[
https://issues.apache.org/jira/browse/PHOENIX-7213?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tanuj Khurana updated PHOENIX-7213:
-----------------------------------
Description:
As discussed in PHOENIX-7117 _*phoenix.query.queueSize*_ limits can cause scan
queries on table with lot of regions to fail. I did some profiling and found
that each task allocates 48 bytes. So I am adding an enhancement to make the
queue size unlimited if we set the limit to -1. We still have a limit on the
number of tasks that can be executed in parallel controlled by the
_*phoenix.query.threadPoolSize*_
If you don't enforce the limit, the only thing that is limiting is the jvm heap
allocated.
was:
As discussed in
[PHOENIX-7117|https://issues.apache.org/jira/browse/PHOENIX-7117]
_*phoenix.query.queueSize*_ limits can cause scan queries on table with lot of
regions to fail. I did some profiling and found that each task allocates 48
bytes. So I am adding an enhancement to not enforce this limit (or an infinite
queue ) if we set the limit to -1. We still have a limit on the number of tasks
that can be executed in parallel controlled by the
_*phoenix.query.threadPoolSize*_
If you don't enforce the limit, the only thing that is limiting is the jvm heap
allocated.
Summary: Add option for unlimited phoenix.query.QueueSize (was: Don't
enforce phoenix.query.queueSize limits)
> Add option for unlimited phoenix.query.QueueSize
> ------------------------------------------------
>
> Key: PHOENIX-7213
> URL: https://issues.apache.org/jira/browse/PHOENIX-7213
> Project: Phoenix
> Issue Type: Improvement
> Reporter: Tanuj Khurana
> Assignee: Tanuj Khurana
> Priority: Major
>
> As discussed in PHOENIX-7117 _*phoenix.query.queueSize*_ limits can cause
> scan queries on table with lot of regions to fail. I did some profiling and
> found that each task allocates 48 bytes. So I am adding an enhancement to
> make the queue size unlimited if we set the limit to -1. We still have a
> limit on the number of tasks that can be executed in parallel controlled by
> the
> _*phoenix.query.threadPoolSize*_
>
> If you don't enforce the limit, the only thing that is limiting is the jvm
> heap allocated.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
