You are viewing a plain text version of this content. The canonical link for it is here.
Posted to proton@qpid.apache.org by Tomáš Šoltys <to...@gmail.com> on 2014/02/07 09:04:34 UTC

Re: Heartbeat

Hi Ken,

I have just opened a new JIRA as you suggested.
https://issues.apache.org/jira/browse/PROTON-512

My idea is that there would be one global idle timeout for whole messenger
and this timeout will be used for all connections.

Regards,
Tomas


2013-11-11 Ken Giusti <kg...@redhat.com>:

> Hi Tomas,
>
> ----- Original Message -----
> > From: "Tomáš Šoltys" <to...@gmail.com>
> > To: proton@qpid.apache.org
> > Sent: Monday, November 11, 2013 6:43:06 AM
> > Subject: Re: Heartbeat
> >
> > Hi Ken,
> >
> > thanks for your answer.
> > And is there a way how to set it for pn_messenger_t
> >
>
> No - not at present.  Messenger doesn't give the application access to the
> transport since Messenger hides all that connection-related stuff.
>
> I'd recommend opening a JIRA requesting this feature so we won't forget
> about it.
>
> https://issues.apache.org/jira/browse/PROTON
>
> thanks.
>
> -K
>
>
>
> > Thanks,
> > Tomas
> >
> >
> > 2013/11/8 Ken Giusti <kg...@redhat.com>
> >
> > > Hi Tomas,
> > >
> > > The C implementation of proton allows you to set an idle time out for a
> > > connection as described in the AMQP 1.0 spec.  This value is used to
> > > generate "null" frames (ie. frames with no bodies) on idle connections
> as
> > > to not expire the timeout.   The connection will be dropped if the
> local
> > > side does not receive any traffic for it's configured timeout period.
> > >
> > > The C api is available on the transport object - see engine.h:
> > >
> > > /* timeout of zero means "no timeout" */
> > > PN_EXTERN pn_millis_t pn_transport_get_idle_timeout(pn_transport_t
> > > *transport);
> > > PN_EXTERN void pn_transport_set_idle_timeout(pn_transport_t *transport,
> > > pn_millis_t timeout);
> > > PN_EXTERN pn_millis_t
> pn_transport_get_remote_idle_timeout(pn_transport_t
> > > *transport);
> > >
> > >
> > > I don't think the java implementation has this - yet.  See
> > > https://issues.apache.org/jira/browse/PROTON-111
> > >
> > >
> > >
> > > ----- Original Message -----
> > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > To: proton@qpid.apache.org
> > > > Sent: Friday, November 8, 2013 6:06:15 AM
> > > > Subject: Heartbeat
> > > >
> > > > Hi,
> > > >
> > > > I am looking for a way how to manually specify heartbeat for a
> > > connection.
> > > >
> > > > Is there a way how to do it using proton?
> > > >
> > > > Thanks,
> > > >
> > > > Tomas Soltys
> > > >
> > >
> > > --
> > > -K
> > >
> >
> >
> >
> > --
> > Tomáš Šoltys
> > tomas.soltys@gmail.com
> > http://www.range-software.com
> > (+420) 776-843-663
> >
>
> --
> -K
>



-- 
Tomáš Šoltys
tomas.soltys@gmail.com
http://www.range-software.com
(+420) 776-843-663

Re: Heartbeat

Posted by Rafael Schloming <rh...@alum.mit.edu>.
No, Messenger.setBlocking(True/False) is independent of
Messenger.setTimeout(...).

--Rafael


On Fri, Feb 7, 2014 at 10:29 AM, Ken Giusti <kg...@redhat.com> wrote:

> Would that force an app to use blocking operations if it desires idle
> timeout?  IOW, should the need for idle monitoring determine the choice of
> blocking vs non-blocking?
>
> I probably don't fully grasp the implications....
>
> -K
>
>
> ----- Original Message -----
> > From: "Rafael Schloming" <rh...@alum.mit.edu>
> > To: proton@qpid.apache.org
> > Sent: Friday, February 7, 2014 9:23:20 AM
> > Subject: Re: Heartbeat
> >
> > There is already a single per messenger timeout for blocking operations.
> It
> > should be a pretty trivial patch to set that same value as the idle
> timeout
> > on each connection.
> >
> > --Rafael
> >
> >
> > On Fri, Feb 7, 2014 at 8:26 AM, Ken Giusti <kg...@redhat.com> wrote:
> >
> > > Thanks Tomáš - I agree with a single per-messenger timeout approach.
> > >
> > >
> > > ----- Original Message -----
> > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > To: proton@qpid.apache.org
> > > > Sent: Friday, February 7, 2014 3:04:34 AM
> > > > Subject: Re: Heartbeat
> > > >
> > > > Hi Ken,
> > > >
> > > > I have just opened a new JIRA as you suggested.
> > > > https://issues.apache.org/jira/browse/PROTON-512
> > > >
> > > > My idea is that there would be one global idle timeout for whole
> > > messenger
> > > > and this timeout will be used for all connections.
> > > >
> > > > Regards,
> > > > Tomas
> > > >
> > > >
> > > > 2013-11-11 Ken Giusti <kg...@redhat.com>:
> > > >
> > > > > Hi Tomas,
> > > > >
> > > > > ----- Original Message -----
> > > > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > > > To: proton@qpid.apache.org
> > > > > > Sent: Monday, November 11, 2013 6:43:06 AM
> > > > > > Subject: Re: Heartbeat
> > > > > >
> > > > > > Hi Ken,
> > > > > >
> > > > > > thanks for your answer.
> > > > > > And is there a way how to set it for pn_messenger_t
> > > > > >
> > > > >
> > > > > No - not at present.  Messenger doesn't give the application
> access to
> > > the
> > > > > transport since Messenger hides all that connection-related stuff.
> > > > >
> > > > > I'd recommend opening a JIRA requesting this feature so we won't
> forget
> > > > > about it.
> > > > >
> > > > > https://issues.apache.org/jira/browse/PROTON
> > > > >
> > > > > thanks.
> > > > >
> > > > > -K
> > > > >
> > > > >
> > > > >
> > > > > > Thanks,
> > > > > > Tomas
> > > > > >
> > > > > >
> > > > > > 2013/11/8 Ken Giusti <kg...@redhat.com>
> > > > > >
> > > > > > > Hi Tomas,
> > > > > > >
> > > > > > > The C implementation of proton allows you to set an idle time
> out
> > > for a
> > > > > > > connection as described in the AMQP 1.0 spec.  This value is
> used
> > > to
> > > > > > > generate "null" frames (ie. frames with no bodies) on idle
> > > connections
> > > > > as
> > > > > > > to not expire the timeout.   The connection will be dropped if
> the
> > > > > local
> > > > > > > side does not receive any traffic for it's configured timeout
> > > period.
> > > > > > >
> > > > > > > The C api is available on the transport object - see engine.h:
> > > > > > >
> > > > > > > /* timeout of zero means "no timeout" */
> > > > > > > PN_EXTERN pn_millis_t
> pn_transport_get_idle_timeout(pn_transport_t
> > > > > > > *transport);
> > > > > > > PN_EXTERN void pn_transport_set_idle_timeout(pn_transport_t
> > > *transport,
> > > > > > > pn_millis_t timeout);
> > > > > > > PN_EXTERN pn_millis_t
> > > > > pn_transport_get_remote_idle_timeout(pn_transport_t
> > > > > > > *transport);
> > > > > > >
> > > > > > >
> > > > > > > I don't think the java implementation has this - yet.  See
> > > > > > > https://issues.apache.org/jira/browse/PROTON-111
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > ----- Original Message -----
> > > > > > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > > > > > To: proton@qpid.apache.org
> > > > > > > > Sent: Friday, November 8, 2013 6:06:15 AM
> > > > > > > > Subject: Heartbeat
> > > > > > > >
> > > > > > > > Hi,
> > > > > > > >
> > > > > > > > I am looking for a way how to manually specify heartbeat for
> a
> > > > > > > connection.
> > > > > > > >
> > > > > > > > Is there a way how to do it using proton?
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > >
> > > > > > > > Tomas Soltys
> > > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > -K
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Tomáš Šoltys
> > > > > > tomas.soltys@gmail.com
> > > > > > http://www.range-software.com
> > > > > > (+420) 776-843-663
> > > > > >
> > > > >
> > > > > --
> > > > > -K
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Tomáš Šoltys
> > > > tomas.soltys@gmail.com
> > > > http://www.range-software.com
> > > > (+420) 776-843-663
> > > >
> > >
> > > --
> > > -K
> > >
> >
>
> --
> -K
>

Re: Heartbeat

Posted by Ken Giusti <kg...@redhat.com>.
Would that force an app to use blocking operations if it desires idle timeout?  IOW, should the need for idle monitoring determine the choice of blocking vs non-blocking?

I probably don't fully grasp the implications....

-K


----- Original Message -----
> From: "Rafael Schloming" <rh...@alum.mit.edu>
> To: proton@qpid.apache.org
> Sent: Friday, February 7, 2014 9:23:20 AM
> Subject: Re: Heartbeat
> 
> There is already a single per messenger timeout for blocking operations. It
> should be a pretty trivial patch to set that same value as the idle timeout
> on each connection.
> 
> --Rafael
> 
> 
> On Fri, Feb 7, 2014 at 8:26 AM, Ken Giusti <kg...@redhat.com> wrote:
> 
> > Thanks Tomáš - I agree with a single per-messenger timeout approach.
> >
> >
> > ----- Original Message -----
> > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > To: proton@qpid.apache.org
> > > Sent: Friday, February 7, 2014 3:04:34 AM
> > > Subject: Re: Heartbeat
> > >
> > > Hi Ken,
> > >
> > > I have just opened a new JIRA as you suggested.
> > > https://issues.apache.org/jira/browse/PROTON-512
> > >
> > > My idea is that there would be one global idle timeout for whole
> > messenger
> > > and this timeout will be used for all connections.
> > >
> > > Regards,
> > > Tomas
> > >
> > >
> > > 2013-11-11 Ken Giusti <kg...@redhat.com>:
> > >
> > > > Hi Tomas,
> > > >
> > > > ----- Original Message -----
> > > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > > To: proton@qpid.apache.org
> > > > > Sent: Monday, November 11, 2013 6:43:06 AM
> > > > > Subject: Re: Heartbeat
> > > > >
> > > > > Hi Ken,
> > > > >
> > > > > thanks for your answer.
> > > > > And is there a way how to set it for pn_messenger_t
> > > > >
> > > >
> > > > No - not at present.  Messenger doesn't give the application access to
> > the
> > > > transport since Messenger hides all that connection-related stuff.
> > > >
> > > > I'd recommend opening a JIRA requesting this feature so we won't forget
> > > > about it.
> > > >
> > > > https://issues.apache.org/jira/browse/PROTON
> > > >
> > > > thanks.
> > > >
> > > > -K
> > > >
> > > >
> > > >
> > > > > Thanks,
> > > > > Tomas
> > > > >
> > > > >
> > > > > 2013/11/8 Ken Giusti <kg...@redhat.com>
> > > > >
> > > > > > Hi Tomas,
> > > > > >
> > > > > > The C implementation of proton allows you to set an idle time out
> > for a
> > > > > > connection as described in the AMQP 1.0 spec.  This value is used
> > to
> > > > > > generate "null" frames (ie. frames with no bodies) on idle
> > connections
> > > > as
> > > > > > to not expire the timeout.   The connection will be dropped if the
> > > > local
> > > > > > side does not receive any traffic for it's configured timeout
> > period.
> > > > > >
> > > > > > The C api is available on the transport object - see engine.h:
> > > > > >
> > > > > > /* timeout of zero means "no timeout" */
> > > > > > PN_EXTERN pn_millis_t pn_transport_get_idle_timeout(pn_transport_t
> > > > > > *transport);
> > > > > > PN_EXTERN void pn_transport_set_idle_timeout(pn_transport_t
> > *transport,
> > > > > > pn_millis_t timeout);
> > > > > > PN_EXTERN pn_millis_t
> > > > pn_transport_get_remote_idle_timeout(pn_transport_t
> > > > > > *transport);
> > > > > >
> > > > > >
> > > > > > I don't think the java implementation has this - yet.  See
> > > > > > https://issues.apache.org/jira/browse/PROTON-111
> > > > > >
> > > > > >
> > > > > >
> > > > > > ----- Original Message -----
> > > > > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > > > > To: proton@qpid.apache.org
> > > > > > > Sent: Friday, November 8, 2013 6:06:15 AM
> > > > > > > Subject: Heartbeat
> > > > > > >
> > > > > > > Hi,
> > > > > > >
> > > > > > > I am looking for a way how to manually specify heartbeat for a
> > > > > > connection.
> > > > > > >
> > > > > > > Is there a way how to do it using proton?
> > > > > > >
> > > > > > > Thanks,
> > > > > > >
> > > > > > > Tomas Soltys
> > > > > > >
> > > > > >
> > > > > > --
> > > > > > -K
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Tomáš Šoltys
> > > > > tomas.soltys@gmail.com
> > > > > http://www.range-software.com
> > > > > (+420) 776-843-663
> > > > >
> > > >
> > > > --
> > > > -K
> > > >
> > >
> > >
> > >
> > > --
> > > Tomáš Šoltys
> > > tomas.soltys@gmail.com
> > > http://www.range-software.com
> > > (+420) 776-843-663
> > >
> >
> > --
> > -K
> >
> 

-- 
-K

Re: Heartbeat

Posted by Rafael Schloming <rh...@alum.mit.edu>.
There is already a single per messenger timeout for blocking operations. It
should be a pretty trivial patch to set that same value as the idle timeout
on each connection.

--Rafael


On Fri, Feb 7, 2014 at 8:26 AM, Ken Giusti <kg...@redhat.com> wrote:

> Thanks Tomáš - I agree with a single per-messenger timeout approach.
>
>
> ----- Original Message -----
> > From: "Tomáš Šoltys" <to...@gmail.com>
> > To: proton@qpid.apache.org
> > Sent: Friday, February 7, 2014 3:04:34 AM
> > Subject: Re: Heartbeat
> >
> > Hi Ken,
> >
> > I have just opened a new JIRA as you suggested.
> > https://issues.apache.org/jira/browse/PROTON-512
> >
> > My idea is that there would be one global idle timeout for whole
> messenger
> > and this timeout will be used for all connections.
> >
> > Regards,
> > Tomas
> >
> >
> > 2013-11-11 Ken Giusti <kg...@redhat.com>:
> >
> > > Hi Tomas,
> > >
> > > ----- Original Message -----
> > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > To: proton@qpid.apache.org
> > > > Sent: Monday, November 11, 2013 6:43:06 AM
> > > > Subject: Re: Heartbeat
> > > >
> > > > Hi Ken,
> > > >
> > > > thanks for your answer.
> > > > And is there a way how to set it for pn_messenger_t
> > > >
> > >
> > > No - not at present.  Messenger doesn't give the application access to
> the
> > > transport since Messenger hides all that connection-related stuff.
> > >
> > > I'd recommend opening a JIRA requesting this feature so we won't forget
> > > about it.
> > >
> > > https://issues.apache.org/jira/browse/PROTON
> > >
> > > thanks.
> > >
> > > -K
> > >
> > >
> > >
> > > > Thanks,
> > > > Tomas
> > > >
> > > >
> > > > 2013/11/8 Ken Giusti <kg...@redhat.com>
> > > >
> > > > > Hi Tomas,
> > > > >
> > > > > The C implementation of proton allows you to set an idle time out
> for a
> > > > > connection as described in the AMQP 1.0 spec.  This value is used
> to
> > > > > generate "null" frames (ie. frames with no bodies) on idle
> connections
> > > as
> > > > > to not expire the timeout.   The connection will be dropped if the
> > > local
> > > > > side does not receive any traffic for it's configured timeout
> period.
> > > > >
> > > > > The C api is available on the transport object - see engine.h:
> > > > >
> > > > > /* timeout of zero means "no timeout" */
> > > > > PN_EXTERN pn_millis_t pn_transport_get_idle_timeout(pn_transport_t
> > > > > *transport);
> > > > > PN_EXTERN void pn_transport_set_idle_timeout(pn_transport_t
> *transport,
> > > > > pn_millis_t timeout);
> > > > > PN_EXTERN pn_millis_t
> > > pn_transport_get_remote_idle_timeout(pn_transport_t
> > > > > *transport);
> > > > >
> > > > >
> > > > > I don't think the java implementation has this - yet.  See
> > > > > https://issues.apache.org/jira/browse/PROTON-111
> > > > >
> > > > >
> > > > >
> > > > > ----- Original Message -----
> > > > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > > > To: proton@qpid.apache.org
> > > > > > Sent: Friday, November 8, 2013 6:06:15 AM
> > > > > > Subject: Heartbeat
> > > > > >
> > > > > > Hi,
> > > > > >
> > > > > > I am looking for a way how to manually specify heartbeat for a
> > > > > connection.
> > > > > >
> > > > > > Is there a way how to do it using proton?
> > > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > Tomas Soltys
> > > > > >
> > > > >
> > > > > --
> > > > > -K
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Tomáš Šoltys
> > > > tomas.soltys@gmail.com
> > > > http://www.range-software.com
> > > > (+420) 776-843-663
> > > >
> > >
> > > --
> > > -K
> > >
> >
> >
> >
> > --
> > Tomáš Šoltys
> > tomas.soltys@gmail.com
> > http://www.range-software.com
> > (+420) 776-843-663
> >
>
> --
> -K
>

Re: Heartbeat

Posted by Ken Giusti <kg...@redhat.com>.
Thanks Tomáš - I agree with a single per-messenger timeout approach.


----- Original Message -----
> From: "Tomáš Šoltys" <to...@gmail.com>
> To: proton@qpid.apache.org
> Sent: Friday, February 7, 2014 3:04:34 AM
> Subject: Re: Heartbeat
> 
> Hi Ken,
> 
> I have just opened a new JIRA as you suggested.
> https://issues.apache.org/jira/browse/PROTON-512
> 
> My idea is that there would be one global idle timeout for whole messenger
> and this timeout will be used for all connections.
> 
> Regards,
> Tomas
> 
> 
> 2013-11-11 Ken Giusti <kg...@redhat.com>:
> 
> > Hi Tomas,
> >
> > ----- Original Message -----
> > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > To: proton@qpid.apache.org
> > > Sent: Monday, November 11, 2013 6:43:06 AM
> > > Subject: Re: Heartbeat
> > >
> > > Hi Ken,
> > >
> > > thanks for your answer.
> > > And is there a way how to set it for pn_messenger_t
> > >
> >
> > No - not at present.  Messenger doesn't give the application access to the
> > transport since Messenger hides all that connection-related stuff.
> >
> > I'd recommend opening a JIRA requesting this feature so we won't forget
> > about it.
> >
> > https://issues.apache.org/jira/browse/PROTON
> >
> > thanks.
> >
> > -K
> >
> >
> >
> > > Thanks,
> > > Tomas
> > >
> > >
> > > 2013/11/8 Ken Giusti <kg...@redhat.com>
> > >
> > > > Hi Tomas,
> > > >
> > > > The C implementation of proton allows you to set an idle time out for a
> > > > connection as described in the AMQP 1.0 spec.  This value is used to
> > > > generate "null" frames (ie. frames with no bodies) on idle connections
> > as
> > > > to not expire the timeout.   The connection will be dropped if the
> > local
> > > > side does not receive any traffic for it's configured timeout period.
> > > >
> > > > The C api is available on the transport object - see engine.h:
> > > >
> > > > /* timeout of zero means "no timeout" */
> > > > PN_EXTERN pn_millis_t pn_transport_get_idle_timeout(pn_transport_t
> > > > *transport);
> > > > PN_EXTERN void pn_transport_set_idle_timeout(pn_transport_t *transport,
> > > > pn_millis_t timeout);
> > > > PN_EXTERN pn_millis_t
> > pn_transport_get_remote_idle_timeout(pn_transport_t
> > > > *transport);
> > > >
> > > >
> > > > I don't think the java implementation has this - yet.  See
> > > > https://issues.apache.org/jira/browse/PROTON-111
> > > >
> > > >
> > > >
> > > > ----- Original Message -----
> > > > > From: "Tomáš Šoltys" <to...@gmail.com>
> > > > > To: proton@qpid.apache.org
> > > > > Sent: Friday, November 8, 2013 6:06:15 AM
> > > > > Subject: Heartbeat
> > > > >
> > > > > Hi,
> > > > >
> > > > > I am looking for a way how to manually specify heartbeat for a
> > > > connection.
> > > > >
> > > > > Is there a way how to do it using proton?
> > > > >
> > > > > Thanks,
> > > > >
> > > > > Tomas Soltys
> > > > >
> > > >
> > > > --
> > > > -K
> > > >
> > >
> > >
> > >
> > > --
> > > Tomáš Šoltys
> > > tomas.soltys@gmail.com
> > > http://www.range-software.com
> > > (+420) 776-843-663
> > >
> >
> > --
> > -K
> >
> 
> 
> 
> --
> Tomáš Šoltys
> tomas.soltys@gmail.com
> http://www.range-software.com
> (+420) 776-843-663
> 

-- 
-K