-
Notifications
You must be signed in to change notification settings - Fork 3.7k
[opt](scanner) Control the degree of parallelism of scanner when only limit involved #39927
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[opt](scanner) Control the degree of parallelism of scanner when only limit involved #39927
Conversation
|
Thank you for your contribution to Apache Doris. Since 2024-03-18, the Document has been moved to doris-website. |
|
clang-tidy review says "All clean, LGTM! 👍" |
| description = { | ||
| "请使用 adaptive_pipeline_task_serial_read_on_limit", | ||
| "Please use adaptive_pipeline_task_serial_read_on_limit"}) | ||
| public long limitRowsForSingleInstance = 8192; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
保持原来的值,10000就行了,万一有一些极限情况,上线后,直接性能下降了
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
FE BE上参数不一样,都改成10000?
6c4ce1b to
8381b41
Compare
|
run buildall |
|
clang-tidy review says "All clean, LGTM! 👍" |
|
clang-tidy review says "All clean, LGTM! 👍" |
TPC-H: Total hot run time: 38150 ms |
TPC-DS: Total hot run time: 193246 ms |
ClickBench: Total hot run time: 31.16 s |
|
run buildall |
|
clang-tidy review says "All clean, LGTM! 👍" |
TPC-H: Total hot run time: 38335 ms |
TPC-DS: Total hot run time: 193613 ms |
ClickBench: Total hot run time: 31.93 s |
|
run buildall |
TPC-H: Total hot run time: 38507 ms |
TPC-DS: Total hot run time: 193612 ms |
ClickBench: Total hot run time: 32.91 s |
ClickBench: Total hot run time: 32.77 s |
… limit involved (apache#39927) For queries like `select * from tbl limit 100;`, we should limit the parallelism to 1. We already have strategies in our code, but we can not control its behavior. So add session variables to control them. Previous pr like apache#33888 apache#36535 already has done something similar, this pr integrates them.
enable_scan_node_run_serial is not working after #39927
enable_scan_node_run_serial is not working after apache#39927
enable_scan_node_run_serial is not working after apache#39927
enable_scan_node_run_serial is not working after apache#39927
# Versions - [x] dev - [x] 3.0 - [x] 2.1 - [ ] 2.0 # Languages - [x] Chinese - [x] English apache/doris#39927
enable_scan_node_run_serial is not working after apache#39927
… limit involved (apache#39927) For queries like `select * from tbl limit 100;`, we should limit the parallelism to 1. We already have strategies in our code, but we can not control its behavior. So add session variables to control them. Previous pr like apache#33888 apache#36535 already has done something similar, this pr integrates them.
… limit involved (apache#39927) For queries like `select * from tbl limit 100;`, we should limit the parallelism to 1. We already have strategies in our code, but we can not control its behavior. So add session variables to control them. Previous pr like apache#33888 apache#36535 already has done something similar, this pr integrates them.
… limit involved (apache#39927) For queries like `select * from tbl limit 100;`, we should limit the parallelism to 1. We already have strategies in our code, but we can not control its behavior. So add session variables to control them. Previous pr like apache#33888 apache#36535 already has done something similar, this pr integrates them.
enable_scan_node_run_serial is not working after apache#39927
Related PR: apache#39927 apache#39927 introduced a new session var `adaptive_pipeline_task_serial_read_on_limit`, its implementation has flaw, reduce scanner parallelism should only happen when there is not conjunction. Since we can not estimate the filtration efficiency.
Related PR: apache#39927 apache#39927 introduced a new session var `adaptive_pipeline_task_serial_read_on_limit`, its implementation has flaw, reduce scanner parallelism should only happen when there is not conjunction. Since we can not estimate the filtration efficiency.
enable_scan_node_run_serial is not working after apache#39927
For queries like
select * from tbl limit 100;, we should limit the parallelism to 1.We already have strategies in our code, but we can not control its behavior. So add session variables to control them.
Previous pr like #33888 #36535 already has done something similar, this pr integrates them.