[
https://issues.apache.org/jira/browse/TAJO-537?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13877237#comment-13877237
]
Hyunsik Choi commented on TAJO-537:
-----------------------------------
+1
The patch looks pretty good for me. I'll commit.
There is one suggestion that can be dealt in a separate jira issue. It would be
great to move the factory construction outside from the NettyClientBase. Then,
factory release becomes possible when a JVM shutdowns.
> After TAJO-522, still OutOfMemoryError: unable to create new native thread
> ---------------------------------------------------------------------------
>
> Key: TAJO-537
> URL: https://issues.apache.org/jira/browse/TAJO-537
> Project: Tajo
> Issue Type: Bug
> Components: distributed query plan
> Affects Versions: 0.8-incubating
> Reporter: Min Zhou
> Assignee: Min Zhou
> Priority: Critical
> Fix For: 0.8-incubating
>
> Attachments: TAJO-537-v2.patch, TAJO-537.patch
>
>
> I ran a big query applied this patch, still OOM. This is because if the
> quantity of different connections is large, Hyoungjun's patch won't help.
> Different connection create different thread pool.
> Actually, we can share the thread pool among all the clients.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
