[jira] [Commented] (IGNITE-4506) Implement thread-per-partition processing for atomic cache updates

2016-12-28 Thread Vladimir Ozerov (JIRA)

[ 
https://issues.apache.org/jira/browse/IGNITE-4506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15783021#comment-15783021
 ] 

Vladimir Ozerov commented on IGNITE-4506:
-

May be we should add {{sendAsync}} methods to {{GridIoManager}} which will 
ensure that message processing happens asynchronously and in correct thread 
pool even for local messages. And your cache logic may simply switch to this 
{{sendAsync}} method. We will certainly need this in other places, so probably 
we should add some common mechanics right now.

> Implement thread-per-partition processing for atomic cache updates
> --
>
> Key: IGNITE-4506
> URL: https://issues.apache.org/jira/browse/IGNITE-4506
> Project: Ignite
>  Issue Type: Task
>  Components: cache
>Reporter: Semen Boikov
>Assignee: Konstantin Dudkov
> Fix For: 2,0
>
>
> Need implement thread-per-partition processing for atomic cache updates.
> For single key requests remote request should be already processed by striped 
> executor, need fix case when update is mapped on local node - in this case it 
> is processed from calling thread.
> For batched updates need implement per-partition mapping and processing.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (IGNITE-4506) Implement thread-per-partition processing for atomic cache updates

2016-12-28 Thread Vladimir Ozerov (JIRA)

[ 
https://issues.apache.org/jira/browse/IGNITE-4506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15783018#comment-15783018
 ] 

Vladimir Ozerov commented on IGNITE-4506:
-

Sam, please see this as well: https://issues.apache.org/jira/browse/IGNITE-4476

> Implement thread-per-partition processing for atomic cache updates
> --
>
> Key: IGNITE-4506
> URL: https://issues.apache.org/jira/browse/IGNITE-4506
> Project: Ignite
>  Issue Type: Task
>  Components: cache
>Reporter: Semen Boikov
>Assignee: Konstantin Dudkov
> Fix For: 2,0
>
>
> Need implement thread-per-partition processing for atomic cache updates.
> For single key requests remote request should be already processed by striped 
> executor, need fix case when update is mapped on local node - in this case it 
> is processed from calling thread.
> For batched updates need implement per-partition mapping and processing.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)