[
https://issues.apache.org/jira/browse/IGNITE-8406?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16488740#comment-16488740
]
Ivan Fedotov edited comment on IGNITE-8406 at 5/24/18 10:03 AM:
----------------------------------------------------------------
[~dpavlov], Hello!
Could you please look at this ticket?
It seems to me that javaDoc of IgniteDataStreamer not entirely correct.
Mapping keys to node happens in load0 method[1] which invoked in addData method
while loading data from buffers always happens in flush[2] and tryFlush[3]
methods. Just about same situation in close(boolean) method. So I think minor
javaDoc changes are appropriate here.
Also according to conversation on user list[4] it would be better to explicitly
indicate about listeners.
[1][https://github.com/apache/ignite/blob/master/modules/core/src/main/java/org/apache/ignite/internal/processors/datastreamer/DataStreamerImpl.java#L872]
[2][https://github.com/apache/ignite/blob/master/modules/core/src/main/java/org/apache/ignite/internal/processors/datastreamer/DataStreamerImpl.java#L1117]
[3][https://github.com/apache/ignite/blob/master/modules/core/src/main/java/org/apache/ignite/internal/processors/datastreamer/DataStreamerImpl.java#L1227]
[4][http://apache-ignite-users.70518.x6.nabble.com/IgniteDataStreamer-flush-returns-before-all-futures-are-completed-td21330.html]
was (Author: ivanan.fed):
[~dpavlov], Hello!
Could you please see at this ticket?
It seems to me that javaDoc of IgniteDataStreamer not entirely correct.
Mapping keys to node happens in load0 method[1] which invoked in addData method
while loading data from buffers always happens in flush[2] and tryFlush[3]
methods. Just about same situation in close(boolean) method. So I think minor
javaDoc changes are appropriate here.
Also according to conversation on user list[4] it would be better to explicitly
indicate about listeners.
[1][https://github.com/apache/ignite/blob/master/modules/core/src/main/java/org/apache/ignite/internal/processors/datastreamer/DataStreamerImpl.java#L872]
[2][https://github.com/apache/ignite/blob/master/modules/core/src/main/java/org/apache/ignite/internal/processors/datastreamer/DataStreamerImpl.java#L1117]
[3][https://github.com/apache/ignite/blob/master/modules/core/src/main/java/org/apache/ignite/internal/processors/datastreamer/DataStreamerImpl.java#L1227]
[4][http://apache-ignite-users.70518.x6.nabble.com/IgniteDataStreamer-flush-returns-before-all-futures-are-completed-td21330.html]
> Update IgniteDataStreamer.flush() javadoc
> -----------------------------------------
>
> Key: IGNITE-8406
> URL: https://issues.apache.org/jira/browse/IGNITE-8406
> Project: Ignite
> Issue Type: Task
> Components: streaming
> Affects Versions: 2.4
> Reporter: Andrey Kuznetsov
> Assignee: Ivan Fedotov
> Priority: Minor
> Fix For: 2.6
>
>
> Current {{flush()}} implementation can throw {{CacheException}} if one or
> more futures previously returned by {{addData()}} have been completed
> exceptionally. This behavior should be described in {{IgniteDataStreamer}}
> javadoc. Also it's worth noting explicitly that all futures completion upon
> return from {{flush}} does not imply all those future listeners have been
> completed by this moment.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
