Re: pg_terminate_backend can terminate background workers and autovacuum launchers

On Wed, Jun 21, 2017 at 7:56 AM, Yugo Nagata <nagata(at)sraoss(dot)co(dot)jp> wrote:
> I have found that we can cancel/terminate autovacuum launchers and
> background worker processes by pg_cancel/terminate_backend function.
> I'm wondering this behavior is not expected and if not I want to fix it.

I think it is expected. Even if we blocked it, those processes have
to cope gracefully with SIGTERM, because anyone with access to the OS
user can kill them that way by hand.

> However, we can terminate background workers by pg_terminate_backend.
> In the following example, I terminated the logical replication launcher,
> and this process did not appear again[1].
>
> postgres=# select pg_terminate_backend(30902);
> pg_terminate_backend
> ----------------------
> t
> (1 row)

That seems to be a bug in logical replication.

> Similarly, we can terminate autovacuum launcher by pg_terminate_backend,
> but a new process is restarted by postmaster in this case.[2]
>
> postgres=# select pg_terminate_backend(30900);
> pg_terminate_backend
> ----------------------
> t
> (1 row)

That is as I would expect.

> [2]
> On the other hand, when we use pg_cancel_backend for autovacuum launcher,
> it causes the following error. I'll report the detail in another thread.
>
> ERROR: can't attach the same segment more than once

I think that's a bug.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company