.
> You can verify this for example in the following way: insert the line
> _System.err.println(sock.getTcpNoDelay());_ at the end of the method
> _org.apache.activemq.transport.tcp.TcpTransport.initialiseSocket_.
> Also you can try my test at
> http://www.nabble.com/High-latency-for-small-messages-problem-t3159901.html
> which makes latency appear if Nagle's algorithm is not disabled.
> This affects versions 4.0, 4.1.0, and 4.2.0 latest snapshot. I guess 4.0.2 is
> affected too.
> *B. Reason*
> After looking at the source code to understand how the options are set, i saw
> that in the method
> org.apache.activemq.transport.tcp.TcpTransportFactory.compositeConfigure the
> following line appears:
> _Map socketOptions = IntrospectionSupport.extractProperties(options,
> "socket.")_;
> and also that the method to set the TcpNoDelay option in class
> _org.apache.activemq.transport.tcp.TcpTransport_ is called _setTcpNoDelay_
> and not _setTcpNoDelayEnabled_
> Therefore, I tried to write _"socket.tcpNoDelay=true"_ instead of
> _"wireFormat.tcpNoDelayEnabled=true"_ and it *worked!* This is a *workaround*.
> But, _"wireFormat.tcpNoDelayEnabled"_ option is not refused (as
> _"wireFormat.tcpNoDelayTypoEnabled"_ would be). The option
> _"tcpNoDelay=true"_ is not refused either, but DOES NOT work either.
> *C. Fix*
> So, please, for the sake of the poor soul who's next going to try to disable
> Nagle's algorithm without noticing any effect, and will start looking for the
> problem elsewhere (and for the sake of consistency also):
> -Either change the documentation
> (http://www.activemq.org/site/tcp-transport-reference.html and
> http://www.activemq.org/site/configuring-wire-formats.html) to say that one
> should use the _socket.tcpNoDelay_ option, or
> -Change the code to obey the documentation (and make it refuse old options
> like just _"tcpNoDelay"_ who have no effect).
> Personally I would change the documentation ASAP so that people know about
> the _"socket.tcpNoDelay=true"_ workaround until next version, and then change
> the code to recognize the _"wireFormat.tcpNoDelayTypoEnabled"_ option. I'd
> like to point out that the own ActiveMQ test cases use
> _"wireFormat.tcpNoDelayTypoEnabled"_ .
> *D. Related issues*
> This is related to a previous post of mine:
> http://www.nabble.com/High-latency-for-small-messages-problem-t3159901.html
> Doing quick synchronized sends showed a 39ms latency if the broker is running
> in Linux, and a 200ms latency if the broker is in Windows (by the way, if
> someone knows the bottom reason for this exact latency times, please tell
> me). Putting the _"wireFormat.tcpNoDelayTypoEnabled"_ to true in order to
> disable Nagle's algorithm had no effect (because ActiveMQ does not recognize
> it).
> Very probably this is also related to issues
> https://issues.apache.org/activemq/browse/AMQ-1143 and
> https://issues.apache.org/activemq/browse/AMQ-1137 (in this one probably
> Helmutt tried to set the wireFormat.tcpNoDelayTypoEnabled without success
> because of this).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
