helix-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Erwin Karbasi <er...@optinity.com>
Subject Re: Messaging
Date Tue, 21 May 2013 21:14:50 GMT
Actually, We'd like to have one node to be active that execute some jobs
and if that node go down other node wold need to take the active (leader)
place and continue the job.
We also would like to connect to the currently leader node and provide to
it tasks that should be done.

As I figured out from that state machines that include in Helix the best
applicable state machine is leader standby.

Your insight would highly appreciated.

Thanks,
Erwin

On Tue, May 21, 2013 at 11:49 PM, Alexandre Porcelli <porcelli@redhat.com>wrote:

> Can you confirm that, if you have just one node active in the cluster
> (first node to startup)... you have 0 messages sent and you don't have
> callbacks executed at all.
>
> Regards,
> ---
> Alexandre Porcelli
> porcelli@redhat.com
>
> On May 21, 2013, at 3:30 PM, Erwin Karbasi <erwin@optinity.com> wrote:
>
> > Thank you a lot for fast turnaround clear and informatics response.
> >
> > Cheers,
> > Erwin
> >
> > Erwin Karbasi
> > AT&T, Senior Software Architect
> >
> >
> > On Tue, May 21, 2013 at 8:05 PM, kishore g <g.kishore@gmail.com> wrote:
> > It uses zookeeper for messaging. The goal of this feature is to
> guarantee successful deliver of message. For example, you can send a
> message even if the node is down and it can process it when it comes back
> up. Other feature is to ensure that the message is processed by the
> recipient only if the send criteria matches when is message is received.
> For example, you can say things like
> >
> >  * Only the current participant should process it, which means if the
> participant dies and comes back up it can ignore that message.
> > *  It should be slave for a resource, if its state has changed after the
> sender sends message it can ignore the message.
> >
> > This feature should not be used to achieve low latency or high
> throughput messaging between components. We do have plans to provide a java
> nio based mechanism to communicate between various components but havent
> scoped it
> >
> > thanks,
> > Kishore G
> >
> >
> > On Tue, May 21, 2013 at 9:29 AM, Erwin Karbasi <erwin@optinity.com>
> wrote:
> > Hello All,
> >
> > Which protocol/mechanism/framework Helix uses for sending messaging
> among nodes as described here:
> > http://helix.incubator.apache.org/tutorial_messaging.html
> >
> > Thanks,
> > Erwin Karbasi
> > AT&T, Senior Software Architect
> >
> >
>
>

Mime
View raw message