Thief node does not belong to task topology [was "Failed to send message"]

[Paolo Di Tommaso]

Hi Val,

I surely can contribute on that issue but I would need a fix as soon as
possible so I will do both. Thus, I will try the dirty path on my own
implementation and to implement a fix for IGNITE-1267
<https://issues.apache.org/jira/browse/IGNITE-1267>.

However also in the above issue the last comment suggests to remove the
check on the task topology and I have any clue of any better alternative.
Could you clarify how do you think it should be fixed?


Cheers,
Paolo



On Tue, Mar 1, 2016 at 11:56 PM, vkulichenko <[hidden email]>
wrote:

> Paolo,
>
> I found the ticket about this issue [1]. How about picking it up and fixing
> instead of implementing your own version of the SPI?
>
> Removing the check completely is wrong, because it's possible that a node
> doesn't belong to the cluster group on which the task was executed. But we
> should check the original predicate instead of collection of nodes sealed
> during the map phase.
>
> [1] https://issues.apache.org/jira/browse/IGNITE-1267
>
> -Val
>
>
>
> --
> View this message in context:
> http://apache-ignite-users.70518.x6.nabble.com/Failed-to-send-message-tp3217p3317.html
> Sent from the Apache Ignite Users mailing list archive at Nabble.com.
>

[Valentin Kulichenko]

Hi Paolo,

See my comment in the ticket.

-Val

On Wed, Mar 2, 2016 at 2:04 AM, Paolo Di Tommaso <[hidden email]>
wrote:

> Hi Val,
>
> I surely can contribute on that issue but I would need a fix as soon as
> possible so I will do both. Thus, I will try the dirty path on my own
> implementation and to implement a fix for IGNITE-1267
> <https://issues.apache.org/jira/browse/IGNITE-1267>.
>
> However also in the above issue the last comment suggests to remove the
> check on the task topology and I have any clue of any better alternative.
> Could you clarify how do you think it should be fixed?
>
>
> Cheers,
> Paolo
>
>
>
> On Tue, Mar 1, 2016 at 11:56 PM, vkulichenko <
> [hidden email]>
> wrote:
>
> > Paolo,
> >
> > I found the ticket about this issue [1]. How about picking it up and
> fixing
> > instead of implementing your own version of the SPI?
> >
> > Removing the check completely is wrong, because it's possible that a node
> > doesn't belong to the cluster group on which the task was executed. But
> we
> > should check the original predicate instead of collection of nodes sealed
> > during the map phase.
> >
> > [1] https://issues.apache.org/jira/browse/IGNITE-1267
> >
> > -Val
> >
> >
> >
> > --
> > View this message in context:
> >
> http://apache-ignite-users.70518.x6.nabble.com/Failed-to-send-message-tp3217p3317.html
> > Sent from the Apache Ignite Users mailing list archive at Nabble.com.
> >
>