[pgpool-hackers: 1421] Re: Item #11, torward pgpool-II 3.6
Yugo Nagata
nagata at sraoss.co.jp
Mon Feb 22 16:29:05 JST 2016
On Mon, 22 Feb 2016 16:02:03 +0900 (JST)
Tatsuo Ishii <ishii at postgresql.org> wrote:
> > Do you mean if we can distiguish TCP connection failure from othres, we
> > can eliminate unnecessary proces exiting on failover?
>
> No. I am talking about a problem caused by a cable is pulled out. See
> the attached mails in the upthread email.
I'm sorry. I had read the attached mails but I wrote it incorrectly.
What I want to ask is; when a failover occurs, if we know this isn't caused
by a cable pllued out, can we handle this without killing the processes?
>
> > For example, don't we need to exit processes in the following cases?
> >
> > a. when using pcp_degenerate_node commend
>
> What is pcp_degenerate_node commend?
It's a mistake. I meant pcp_detach_node. This command is used to
degenerate the backend manually, and in this case, the failover
isn't caused by a cable plugged out.
>
> > b. when the backend send ADMIN_SHUTDOWN_ERROR_CODE ("57P01") or
> > CRASH_SHUTDOWN_ERROR_CODE "57P02"
>
> Not related to the topic I'm arguing.
When these error/notice messages are sent, a failover occurs by calling
notice_backend_error(). However, this is caused administrative shutdown
of backend and not a cable un-plugged. I think we don't need kill
processes in this case.
(pool_process_query.c)
4521 /*
4522 * admin shutdown postmaster or postmaster goes down
4523 */
4524 r = detect_postmaster_down_error(CONNECTION(backend, i), MAJOR(backend));
4525 if (r == SPECIFIED_ERROR)
4526 {
4527 ereport(LOG,
4528 (errmsg("reading and processing packets"),
4529 errdetail("postmaster on DB node %d was shutdown by administrative command", i)));
4530 /* detach backend node. */
4531 was_error = 1;
4532 if (!VALID_BACKEND(i))
4533 break;
4534 notice_backend_error(i);
4535 sleep(5);
4536 break;
4537 }
>
> > c. when we can confirm that there are no problem on TCP connection
> > from ping result
> >
> > Of course, this is the exception when the backend is used by the current session.
>
> Yes.
>
> Best regards,
> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_en.php
> Japanese:http://www.sraoss.co.jp
--
Yugo Nagata <nagata at sraoss.co.jp>
More information about the pgpool-hackers
mailing list