Re: Foreign Key locking / deadlock issue.... v2
От | rob stone |
---|---|
Тема | Re: Foreign Key locking / deadlock issue.... v2 |
Дата | |
Msg-id | 1521805381.3002.8.camel@gmail.com обсуждение исходный текст |
Ответ на | RE: Foreign Key locking / deadlock issue.... v2 (HORDER Phil <Phil.Horder@uk.thalesgroup.com>) |
Ответы |
RE: Foreign Key locking / deadlock issue.... v2
|
Список | pgsql-general |
Hello Phil, I've run your sample script on 9.6.5 and 10.3. The only thing that I added was a commit; after the initial inserts just to ensure the rows were saved. No errors were reported for either version. The output of \dp after running was:- Access privileges Schema | Name | Type | Access privileges | Column privileges | Policies --------+------+-------+-------------------+-------------------+------- ----------- public | eln | table | | | public | pl | table | | | security_policy:+ | | | | | (u): true --> including the FOR ALL in the create policy statement as well as WITH CHECK(true). Access privileges Schema | Name | Type | Access privileges | Column privileges | Policies --------+------+-------+-------------------+-------------------+------- ----------- public | eln | table | | | public | pl | table | | | security_policy:+ | | | | | (u): true + | | | | | (c): true The only mystery is what happens here:- <snip> -- …. Pause while other processing happens ….. (commit;) -- Child table processing – occurs often & quickly. Starts after parent update. <\snip> I'd like to know more about RLS and trying to de-bug your script. On a production application you'd be testing for errors and raising exceptions so as to inform users that a problem occurred. So, without knowing what occurs during "Pause while other processing happens" I can't help any further. Cheers, Rob
В списке pgsql-general по дате отправления: