Re: Shouldn't postgres_fdw report warning when it gives up getting result from foreign server?

From: Bharath Rupireddy <bharath(dot)rupireddyforpostgres(at)gmail(dot)com>
To: Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>
Cc: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Shouldn't postgres_fdw report warning when it gives up getting result from foreign server?
Date: 2021-12-03 14:04:21
Message-ID: CALj2ACUPJ9+XLsx0ws9yQ0o=vRCM4Eq8UgUrXbCxV3aNQ3+ibw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Fri, Dec 3, 2021 at 2:26 PM Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com> wrote:
> > For a cancel request maybe we can just say without te errcontext:
> > ereport(WARNING,
> > (errmsg("could not get result of cancel
> > request due to timeout")));
> >
> > See the below existing message using "cancel request":
> > errmsg("could not send cancel request: %s",
> >
> > For SQL command we can say:
> > ereport(WARNING,
> > (errmsg("could not get query result due to
> > timeout"),
> > query ? errcontext("remote SQL command:
> > %s", query) : 0));
>
> I wonder how pgfdw_get_cleanup_result() can determine which
> log message to report. Probably we can add new boolean argument
> to pgfdw_get_cleanup_result() so that it should be set to true
> for cancel request case, but false for query case. Then
> pgfdw_get_cleanup_result() can decide wihch message to log
> based on that argument. But it seems not good design to me.
> Thought?

Let's not use the boolean just for the cancel request which isn't
scalable IMO. Maybe a macro/enum?

Otherwise, we could just do, although it doesn't look elegant:

if (pgfdw_get_cleanup_result(conn, endtime, &result, "(cancel request)"))

if (strcmp(query, "(cancel request)") == 0)
WARNING without "remote SQL command:
else
WARNING with "remote SQL command:

Regards,
Bharath Rupireddy.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Ashutosh Sharma 2021-12-03 14:08:45 Re: [Proposal] Fully WAL logged CREATE DATABASE - No Checkpoints
Previous Message Andrei Zubkov 2021-12-03 14:03:46 Re: [PATCH] Tracking statements entry timestamp in pg_stat_statements