I didn't vote +1 bcos I was investigating a potential issue where the
java client getting disconnected due to not completing the protocol
negotiation on time.
Max negotiate time is a new feature introduced by the c++ broker as a
security measure to terminate connections that does not open an amqp
connection within a given time limit.
After investigation with Petr and Leonid we have found the following
information.

1. This happens only on RHEL6 (tested on RHEL5 and 6 ... not sure
about Windows or Solaris)
2. When it happens, a ~3 sec delay is observed btw packets (found
using a protocol trace)
3. This can be reproduced with jars as far as the 0.14 release.

IMO this is not a blocker, but we might need to release note it.
Clearly the default max-negotiate-timeout is not enough for the java client.

However the more serious issue for me is, that this delay **may** be
happening during normal operations there by causing a perf issue.
We haven't had a chance to investigate this aspect fully.

Should we release note this via a JIRA? or should we add it to the
release email and a webpage.

Regards,

Rajith

On Mon, Aug 27, 2012 at 1:30 PM, Justin Ross <jr...@redhat.com> wrote:
> Three votes--I didn't know that, :).  In any case, I feel better with your
> vote.
>
> Now that we have a good crosssection of contributors, I'm ready to close the
> vote end-of-business today (US east coast time).
>
> Justin
>
>
> On Mon, 27 Aug 2012, Robbie Gemmell wrote:
>
>> +1
>>
>> (I have the maven artifacts staged for release, and hadn't voted simply
>> because I coudlnt before the mentioned Friday deadline and you already had
>> the required three +1s to release it if you wanted anyway :P)
>>
>> Robbie
>>
>> On 20 August 2012 16:53, Justin Ross <jr...@redhat.com> wrote:
>>
>>> Hi, folks.  It's time again to vote on our release.  The proposed final
>>> RC
>>> (same revision as RC3, minus the -rc3 version suffix) is available here:
>>>
>>>
>>> http://people.apache.org/~**jross/qpid-0.18/<http://people.apache.org/%7Ejross/qpid-0.18/>
>>>
>>> If you favor releasing the RC3 bits as 0.18, vote +1.  If you have a
>>> reason to think that RC3 is not ready for frelease, vote -1.
>>>
>>> I'll close the vote on Friday.
>>>
>>> Thank you, Alex, Gordon, and Darryl, for testing the RC3 bits and posting
>>> what you found.  And thanks again to Robbie for picking up release tasks
>>> when I could not.
>>>
>>> Thanks!
>>> Justin
>>>
>>> ---
>>> 0.18 release page:
>>> https://cwiki.apache.org/qpid/**018-release.html<https://cwiki.apache.org/qpid/018-release.html>
>>>
>>> ------------------------------**------------------------------**---------
>>> To unsubscribe, e-mail:
>>> dev-unsubscribe@qpid.apache.**org<dev-unsubscr...@qpid.apache.org>
>>> For additional commands, e-mail: dev-h...@qpid.apache.org
>>>
>>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
>
> For additional commands, e-mail: dev-h...@qpid.apache.org
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org

Reply via email to