refactor(jdbc-sink): execute statements in batch and set isolation level to RC #12250

StrikeW · 2023-09-13T03:12:33Z

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

Explicitly set transaction isolation level to READ COMMITTED (same as before).
Fix #12061, group SQL statements to a stagingStatements map, and execute them after processing a chunk. And we execute DELETE first to prevent accidentally deletions.

Performance numbers (sink a tpch orders table to postgres):
6k rows/s for each parallelism (4 parallelisms in total)

Checklist

I have written necessary rustdoc comments
I have added necessary unit tests and integration tests
I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features Sqlsmith: Sql feature generation #7934).
My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future).
All checks passed in ./risedev check (or alias, ./risedev c)
My PR changes performance-critical code. (Please run macro/micro-benchmarks and show the results.)

My PR contains critical fixes that are necessary to be merged into the latest release. (Please check out the details)

Documentation

My PR needs documentation updates. (Please use the Release note section below to summarize the impact on users)

Release note

If this PR includes changes that directly affect users or other significant modifications relevant to the community, kindly draft a release note to provide a concise summary of these changes. Please prioritize highlighting the impact these changes will have on users.

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java

chenzl25 · 2023-09-13T03:32:47Z

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java

+ // execute staging statement in batch after all rows are prepared,
+ // and we need to delete first to prevent accidentally deletion
+ var deleteStmt = stagingStatements.remove(OpType.DELETE);
+ if (deleteStmt != null) {
+ executeStatement(deleteStmt);
+ }
+
+ var upsertStmt = stagingStatements.remove(OpType.UPSERT);
+ if (upsertStmt != null) {
+ executeStatement(upsertStmt);
+ }
+
+ var insertStmt = stagingStatements.remove(OpType.INSERT);
+ if (insertStmt != null) {
+ executeStatement(insertStmt);
+ }


IIRC we have discussed executing deletes before inserts in sink executor before, but we should do a compaction in the upstream? cc @st1page

I am not quite sure whether the execution order is right here, since our e2e test lacks of delete/update workload.

@chenzl25 yes, The compaction logic has been implemented in #11070. And I will continue developing that in another PR later.

codecov · 2023-09-13T03:34:02Z

Codecov Report

Merging #12250 (e647d5b) into main (ae4b1f8) will decrease coverage by 0.01%.
Report is 4 commits behind head on main.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main   #12250      +/-   ##
==========================================
- Coverage   69.96%   69.95%   -0.01%     
==========================================
  Files        1411     1411              
  Lines      235305   235305              
==========================================
- Hits       164620   164600      -20     
- Misses      70685    70705      +20

Flag	Coverage Δ
rust	`69.95% <ø> (-0.01%)`	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

see 8 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java

StrikeW · 2023-09-13T09:50:30Z

Add the performance of main branch. Test setup: c5.4xlarge, release build, 4 parallelisms, S3 cc @hzxa21

…jdbc-sink

hzxa21

Rest LGTM

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java

wenym1

Rest LGTM

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java

…jdbc-sink

chenzl25

LGTM. So it is about 2x throughput improvement achieved by batching, right?

StrikeW · 2023-09-14T10:26:27Z

LGTM. So it is about 2x throughput improvement achieved by batching, right?

The main branch is about 4k rows/s, improvement is about 1.5x.

…vel to RC (#12250)

commit c82fc9c Author: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Date: Mon Sep 18 08:37:33 2023 +0000 chore(deps): Bump chrono from 0.4.30 to 0.4.31 (#12359) Signed-off-by: dependabot[bot] <[email protected]> Signed-off-by: Runji Wang <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Runji Wang <[email protected]> Co-authored-by: TennyZhuang <[email protected]> commit cbdc1ac Author: Huangjw <[email protected]> Date: Mon Sep 18 16:22:35 2023 +0800 chore(ci): move release jobs to main-cron pipeline (#12339) commit b37a19c Author: Yuhao Su <[email protected]> Date: Mon Sep 18 16:18:01 2023 +0800 feat(dashboard): add memory profiling (#12052) commit 71d8170 Author: TennyZhuang <[email protected]> Date: Mon Sep 18 15:58:26 2023 +0800 refactor(expr): allow defining functions in frontend (#12287) Signed-off-by: TennyZhuang <[email protected]> Co-authored-by: zwang28 <[email protected]> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> commit cedaec9 Author: Dylan <[email protected]> Date: Mon Sep 18 15:54:10 2023 +0800 feat(optimizer): support agg group by simplify rule (#12349) commit 71d9b0b Author: Noel Kwan <[email protected]> Date: Mon Sep 18 15:32:00 2023 +0800 feat(meta): update StreamJob status on finish (#12342) commit 784fe56 Author: zwang28 <[email protected]> Date: Mon Sep 18 14:47:49 2023 +0800 fix(backup): ensure correct delta log order (#12371) commit 711ecd5 Author: congyi wang <[email protected]> Date: Mon Sep 18 14:11:24 2023 +0800 feat(state_table): add iterator sub range under a certain pk prefix (#12251) commit 1877aed Author: xiangjinwu <[email protected]> Date: Mon Sep 18 13:49:15 2023 +0800 refactor(sink): impl SinkFormatter for AppendOnly and Upsert (#12321) commit f304ed2 Author: xxchan <[email protected]> Date: Sun Sep 17 20:20:17 2023 +0800 revert: Revert "chore: add platforms to hakari (#12333)" (#12363) commit a975d93 Author: Bohan Zhang <[email protected]> Date: Sun Sep 17 19:04:24 2023 +0800 fix: handle kafka sink message timeout error (#12350) commit 8ef74ad Author: Runji Wang <[email protected]> Date: Sat Sep 16 12:16:02 2023 +0800 fix(udf): handle visibility of input chunks in UDTF (#12357) Signed-off-by: Runji Wang <[email protected]> commit 31fdc26 Author: Xu <[email protected]> Date: Fri Sep 15 21:01:14 2023 -0400 feat(expr): switch to `fancy-regex` crate & update the original version (#12329) Co-authored-by: xzhseh <[email protected]> commit 0032145 Author: Runji Wang <[email protected]> Date: Fri Sep 15 16:57:25 2023 +0800 refactor(expr): support variadic function in `#[function]` macro (#12178) Signed-off-by: Runji Wang <[email protected]> commit 467ba4b Author: stonepage <[email protected]> Date: Fri Sep 15 16:28:13 2023 +0800 fix: stream backfill executor use correct schema (#12314) Co-authored-by: Noel Kwan <[email protected]> commit c443197 Author: Dylan <[email protected]> Date: Fri Sep 15 16:22:13 2023 +0800 feat(optimizer): support correlated column in order by (#12341) commit 8a36ca3 Author: Noel Kwan <[email protected]> Date: Fri Sep 15 16:11:03 2023 +0800 feat(meta): Add `creating_status` field for stream jobs (#12330) commit bf5b14e Author: zwang28 <[email protected]> Date: Fri Sep 15 16:06:17 2023 +0800 chore: lift decoding message size limit for ddl client (#12340) commit c0060b2 Author: zwang28 <[email protected]> Date: Fri Sep 15 15:32:14 2023 +0800 feat(meta): add hummock config relevant tables to rw_catalog (#12337) commit 59bb645 Author: xxchan <[email protected]> Date: Fri Sep 15 14:54:54 2023 +0800 chore: add platforms to hakari (#12333) Signed-off-by: Runji Wang <[email protected]> Co-authored-by: Runji Wang <[email protected]> commit 7baa27f Author: Bugen Zhao <[email protected]> Date: Fri Sep 15 14:00:14 2023 +0800 chore: split full debug info for release build (#12255) Signed-off-by: Bugen Zhao <[email protected]> commit a99e6f3 Author: Richard Chien <[email protected]> Date: Fri Sep 15 13:58:19 2023 +0800 fix(stream): fix pk indices of GroupTopN executors (#12304) Signed-off-by: Richard Chien <[email protected]> commit 43c010e Author: Croxx <[email protected]> Date: Fri Sep 15 11:59:41 2023 +0800 chore: fix comment and metrics (#12331) Signed-off-by: MrCroxx <[email protected]> commit 214118b Author: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Date: Fri Sep 15 10:03:14 2023 +0800 chore(deps): Bump serde_json from 1.0.106 to 1.0.107 (#12322) Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> commit 41ebb2a Author: Xu <[email protected]> Date: Thu Sep 14 22:02:08 2023 -0400 fix(regexp): substraction overflow when incorrectly speicifying `start` (#12325) commit a566cfe Author: Xu <[email protected]> Date: Thu Sep 14 12:58:35 2023 -0400 feat(expr): add `array_sum` (#12162) Signed-off-by: Runji Wang <[email protected]> Co-authored-by: Runji Wang <[email protected]> commit 28bbf10 Author: Croxx <[email protected]> Date: Fri Sep 15 00:40:27 2023 +0800 fix(ci): exclude tikv-jemalloc-sys in hakari check (#12320) Signed-off-by: MrCroxx <[email protected]> commit 5aa5a47 Author: zwang28 <[email protected]> Date: Thu Sep 14 21:02:01 2023 +0800 feat(meta): add hummock version relevant tables to rw_catalog (#12309) commit a740364 Author: Huangjw <[email protected]> Date: Thu Sep 14 19:11:04 2023 +0800 chore(ci): install locales in prebuilt image (#12311) Signed-off-by: Bugen Zhao <[email protected]> Co-authored-by: Bugen Zhao <[email protected]> commit 0e72056 Author: StrikeW <[email protected]> Date: Thu Sep 14 18:42:34 2023 +0800 refactor(jdbc-sink): execute statements in batch and set isolation level to RC (#12250) commit 827ed5e Author: Dylan <[email protected]> Date: Thu Sep 14 17:31:41 2023 +0800 refactor(connector): migrate cdc source metric from connector to compute (#12283) commit a934185 Author: Dylan <[email protected]> Date: Thu Sep 14 17:31:04 2023 +0800 fix(optimizer): relax scan predicate pull up mapping inverse restriction (#12308) commit db0c099 Author: Dylan <[email protected]> Date: Thu Sep 14 17:30:28 2023 +0800 feat(stream): handling watermark in temporal join (#12302) commit 1ecea63 Author: Bugen Zhao <[email protected]> Date: Thu Sep 14 16:43:14 2023 +0800 refactor(risedev): split the steps for building and running playground (#12279) Signed-off-by: Bugen Zhao <[email protected]> Co-authored-by: xxchan <[email protected]> commit ae4b1f8 Author: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Date: Thu Sep 14 08:41:29 2023 +0000 chore(deps): Bump clap from 4.4.2 to 4.4.3 (#12245) Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Bugen Zhao <[email protected]> commit 7ca370a Author: Croxx <[email protected]> Date: Thu Sep 14 16:24:19 2023 +0800 feat(refill): fetch whole sst file when refilling (#12265) Signed-off-by: MrCroxx <[email protected]> commit ec129b6 Author: Yuhao Su <[email protected]> Date: Thu Sep 14 16:04:37 2023 +0800 chore: use cfg! to instead of #cfg[] for jemalloc control policy (#12307) commit 9814af8 Author: Runji Wang <[email protected]> Date: Thu Sep 14 14:45:14 2023 +0800 feat(expr): add `pg_sleep` function (#12294) Signed-off-by: Runji Wang <[email protected]> commit 4525e67 Author: Noel Kwan <[email protected]> Date: Thu Sep 14 14:38:03 2023 +0800 feat(stream): support source throttling (#12295) commit 5ffd58d Author: Dylan <[email protected]> Date: Thu Sep 14 14:35:03 2023 +0800 refactor(connector): replace validate source rpc with jni (#12270) commit 888f2dd Author: Eric Fu <[email protected]> Date: Thu Sep 14 14:32:59 2023 +0800 fix: panic when dumping memory profile (#12276) Signed-off-by: Little-Wallace <[email protected]>

StrikeW added 3 commits September 11, 2023 17:11

optimize jdbc sink

e5c0c17

disable autocommit

4f42227

fix

f9283d5

StrikeW requested review from hzxa21, chenzl25, fuyufjh and wenym1 September 13, 2023 03:12

github-actions bot added the type/refactor label Sep 13, 2023

chenzl25 reviewed Sep 13, 2023

View reviewed changes

wenym1 reviewed Sep 13, 2023

View reviewed changes

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java Outdated Show resolved Hide resolved

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java Outdated Show resolved Hide resolved

StrikeW added 3 commits September 14, 2023 10:16

refine

dffb651

Merge remote-tracking branch 'origin/main' into siyuan/v1.3-refactor-…

61c8a2d

…jdbc-sink

fix

582acc5

StrikeW requested review from chenzl25 and wenym1 September 14, 2023 03:37

hzxa21 reviewed Sep 14, 2023

View reviewed changes

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java Outdated Show resolved Hide resolved

StrikeW added the ci/run-java-connector-node-integration-tests label Sep 14, 2023

update level to RC

7c7eb52

wenym1 reviewed Sep 14, 2023

View reviewed changes

java/connector-node/risingwave-sink-jdbc/src/main/java/com/risingwave/connector/JDBCSink.java Outdated Show resolved Hide resolved

StrikeW added 2 commits September 14, 2023 17:03

refine

1bed88d

Merge remote-tracking branch 'origin/main' into siyuan/v1.3-refactor-…

352e739

…jdbc-sink

StrikeW changed the title ~~refactor(jdbc-sink): execute statements in batch and modify isolation level~~ refactor(jdbc-sink): execute statements in batch and set isolation level to RC Sep 14, 2023

wenym1 approved these changes Sep 14, 2023

View reviewed changes

chenzl25 approved these changes Sep 14, 2023

View reviewed changes

update jdbc sink test

e647d5b

StrikeW enabled auto-merge September 14, 2023 10:35

StrikeW added this pull request to the merge queue Sep 14, 2023

Merged via the queue into main with commit 0e72056 Sep 14, 2023
28 of 29 checks passed

StrikeW deleted the siyuan/v1.3-refactor-jdbc-sink branch September 14, 2023 11:04

Li0k pushed a commit that referenced this pull request Sep 15, 2023

refactor(jdbc-sink): execute statements in batch and set isolation le…

b8d9ad5

…vel to RC (#12250)

StrikeW mentioned this pull request Oct 9, 2023

Don't use transaction in JDBC sink #12089

Closed

st1page mentioned this pull request Jun 7, 2024

fix(sink): do compaction in chunk for upsert sink #17174

Merged

9 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

refactor(jdbc-sink): execute statements in batch and set isolation level to RC #12250

refactor(jdbc-sink): execute statements in batch and set isolation level to RC #12250

StrikeW commented Sep 13, 2023 •

edited

Loading

chenzl25 Sep 13, 2023

StrikeW Sep 13, 2023

st1page Sep 13, 2023

codecov bot commented Sep 13, 2023 •

edited

Loading

StrikeW commented Sep 13, 2023 •

edited

Loading

hzxa21 left a comment

wenym1 left a comment

chenzl25 left a comment

StrikeW commented Sep 14, 2023

refactor(jdbc-sink): execute statements in batch and set isolation level to RC #12250

refactor(jdbc-sink): execute statements in batch and set isolation level to RC #12250

Conversation

StrikeW commented Sep 13, 2023 • edited Loading

What's changed and what's your intention?

Checklist

Documentation

Release note

chenzl25 Sep 13, 2023

Choose a reason for hiding this comment

StrikeW Sep 13, 2023

Choose a reason for hiding this comment

st1page Sep 13, 2023

Choose a reason for hiding this comment

codecov bot commented Sep 13, 2023 • edited Loading

Codecov Report

StrikeW commented Sep 13, 2023 • edited Loading

hzxa21 left a comment

Choose a reason for hiding this comment

wenym1 left a comment

Choose a reason for hiding this comment

chenzl25 left a comment

Choose a reason for hiding this comment

StrikeW commented Sep 14, 2023

StrikeW commented Sep 13, 2023 •

edited

Loading

codecov bot commented Sep 13, 2023 •

edited

Loading

StrikeW commented Sep 13, 2023 •

edited

Loading