Re: Problems with pipelining requests
Oups again.
Patrick Julien wrote:
> This does not seem to be printing out bytes, you sure I have the right
> package? Or did the procedure change from the first one?
>
> On Tue, Dec 16, 2008 at 1:54 PM, Jeanfrancois Arcand
> <Jeanfrancois.Arcand_at_sun.com> wrote:
>> Salut,
>>
>> Patrick Julien wrote:
>>> OK, so here's the log, what's interesting here is that I can see that
>>> there is an attempt to handle an OP_READ but that is followed
>>> immediately by a socket close on the server.
>> Looks like there was no bytes available to read. I'm attaching a new patch
>> which will display the bytes read. maybe the client has trouble sending more
>> data?
>>
>> A+
>>
>> -- jeanfrancoois
>>
>>
>>> On Tue, Dec 16, 2008 at 11:56 AM, Jeanfrancois Arcand
>>> <Jeanfrancois.Arcand_at_sun.com> wrote:
>>>> My bad. Proper patch attached.
>>>>
>>>> -- Jeanfrancois
>>>>
>>>> Patrick Julien wrote:
>>>>> You sure about these instructions? This jar file only contains a
>>>>> META-INF directory.
>>>>>
>>>>> On Tue, Dec 16, 2008 at 11:10 AM, Jeanfrancois Arcand
>>>>> <Jeanfrancois.Arcand_at_sun.com> wrote:
>>>>>> Salut,
>>>>>>
>>>>>> I'm attaching the patch which add more logging (let's start with that
>>>>>> one).
>>>>>> Just do:
>>>>>>
>>>>>> % cd ${glassfish.home}/lib
>>>>>> % unzip grizzly-1.0...
>>>>>> % cp appserv-rt.jar appserv-rt.jar.org
>>>>>> % jar uvf appserv-rt.jar com
>>>>>>
>>>>>> restart and execute your test.
>>>>>>
>>>>>> A+
>>>>>>
>>>>>> -- Jeanfrancois
>>>>>>
>>>>>> Jeanfrancois Arcand wrote:
>>>>>>> Salut,
>>>>>>>
>>>>>>> Patrick Julien wrote:
>>>>>>>> On Mon, Dec 15, 2008 at 3:32 PM, Jeanfrancois Arcand
>>>>>>>> <Jeanfrancois.Arcand_at_sun.com> wrote:
>>>>>>>>> Patrick Julien wrote:
>>>>>>>>>>>> Add -Dcom.sun.enterprise.web.connector.grizzly.snoopEnabled=true
>>>>>>>>>>>> and
>>>>>>>>>>>> send
>>>>>>>>>>>> the log here. But I doubt you will find http pipelining support
>>>>>>>>>>>> with
>>>>>>>>>>>> Tomcat
>>>>>>>>>>>> (the parser code here is from them)...I think Jetty might work,
>>>>>>>>>>>> but
>>>>>>>>>>>> I'm
>>>>>>>>>>>> clearly not sure....
>>>>>>>>>> Wait a minute, that doesn't compute for me. Both glassfish and
>>>>>>>>>> Tomcat
>>>>>>>>>> documentation refer to pipelining via the "maxKeepAliveRequests"
>>>>>>>>>> property.
>>>>>>>>>>
>>>>>>>>>> You can see it in the documentation of Tomcat here:
>>>>>>>>>>
>>>>>>>>>> http://tomcat.apache.org/tomcat-6.0-doc/config/http.html
>>>>>>>>>>
>>>>>>>>>> and for glassfish, here:
>>>>>>>>>>
>>>>>>>>>> http://docs.sun.com/app/docs/doc/820-4507/abhco?a=view
>>>>>>>>> OK...http pipelining is not what I means :-) Http pipelining usually
>>>>>>>>> refer
>>>>>>>>> to the ability to sent multiple requests without waiting for the
>>>>>>>>> response.
>>>>>>>>> So you can send R1,R2,R3 in one chunk and then wait for un ordered
>>>>>>>>> response.
>>>>>>>>> I was under the impression you were talking about that behavior.
>>>>>>>> Yes, this is what I am talking about. Send 10 requests, then start
>>>>>>>> accepting responses.
>>>>>>> OK then the links are not related.
>>>>>>>
>>>>>>>>> What you mean here based on the above link is the maximum number of
>>>>>>>>> requests
>>>>>>>>> a client can do on an open connection. The client does R1, wait for
>>>>>>>>> Res1,
>>>>>>>>> then Req2, wait for Req2, etc.
>>>>>>>> No, I really mean sending multiple requests and getting back
>>>>>>>> responses
>>>>>>>> in one shot, if the documentation provided here means something
>>>>>>>> different, than it most certainly fooled me.
>>>>>>>>
>>>>>>>>> Now I understand your issue. I think your request is invalid and
>>>>>>>>> Grizzly
>>>>>>>>> is
>>>>>>>>> waiting for more bytes. Is the snippet you wrote previously contains
>>>>>>>>> the
>>>>>>>>> entire request? I will try to reproduce the issue using that one and
>>>>>>>>> telnet.
>>>>>>>> No, as I said before, I am getting 2 to 4 responses then nothing from
>>>>>>>> the 10 requests. Just a long blocking read followed by a disconnect.
>>>>>>> Ok let me instrument a patch for you to see what happening. I hope I
>>>>>>> don't
>>>>>>> have to drive to see the problem live (at least not today as I will
>>>>>>> need
>>>>>>> a
>>>>>>> boat :-)). Gives me 1 hour or so.
>>>>>>>
>>>>>>> A+
>>>>>>>
>>>>>>> -- Jeanfrancois
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>>> A+
>>>>>>>>>
>>>>>>>>> -- jeanfrancois
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> ---------------------------------------------------------------------
>>>>>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_grizzly.dev.java.net
>>>>>>>>>> For additional commands, e-mail: users-help_at_grizzly.dev.java.net
>>>>>>>>>>
>>>>>>>>> ---------------------------------------------------------------------
>>>>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_grizzly.dev.java.net
>>>>>>>>> For additional commands, e-mail: users-help_at_grizzly.dev.java.net
>>>>>>>>>
>>>>>>>>>
>>>>>>> ---------------------------------------------------------------------
>>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_grizzly.dev.java.net
>>>>>>> For additional commands, e-mail: users-help_at_grizzly.dev.java.net
>>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_grizzly.dev.java.net
>>>>>> For additional commands, e-mail: users-help_at_grizzly.dev.java.net
>>>>>>
>>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe_at_grizzly.dev.java.net
>>>> For additional commands, e-mail: users-help_at_grizzly.dev.java.net
>>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe_at_grizzly.dev.java.net
>>> For additional commands, e-mail: users-help_at_grizzly.dev.java.net
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_grizzly.dev.java.net
>> For additional commands, e-mail: users-help_at_grizzly.dev.java.net
>>
>
>
>