Skip to content
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

[clickhouse] can't show progress for some query #1216

Closed
BohuTANG opened this issue Jul 28, 2021 · 10 comments · Fixed by #3547
Closed

[clickhouse] can't show progress for some query #1216

BohuTANG opened this issue Jul 28, 2021 · 10 comments · Fixed by #3547
Labels
C-bug Category: something isn't working C-clickhouse Category: Clickhouse good first issue Category: good first issue

Comments

@BohuTANG
Copy link
Member

BohuTANG commented Jul 28, 2021

Summary

If we run this query in ClickHouse client:

SELECT number FROM numbers_mt(10000000000) ORDER BY number DESC LIMIT 100;

It can not show the progress, but the group by query is ok:

datafuse :) SELECT max(number), sum(number) FROM numbers_mt(1000000000) GROUP BY number % 3, number % 4, number % 5;

It seems that only the group by query working.

@BohuTANG BohuTANG added C-bug Category: something isn't working C-clickhouse Category: Clickhouse labels Jul 28, 2021
@compasses
Copy link

compasses commented Jul 29, 2021

BTW, just kindly remind, I did some performance test against clickhouse follow the guide from https://datafuse.rs/overview/performance/, all use the newest master branch, but i have a very different result with that, some cases clickhouse is better :)

@BohuTANG
Copy link
Member Author

BTW, just kindly remind, I did some performance test against clickhouse follow the guide from https://datafuse.rs/overview/performance/, all use the newest master branch, but i have a very different result with that, mostly clickhouse is better :)

Hi!
Could you give the env and results of your test?

@compasses
Copy link

Hi,
Just some case, before i make some mis config, you can just check the two case appeared in this issue.

Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz 48 core:

image
image

image
image

@BohuTANG
Copy link
Member Author

BohuTANG commented Jul 29, 2021

Hmm, how do you build the datafuse?
The performance binary build with make build and test on Hardware: AMD Ryzen 7 PRO 4750U, 8 CPU Cores, 16 Threads

@compasses
Copy link

Hmm, how do you build the datafuse?

Just run with: cargo run --bin fuse-query --release

instead of make build, seems no difference.

@BohuTANG
Copy link
Member Author

BohuTANG commented Jul 29, 2021

You could try

RUSTFLAGS="-C target-cpu=native" cargo build --bin=fuse-query --features allocator --release

and test again.

@BohuTANG
Copy link
Member Author

BohuTANG commented Jul 29, 2021

On my local, I have set set max_threads=16 for clickhouse, but it seems slower than the default(max_threads=8).
I will re-measure the performance against the latest version ClickHouse, there is already a task here: #1129

Thanks for your report @compasses

@compasses
Copy link

Yes, that's meet my test to

RUSTFLAGS="-C target-cpu=native" cargo build --bin=fuse-query --features allocator --release

That's not work on my environment:

.cargo/registry/src/github.com-1ecc6299db9ec823/snmalloc-sys-0.2.27/snmalloc/src/override/../mem/../mem/../pal/pal_posix.h:294:23: error: ‘getentropy’ was not declared in this scope; did you mean ‘get_entropy64’?
    294 |         if (getentropy(&result, sizeof(result)) != 0)
        |             ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
        |             get_entropy64
  make[3]: *** [CMakeFiles/snmallocshim-1mib-rust.dir/src/override/rust.cc.o] Error 1
  make[2]: *** [CMakeFiles/snmallocshim-1mib-rust.dir/all] Error 2
  make[1]: *** [CMakeFiles/snmallocshim-1mib-rust.dir/rule] Error 2
  make: *** [snmallocshim-1mib-rust] Error 2
  thread 'main' panicked at '
  command did not execute successfully, got: exit status: 2

BTW, where I can find the newest progress of this project? Currently I cannot create a local table and do some insert & query test right ?

@BohuTANG
Copy link
Member Author

BohuTANG commented Jul 31, 2021

@compasses

  1. snmalloc has few problems here, what's your os version?
  2. We have roadmap to track the progress Roadmap 2021 #746
  3. local table insert is working, there's an example for the memory engine https://datafuse.rs/sqlstatement/data-manipulation-language-dml/dml-insert/
    and, the remote table is working too but has not documents yet, I think it will done soon

@compasses
Copy link

cool!big thanks, the os version: SMP Debian 4.14.81.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
C-bug Category: something isn't working C-clickhouse Category: Clickhouse good first issue Category: good first issue
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants