You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In large schemas, loading check constraints can cause significant querying overhead (see also #8463). When debug logging a code generation run on a large Oracle schema, we've seen:
15:03:04,921 [DEBUG] Executing query : select "ac"."OWNER", "ac"."TABLE_NAME", "ac"."CONSTRAINT_NAME", "ac"."SEARCH_CONDITION" from "SYS"."ALL_CONSTRAINTS" "ac" where ("ac"."CONSTRAINT_TYPE" = ? and "ac"."OWNER" in (?, ?, ?, ?))
15:03:29,559 [DEBUG] Fetched result : +-----+----------+---------------+---------------------+
15:03:29,559 [DEBUG] : |OWNER|TABLE_NAME|CONSTRAINT_NAME|SEARCH_CONDITION |
15:03:29,560 [DEBUG] : +-----+----------+---------------+---------------------+
15:03:29,560 [DEBUG] : |SYS |TS$ |SYS_C0061 |CONTENTS$ IS NOT NULL|
15:03:29,560 [DEBUG] : |SYS |TS$ |SYS_C0060 |ONLINE$ IS NOT NULL |
15:03:29,561 [DEBUG] : |SYS |UNDO$ |SYS_C0052 |NAME IS NOT NULL |
15:03:29,561 [DEBUG] : |SYS |SEG$ |SYS_C0043 |MAXEXTS IS NOT NULL |
15:03:29,561 [DEBUG] : |SYS |TS$ |SYS_C0069 |DFLEXTPCT IS NOT NULL|
15:03:29,562 [DEBUG] : +-----+----------+---------------+---------------------+
15:03:29,562 [DEBUG] : |...record(s) truncated...
15:03:29,562 [DEBUG] Fetched row(s) : 14005
The time between the beginning and the end of result fetching has not been considered for a log message via org.jooq.meta.AbstractDatabase$1$SQLPerformanceWarning, so this kind of problem has gone undetected so far.
We should log this as well, so users can assess whether to possibly change their code generation configuration to prevent such large result sets.
This logging should be accompanied by a new configuration flag <logSlowResultsAfterSeconds/>. See #8465
The text was updated successfully, but these errors were encountered:
In large schemas, loading check constraints can cause significant querying overhead (see also #8463). When debug logging a code generation run on a large Oracle schema, we've seen:
The time between the beginning and the end of result fetching has not been considered for a log message via
org.jooq.meta.AbstractDatabase$1$SQLPerformanceWarning
, so this kind of problem has gone undetected so far.We should log this as well, so users can assess whether to possibly change their code generation configuration to prevent such large result sets.
This logging should be accompanied by a new configuration flag
<logSlowResultsAfterSeconds/>
. See #8465The text was updated successfully, but these errors were encountered: