Re: Invalidate the subscription worker in cases where a user loses their superuser status

Поиск
Список
Период
Сортировка
От Peter Smith
Тема Re: Invalidate the subscription worker in cases where a user loses their superuser status
Дата
Msg-id CAHut+PuK3Q9i0e0co+CZv=Wnkjh1e96VNxYPg_85DZ7_G0b+hg@mail.gmail.com
обсуждение исходный текст
Ответ на Re: Invalidate the subscription worker in cases where a user loses their superuser status  (vignesh C <vignesh21@gmail.com>)
Ответы Re: Invalidate the subscription worker in cases where a user loses their superuser status  (vignesh C <vignesh21@gmail.com>)
Re: Invalidate the subscription worker in cases where a user loses their superuser status  (vignesh C <vignesh21@gmail.com>)
Список pgsql-hackers
Here are some comments for patch v2-0001.

======
src/backend/replication/logical/worker.c

1. maybe_reread_subscription

ereport(LOG,
        (errmsg("logical replication worker for subscription \"%s\"
will restart because of a parameter change",
                MySubscription->name)));

Is this really a "parameter change" though? It might be a stretch to
call the user role change a subscription parameter change. Perhaps
this case needs its own LOG message?

======
src/include/catalog/pg_subscription.h

2.
  char    *origin; /* Only publish data originating from the
  * specified origin */
+ bool isownersuperuser; /* Is subscription owner superuser? */
 } Subscription;


Is a new Subscription field member really necessary? The Subscription
already has 'owner' -- why doesn't function
maybe_reread_subscription() just check:

(!superuser_arg(sub->owner) && superuser_arg(MySubscription->owner))

======
src/test/subscription/t/027_nosuperuser.pl

3.
# The apply worker should get restarted after the superuser prvileges are
# revoked for subscription owner alice.

typo

/prvileges/privileges/

~

4.
+# After the user becomes non-superuser the apply worker should be restarted and
+# it should fail with 'password is required' error as password option is not
+# part of the connection string.

/as password option/because the password option/

======
Kind Regards,
Peter Smith.
Fujitsu Australia.



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Laurenz Albe
Дата:
Сообщение: Re: Regression in COPY FROM caused by 9f8377f7a2
Следующее
От: Michael Paquier
Дата:
Сообщение: Fail hard if xlogreader.c fails on out-of-memory