-
Notifications
You must be signed in to change notification settings - Fork 5.8k
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
information_schema.cluster_slow_query returns wrong result #56100
Comments
Only return data for the same date MySQL root@127.0.0.1:test> select time,query from information_schema.cluster_slow_query where time > '2024-09-06 00:00:01.000000' limit 1;
+----------------------------+-------------------------------------+
| time | query |
+----------------------------+-------------------------------------+
| 2024-09-06 14:54:33.064918 | CREATE DATABASE IF NOT EXISTS test; |
+----------------------------+-------------------------------------+
1 row in set
Time: 0.102s
MySQL root@127.0.0.1:test> select time,query from information_schema.cluster_slow_query where time > '2024-09-15 00:00:01.000000' limit 1;
+----------------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| time | query |
+----------------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| 2024-09-15 00:00:12.485027 | select job_meta, processing from mysql.tidb_ddl_job where job_id in (select min(job_id) from mysql.tidb_ddl_job group by schema_ids, table_ids, processing) and not reorg order by processing desc, job_id; |
+----------------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set
Time: 0.017s
MySQL root@127.0.0.1:test> select time,query from information_schema.cluster_slow_query where time > '2024-09-16 00:00:01.000000' limit 1;
+----------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| time | query |
+----------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| 2024-09-16 00:04:56.937909 | INSERT HIGH_PRIORITY INTO mysql.tidb VALUES ('tikv_gc_leader_lease', '20240916-00:06:51.931 +0800', 'Current GC worker leader lease. (DO NOT EDIT)') ON DUPLICATE KEY UPDATE variable_value = '20240916-00:06:51.931 +0800', comment = 'Current GC worker leader lease. (DO NOT EDIT)'; |
+----------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ |
/assign |
/assign |
The root cause of this issue may lie in the unreasonable implementation of #14840. At that time, I was worried that unreasonable queries would cause a lot of slow log file parsing. which would consume a lot of TIDB CPU resources. Specifically, when querying the
@songrijie PTAL, do you think the above (current) behavior is reasonable? And if not, What is reasonable behavior? |
@crazycs520 I guess it should be up to the caller, what range they want to select. Even with the current limitation they could set any range they want and cause lots of parsing (if they specify both lower and upper bounds), so current logic doesn't really prevent this situation, but makes the behavior very confusing (also it's not documented anywhere as far as I understand). |
tidb query is limited by max_execution_time and tidb_mem_quota_query. Query on information_schema.cluster_slow_query should follow same rule, so it's easy for user to understand and manage. |
I prefer to force the user to specify the time range instead of any implicit conditions behavior. That is, if a user does not specify a start or end time, an error should returned. |
The option with explicit error (if that behavior is documented) could be better, but what about cases when a user specified both start and end times, but the range is still too broad to handle efficiently? It's not different from not specifying start or end time by end result. |
I think a better option would be to rely on quotas / time limits as @mzhang77 pointed out above, without any SQL-level restrictions. |
Bug Report
Please answer these questions before submitting your issue. Thanks!
1. Minimal reproduce step (Required)
2. What did you expect to see? (Required)
The second query should return result if the first query returns result
3. What did you see instead (Required)
4. What is your TiDB version? (Required)
The text was updated successfully, but these errors were encountered: