You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by Martín Tabak <ma...@gmail.com> on 2020/04/15 16:38:48 UTC
Help request
Hi community!
I'm facing a disconnection issue using the Qpid python client v0.30.0 for
jobs taking > 40 seconds.
The issue is described on this thread:
https://stackoverflow.com/q/60857609/597394
And there's a MRE here: https://github.com/martintuk/qpid-python-mre.
This is of extreme importance to the project I'm working now, so Id be
really grateful if I could get some help to identify if it's due to a
configuration issue or an actual bug on the library.
Best,
Martin
Re: Help request
Posted by Martín Tabak <ma...@gmail.com>.
Hey Gordon,
I went for the background thread, and is running smoothly now.
Thanks so much for the quick reply!
Martin
El mié., 15 de abril de 2020 15:59, Gordon Sim <gs...@redhat.com> escribió:
> On 15/04/2020 5:38 pm, Martín Tabak wrote:
> > Hi community!
> >
> > I'm facing a disconnection issue using the Qpid python client v0.30.0 for
> > jobs taking > 40 seconds.
> >
> > The issue is described on this thread:
> > https://stackoverflow.com/q/60857609/597394
> >
> > And there's a MRE here: https://github.com/martintuk/qpid-python-mre.
> >
> > This is of extreme importance to the project I'm working now, so Id be
> > really grateful if I could get some help to identify if it's due to a
> > configuration issue or an actual bug on the library.
>
> The reason it is not reconnecting is beacuse when you finally allow the
> event loop to run again (after the sleep) it processes the connection
> close. Reconnect at present for the python client is only triggered when
> the transport under an active connection is lost.
>
> Ideally there would be a way to indicate a list of connection error
> conditions that would also trigger reconnect, but I don't see anything
> there at present.
>
> However, the biggest issue I think is that you are blocking the event
> thread. A better approach would be to use separate threads for the event
> loop and message processing and coordinate via a Queue or similar.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
> For additional commands, e-mail: dev-help@qpid.apache.org
>
>
Re: Help request
Posted by Martín Tabak <ma...@gmail.com>.
Hey Gordon,
I went for the background thread, and is running smoothly now.
Thanks so much for the quick reply!
Martin
El mié., 15 de abril de 2020 15:59, Gordon Sim <gs...@redhat.com> escribió:
> On 15/04/2020 5:38 pm, Martín Tabak wrote:
> > Hi community!
> >
> > I'm facing a disconnection issue using the Qpid python client v0.30.0 for
> > jobs taking > 40 seconds.
> >
> > The issue is described on this thread:
> > https://stackoverflow.com/q/60857609/597394
> >
> > And there's a MRE here: https://github.com/martintuk/qpid-python-mre.
> >
> > This is of extreme importance to the project I'm working now, so Id be
> > really grateful if I could get some help to identify if it's due to a
> > configuration issue or an actual bug on the library.
>
> The reason it is not reconnecting is beacuse when you finally allow the
> event loop to run again (after the sleep) it processes the connection
> close. Reconnect at present for the python client is only triggered when
> the transport under an active connection is lost.
>
> Ideally there would be a way to indicate a list of connection error
> conditions that would also trigger reconnect, but I don't see anything
> there at present.
>
> However, the biggest issue I think is that you are blocking the event
> thread. A better approach would be to use separate threads for the event
> loop and message processing and coordinate via a Queue or similar.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
> For additional commands, e-mail: dev-help@qpid.apache.org
>
>
Re: Help request
Posted by Gordon Sim <gs...@redhat.com>.
On 15/04/2020 5:38 pm, Martín Tabak wrote:
> Hi community!
>
> I'm facing a disconnection issue using the Qpid python client v0.30.0 for
> jobs taking > 40 seconds.
>
> The issue is described on this thread:
> https://stackoverflow.com/q/60857609/597394
>
> And there's a MRE here: https://github.com/martintuk/qpid-python-mre.
>
> This is of extreme importance to the project I'm working now, so Id be
> really grateful if I could get some help to identify if it's due to a
> configuration issue or an actual bug on the library.
The reason it is not reconnecting is beacuse when you finally allow the
event loop to run again (after the sleep) it processes the connection
close. Reconnect at present for the python client is only triggered when
the transport under an active connection is lost.
Ideally there would be a way to indicate a list of connection error
conditions that would also trigger reconnect, but I don't see anything
there at present.
However, the biggest issue I think is that you are blocking the event
thread. A better approach would be to use separate threads for the event
loop and message processing and coordinate via a Queue or similar.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@qpid.apache.org
For additional commands, e-mail: users-help@qpid.apache.org
Re: Help request
Posted by Gordon Sim <gs...@redhat.com>.
On 15/04/2020 5:38 pm, Martín Tabak wrote:
> Hi community!
>
> I'm facing a disconnection issue using the Qpid python client v0.30.0 for
> jobs taking > 40 seconds.
>
> The issue is described on this thread:
> https://stackoverflow.com/q/60857609/597394
>
> And there's a MRE here: https://github.com/martintuk/qpid-python-mre.
>
> This is of extreme importance to the project I'm working now, so Id be
> really grateful if I could get some help to identify if it's due to a
> configuration issue or an actual bug on the library.
The reason it is not reconnecting is beacuse when you finally allow the
event loop to run again (after the sleep) it processes the connection
close. Reconnect at present for the python client is only triggered when
the transport under an active connection is lost.
Ideally there would be a way to indicate a list of connection error
conditions that would also trigger reconnect, but I don't see anything
there at present.
However, the biggest issue I think is that you are blocking the event
thread. A better approach would be to use separate threads for the event
loop and message processing and coordinate via a Queue or similar.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org