You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ignite.apache.org by drosso <dr...@inwind.it> on 2019/01/25 11:25:25 UTC
Data Streamers and Topology Change
Hi,
I'm experiencing a behavior that I don't know if it's correct or not.
I start with a Client Node and 2 Server Nodes: the Client node inserts 200
records into a cache using a Data Streamer: the logs correctly show that the
records are processed (almost) evenly by the 2 server nodes (I've added a
log with the NodeId, just to be sure that the streamer works as expected).
The cache is defined as PARTITIONED with backup copies = 1 and native
persistence is enabled.
All Nodes are deployed on the same physical node.
Now, if I add another Server Node and use the same Client node to inject
another bunch of 200 records, I would expect the records to be processed by
3 Server Nodes: what I see, instead, is that the records continue to be
processed by the 2 original Server Nodes.
Is this the expected behavior ? If it is, this means that Streamers are
created one and for all when the Client Node connects to the Grid for the
first time ?
Thank you for any feedback.
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Re: Data Streamers and Topology Change
Posted by drosso <dr...@inwind.it>.
No, I didn't... And that was the problem!
Thank you so much for your hint, now it works!
Bye
Davide
--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Re: Data Streamers and Topology Change
Posted by Sergey Antonov <an...@gmail.com>.
Hi, did add you third server node to baseline topology?
https://apacheignite.readme.io/docs/baseline-topology
пт, 25 янв. 2019 г. в 14:25, drosso <dr...@inwind.it>:
> Hi,
> I'm experiencing a behavior that I don't know if it's correct or not.
> I start with a Client Node and 2 Server Nodes: the Client node inserts 200
> records into a cache using a Data Streamer: the logs correctly show that
> the
> records are processed (almost) evenly by the 2 server nodes (I've added a
> log with the NodeId, just to be sure that the streamer works as expected).
> The cache is defined as PARTITIONED with backup copies = 1 and native
> persistence is enabled.
> All Nodes are deployed on the same physical node.
>
> Now, if I add another Server Node and use the same Client node to inject
> another bunch of 200 records, I would expect the records to be processed by
> 3 Server Nodes: what I see, instead, is that the records continue to be
> processed by the 2 original Server Nodes.
>
> Is this the expected behavior ? If it is, this means that Streamers are
> created one and for all when the Client Node connects to the Grid for the
> first time ?
> Thank you for any feedback.
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>
--
BR, Sergey Antonov