MySQL Handler Kill Query #5448

TCeason · 2022-05-18T13:57:27Z

I hereby agree to the terms of the CLA available at: https://databend.rs/dev/policies/cla/

Summary

MySQL Client support according to "Ctrl C" interrupt a long query.
Link to discussion: #5405

Changelog

New Feature

Related Issues

Fixes #4871

vercel · 2022-05-18T13:57:33Z

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated
databend	✅ Ready (Inspect)	Visit Preview	May 19, 2022 at 8:32AM (UTC)

mergify · 2022-05-18T13:57:59Z

Thanks for the contribution!
I have applied any labels matching special text in your PR Changelog.

Please review the labels and make any necessary changes.

sundy-li · 2022-05-18T14:22:13Z

query/src/sessions/session.rs

@@ -63,16 +65,38 @@ impl Session {
        let session_settings = Settings::try_create(&conf)?;
        let ref_count = Arc::new(AtomicUsize::new(0));
        let status = Arc::new(Default::default());
+        match typ {
+            SessionType::MySQL => {
+                let mysql_conn_id = session_mgr


It's incorrect, we must save the mysql_conn_id in fetch_add

saved in there: self.mysql_basic_conn_id.fetch_add(1, Ordering::Relaxed);

Maybe we can directly use session_mgr.mysql_basic_conn_id?

Must save the return value of fetch_add, see https://doc.rust-lang.org/std/sync/atomic/struct.AtomicU32.html#method.fetch_add

Otherwise, you can't ensure the load value is not changed.

sundy-li · 2022-05-19T01:56:51Z

Maybe we can add a test with native_client ?

TCeason · 2022-05-19T03:34:06Z

Maybe we can add a test with native_client ?

Like this tests/suites/1_stateful/02_query/02_0000_kill_query.py?

sundy-li · 2022-05-19T05:21:51Z

tests/suites/1_stateful/02_query/02_0000_kill_query.py

+    time.sleep(5)
+    mycursor.execute("SELECT * FROM system.processes WHERE extra_info like '%select * from numbers(999999999)%' and extra_info not like '%system.processes%'")
+    res = mycursor.fetchone()
+    assert res is None


result file is empty, the test will be skipped.

BohuTANG · 2022-05-19T08:10:27Z

02_0000_kill_query:                                                     [ FAIL ] - return code 1
, result:

Traceback (most recent call last):
  File "/home/runner/work/databend/databend/tests/helpers/uexpect.py", line 178, in read
    data += self.queue.get(timeout=timeleft)
  File "/usr/lib/python3.8/queue.py", line 178, in get
    raise Empty
_queue.Empty

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/runner/work/databend/databend/tests/helpers/uexpect.py", line 152, in expect
    data = self.read(timeout=timeleft, raise_exception=True)
  File "/home/runner/work/databend/databend/tests/helpers/uexpect.py", line 186, in read
    raise TimeoutError(timeout)
uexpect.TimeoutError: Timeout 60.000s

Link to discussion: databendlabs#5405

BohuTANG · 2022-05-19T10:30:51Z

From my test on local, the server(MySQL handler) received the kill query from the client, but databend-query select cannot be canceled:

2022-05-19T10:25:28.715810Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: select sum(number) from numbers(10000000000)
2022-05-19T10:25:30.322079Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46846)
2022-05-19T10:25:30.626255Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9
2022-05-19T10:25:33.547171Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46848)
2022-05-19T10:25:33.765910Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9
2022-05-19T10:26:16.341979Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46850)
2022-05-19T10:26:16.625610Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9
2022-05-19T10:26:16.692260Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46852)
2022-05-19T10:26:16.933850Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9

TCeason · 2022-05-19T13:19:17Z

From my test on local, the server(MySQL handler) received the kill query from the client, but databend-query select cannot be canceled:

2022-05-19T10:25:28.715810Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: select sum(number) from numbers(10000000000)
2022-05-19T10:25:30.322079Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46846)
2022-05-19T10:25:30.626255Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9
2022-05-19T10:25:33.547171Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46848)
2022-05-19T10:25:33.765910Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9
2022-05-19T10:26:16.341979Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46850)
2022-05-19T10:26:16.625610Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9
2022-05-19T10:26:16.692260Z  INFO databend_query::servers::mysql::mysql_handler: MySQL connection coming: Ok(127.0.0.1:46852)
2022-05-19T10:26:16.933850Z  INFO databend_query::servers::mysql::mysql_interactive_worker: Normal query: KILL QUERY 9

Actually this pr just do one thing : convert mysql_conn_id to session_id.

I do this test:

in session 1

select sum(number) from numbers(10000000000);

in session 2

root@mysqldb 21:15:57 [(none)]> show processlist;
+--------------------------------------+-------+-----------------+------+-------+----------+----------------------------------------------+--------------+------------------------+-------------------------+-------------------------+--------------------------+---------------------+
| id                                   | type  | host            | user | state | database | extra_info                                   | memory_usage | dal_metrics_read_bytes | dal_metrics_write_bytes | scan_progress_read_rows | scan_progress_read_bytes | mysql_connection_id |
+--------------------------------------+-------+-----------------+------+-------+----------+----------------------------------------------+--------------+------------------------+-------------------------+-------------------------+--------------------------+---------------------+
| 6297888a-668f-4e20-a7ce-adc8fadbed85 | MySQL | 127.0.0.1:51820 | root | Query | default  | select sum(number) from numbers(10000000000) |            0 |                      0 |                       0 |               124320000 |                994560000 |                  28 |
| 0f92fe22-0015-47ac-85c5-5463a94f9cbe | MySQL | 127.0.0.1:51814 | root | Query | default  | show processlist                             |            0 |                      0 |                       0 |                       0 |                        0 |                  25 |
| 3ac8bcbf-6a13-4859-b276-7f81570fb871 | MySQL | 127.0.0.1:51802 | root | Idle  | default  | NULL                                         |            0 |                   NULL |                    NULL |                    NULL |                     NULL |                  20 |
+--------------------------------------+-------+-----------------+------+-------+----------+----------------------------------------------+--------------+------------------------+-------------------------+-------------------------+--------------------------+---------------------+
3 rows in set (0.10 sec)
Read 3 rows, 631.00 B in 0.052 sec., 57.85 rows/sec., 11.88 KiB/sec.

root@mysqldb 21:16:26 [(none)]> kill query '6297888a-668f-4e20-a7ce-adc8fadbed85';
Query OK, 0 rows affected (0.07 sec)
Read 0 rows, 0.00 B in 0.003 sec., 0 rows/sec., 0.00 B/sec.

And then the session 1 's query looks like not exit, but if already exit.


root@mysqldb 21:16:33 [(none)]> show processlist;
+--------------------------------------+-------+-----------------+------+-------+----------+------------------+--------------+------------------------+-------------------------+-------------------------+--------------------------+---------------------+
| id                                   | type  | host            | user | state | database | extra_info       | memory_usage | dal_metrics_read_bytes | dal_metrics_write_bytes | scan_progress_read_rows | scan_progress_read_bytes | mysql_connection_id |
+--------------------------------------+-------+-----------------+------+-------+----------+------------------+--------------+------------------------+-------------------------+-------------------------+--------------------------+---------------------+
| 6297888a-668f-4e20-a7ce-adc8fadbed85 | MySQL | 127.0.0.1:51820 | root | Idle  | default  | NULL             |            0 |                   NULL |                    NULL |                    NULL |                     NULL |                  28 |
| 0f92fe22-0015-47ac-85c5-5463a94f9cbe | MySQL | 127.0.0.1:51814 | root | Query | default  | show processlist |            0 |                      0 |                       0 |                       0 |                        0 |                  25 |
| 3ac8bcbf-6a13-4859-b276-7f81570fb871 | MySQL | 127.0.0.1:51802 | root | Idle  | default  | NULL             |            0 |                   NULL |                    NULL |                    NULL |                     NULL |                  20 |
+--------------------------------------+-------+-----------------+------+-------+----------+------------------+--------------+------------------------+-------------------------+-------------------------+--------------------------+---------------------+
3 rows in set (0.17 sec)
Read 3 rows, 586.00 B in 0.094 sec., 31.86 rows/sec., 6.08 KiB/sec.

TCeason · 2022-05-19T13:21:12Z

maybe you could off the new processor.

root@mysqldb 21:19:33 [(none)]> SET enable_new_processor_framework=0;
Query OK, 0 rows affected (0.12 sec)
Read 0 rows, 0.00 B in 0.020 sec., 0 rows/sec., 0.00 B/sec.

root@mysqldb 21:19:37 [(none)]> select sum(number) from numbers(10000000000);
^C^C -- query aborted
ERROR 1152 (08S01): Code: 1043, displayText = Aborted query, because the server is shutting down or the query was killed.
root@mysqldb 21:19:40 [(none)]>

I met a similar situation, but I haven't submitted the issue yet.

TCeason · 2022-05-19T14:34:13Z

maybe you could off the new processor.

root@mysqldb 21:19:33 [(none)]> SET enable_new_processor_framework=0;
Query OK, 0 rows affected (0.12 sec)
Read 0 rows, 0.00 B in 0.020 sec., 0 rows/sec., 0.00 B/sec.

root@mysqldb 21:19:37 [(none)]> select sum(number) from numbers(10000000000);
^C^C -- query aborted
ERROR 1152 (08S01): Code: 1043, displayText = Aborted query, because the server is shutting down or the query was killed.
root@mysqldb 21:19:40 [(none)]>

I met a similar situation, but I haven't submitted the issue yet.

And I find if let the query in session 1 execute a long time, and kill it in session 2 the session 1 will at aborting state.

TCeason requested a review from BohuTANG as a code owner May 18, 2022 13:57

databend-bot added the need-review label May 18, 2022

mergify bot added the pr-feature this PR introduces a new feature to the codebase label May 18, 2022

vercel bot deployed to Preview May 18, 2022 13:58 View deployment

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from a675b5a to c571dd1 Compare May 18, 2022 14:01

vercel bot deployed to Preview May 18, 2022 14:02 View deployment

BohuTANG requested review from zhang2014 and sundy-li and removed request for BohuTANG May 18, 2022 14:03

sundy-li reviewed May 18, 2022

View reviewed changes

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from c571dd1 to da493c2 Compare May 18, 2022 15:01

vercel bot deployed to Preview May 18, 2022 15:02 View deployment

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from da493c2 to b9e58a9 Compare May 19, 2022 03:33

vercel bot deployed to Preview May 19, 2022 03:34 View deployment

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from b9e58a9 to cfd36f2 Compare May 19, 2022 04:36

vercel bot deployed to Preview May 19, 2022 04:38 View deployment

sundy-li reviewed May 19, 2022

View reviewed changes

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from cfd36f2 to 53ff418 Compare May 19, 2022 05:41

vercel bot deployed to Preview May 19, 2022 05:42 View deployment

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from 2139a16 to a1c4afe Compare May 19, 2022 06:33

vercel bot deployed to Preview May 19, 2022 06:34 View deployment

sundy-li approved these changes May 19, 2022

View reviewed changes

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from a1c4afe to 354b261 Compare May 19, 2022 08:30

MySQL Handler Kill Query

c7b0b05

Link to discussion: databendlabs#5405

vercel bot deployed to Preview May 19, 2022 08:31 View deployment

TCeason force-pushed the ISSUE-4871/handle_mysql_kill branch from 354b261 to c7b0b05 Compare May 19, 2022 08:31

vercel bot deployed to Preview May 19, 2022 08:32 View deployment

BohuTANG merged commit 40114c6 into databendlabs:main May 19, 2022

BohuTANG mentioned this pull request May 19, 2022

flaky 02_0000_kill_query #5472

Closed

sundy-li mentioned this pull request May 20, 2022

Feature: make executor.finish() respect abort signal #5481

Closed

TCeason deleted the ISSUE-4871/handle_mysql_kill branch June 9, 2022 02:32

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

MySQL Handler Kill Query #5448

MySQL Handler Kill Query #5448

TCeason commented May 18, 2022 •

edited by BohuTANG

Loading

vercel bot commented May 18, 2022 •

edited

Loading

mergify bot commented May 18, 2022

sundy-li May 18, 2022

TCeason May 18, 2022

sundy-li May 18, 2022

sundy-li commented May 19, 2022

TCeason commented May 19, 2022

sundy-li May 19, 2022

BohuTANG commented May 19, 2022

BohuTANG commented May 19, 2022 •

edited

Loading

TCeason commented May 19, 2022

TCeason commented May 19, 2022

TCeason commented May 19, 2022

MySQL Handler Kill Query #5448

MySQL Handler Kill Query #5448

Conversation

TCeason commented May 18, 2022 • edited by BohuTANG Loading

Summary

Changelog

Related Issues

vercel bot commented May 18, 2022 • edited Loading

mergify bot commented May 18, 2022

sundy-li May 18, 2022

Choose a reason for hiding this comment

TCeason May 18, 2022

Choose a reason for hiding this comment

sundy-li May 18, 2022

Choose a reason for hiding this comment

sundy-li commented May 19, 2022

TCeason commented May 19, 2022

sundy-li May 19, 2022

Choose a reason for hiding this comment

BohuTANG commented May 19, 2022

BohuTANG commented May 19, 2022 • edited Loading

TCeason commented May 19, 2022

TCeason commented May 19, 2022

TCeason commented May 19, 2022

TCeason commented May 18, 2022 •

edited by BohuTANG

Loading

vercel bot commented May 18, 2022 •

edited

Loading

BohuTANG commented May 19, 2022 •

edited

Loading