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
parallel_view_processing behavior has changed. #10241
Comments
The performance impact is visible with a high number of MVs.
18.14.18 Elapsed: 0.005 sec. Processed 100.00 thousand rows
19.13.7.57 Elapsed: 0.011 sec. Processed 100.00 thousand rows,
All inserts are executed by the same thread [ 68 ] |
The change happened in September 2019 |
Broken in #3796 if I'm not mistaken. |
This is not fixed. |
yes, no parallelism during MV inserts (one thread [ 202101 ] for all MV) 20.5.1.3496 insert into testX select number from numbers(100000) settings max_block_size=1000000; parallel_view_processing=0 Elapsed: 0.010 sec. parallel_view_processing=1 Elapsed: 0.010 sec. / EXPECTED 0.005 sec.
|
parallel_view_processing=1
18.14.18
10
10
0
10
19.13.7.57
10
0
0
0
18.14.18
Threads 209, 211 did insert before Exception
19.13.7.57
Threads 120, 81 never inserted
We suspect it impacts performance because INSERTS start after all SELECTS (for 1 block ?) finished.
The text was updated successfully, but these errors were encountered: