| From: | Thomas Munro <thomas(dot)munro(at)gmail(dot)com> |
|---|---|
| To: | 2023103793(at)ruc(dot)edu(dot)cn, pgsql-bugs(at)lists(dot)postgresql(dot)org |
| Subject: | Re: BUG #18882: Unexpected write skew at SERIALIZABLE with psycopg2.connect() |
| Date: | 2025-04-08 11:09:54 |
| Message-ID: | CA+hUKGKp7Cj3q=fPCKKS9JSW-KiGbVNpBcRo0-nd+z_5cedoZA@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs |
On Tue, Apr 8, 2025 at 9:18 PM PG Bug reporting form
<noreply(at)postgresql(dot)org> wrote:
> cur1.execute("SET SESSION CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL
> SERIALIZABLE;")
> cur1.execute("START TRANSACTION;")
I think it sends an implicit BEGIN, so you're already in a
transaction, and then START TRANSACTION (= BEGIN) is ignored, and SET
SESSION will affect the *next* transaction. I bet your log says
"WARNING: there is already a transaction in progress". I think there
is a psycopg2 way to request SERIALIZABLE, something like
conn1.set_isolation_level(extensions.ISOLATION_LEVEL_SERIALIZABLE), or
you could probably use SET TRANSACTION ISOLATION LEVEL SERIALIZABLE
instead of the SESSION version.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | PG Bug reporting form | 2025-04-08 12:00:14 | BUG #18883: Epoch issue |
| Previous Message | PG Bug reporting form | 2025-04-08 07:10:11 | BUG #18882: Unexpected write skew at SERIALIZABLE with psycopg2.connect() |